1 A TWO SYSTEM, TWO AFFECT FRAMEWORK OF CONSUMPTION EXPERIENCE By DENNIS DANNY ONG UY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE O F DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Dennis Danny Ong Uy
3 To my Mom who passed away unexpectedly, and to my dad, whom I hope to share more memories in this finite time together
4 ACKNOWLEDGMENTS I thank my mom (Christ ine, deceased) and my dad (Peter) for raising me to become who I am today, and for their ability to squeeze out so much out of so little. I also thank my sister (Joy) for her unwavering support. I thank Chris Janiszewski for his advice and for guiding me i n the process of acquiring scientific knowledge. I also thank my doctoral dissertation committee members: Andreas Keil, Robyn LeBoeuf, and Alan Sawyer for their helpful comments. Lastly, I thank Albert Einstein, whose papers in 1905 inspired my way of thinking about the cosmos and our place in it.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 10 LIST OF DEFINITIONS ................................................................................................. 11 ABSTRACT ................................................................................................................... 12 CHAPTER 1 A TWO AFFECT, TWO CONSUMPTION SYSTEM FRAMEWORK ...................... 14 Introduction ............................................................................................................. 14 Theoretical Framework ........................................................................................... 14 2 TWO CONSUMPTION SYSTEMS.......................................................................... 22 System 1 The Sensory Consumption System ...................................................... 22 A Broader View o f Sensory Consumption ........................................................ 23 Structure of the sensory consumption system ........................................... 24 The opponent process structure of sensory channels ................................ 24 The opponent process structure of the visual channel ............................... 25 The gustatory channel ............................................................................... 27 Bottom up pathways of gustatory processing ............................................ 27 Two schools of thought in gustatory encoding ........................................... 28 Reconciling the two schools of thought ...................................................... 29 Predictive property of sensory processing ................................................. 31 Response properties of the sensory consumption system pleasure pathway .................................................................................................. 33 Two types of sensitization and habituation implications for contrast and pleasure generation ......................................................................... 35 System 2 The Semantic Consumption System ...................................................... 38 Structure of the Semantic Consumption System ........................................ 40 3 THE TWO AFFECTIVE SYSTEMS ........................................................................ 49 Hedonic Affective System ....................................................................................... 50 The Heuristic Value of Sensory Qualities in the TwoAffect Organism ................... 52 4 OPERATING PROPERTIES OF THE TWO SYSTEM CONSUMER ...................... 55 Four Pathways ........................................................................................................ 55
6 First Pathway: Sensory to Hedonic Affect Pathway .......................................... 55 Second Pathway: Response Properties of the Semantic Consumption System Semantic Generation Pathway ...................................................... 58 The role of B ayesian prediction in pleasure ............................................... 63 5 AFFECT GATING ................................................................................................... 65 How Affect States Influence Relevance of Sensory Channels ................................ 65 Two Types of Affect ................................................................................................ 68 Affective Gating ................................................................................................ 70 Negative State Affect ................................................................................. 71 Positive State Affect ................................................................................... 72 Evidence for Affective Gating ..................................................................... 74 Hypotheses ...................................................................................................... 75 Experiment 1 ........................................................................................................... 77 Design and Procedure ...................................................................................... 77 Design ........................................................................................................ 77 Procedure ......................................................................................................... 77 Product description coding ......................................................................... 79 Velten procedure manipulation check ........................................................ 79 Results ............................................................................................................. 80 Discussion ........................................................................................................ 80 Experiment 2 ........................................................................................................... 81 Design and Procedure ...................................................................................... 82 Results ............................................................................................................. 82 Primary analysis ......................................................................................... 82 Mediation analysis for negative affective state participants ....................... 84 Lack of mediation for positive affective state participants .......................... 85 Discussion ........................................................................................................ 85 Experiment 3 ........................................................................................................... 86 Design and Procedure ...................................................................................... 87 Design ........................................................................................................ 87 Procedure .................................................................................................. 87 Results ............................................................................................................. 87 Primary analysis ......................................................................................... 87 Mediation analysis for negative affective state participants ....................... 89 Discussion ........................................................................................................ 90 Experiment 4 ........................................................................................................... 91 Design and Procedure ...................................................................................... 91 Design .............................................................................................................. 91 Procedure .................................................................................................. 92 Results ............................................................................................................. 92 Primary analysis ......................................................................................... 92 Mediati on analysis for negative affective state participants ....................... 93 Discussion ........................................................................................................ 94 Experiment 5a ......................................................................................................... 96 Design and Procedure ...................................................................................... 96
7 Design ........................................................................................................ 96 Procedure .................................................................................................. 97 Results ............................................................................................................. 97 Discussion ........................................................................................................ 97 Experiment 5b ......................................................................................................... 97 Design and Proc edure ...................................................................................... 98 Design ........................................................................................................ 98 Procedure and Stimuli ................................................................................ 98 Results ............................................................................................................. 98 Primary analysis ......................................................................................... 98 Mediation analysis for negative affective state participants ..................... 100 Discussion ...................................................................................................... 100 General Discussion ............................................................................................... 101 6 AFFECT TUNING ................................................................................................. 108 Affective Influences on Interpreting Sensory Qualities As Rewarding Or Punishing ........................................................................................................... 108 Taste Perception ................................................................................................... 110 Affective System ................................................................................................... 113 Experiment 1 ......................................................................................................... 116 Design and Procedure .................................................................................... 117 Design ...................................................................................................... 117 Procedure ................................................................................................ 117 Coding ..................................................................................................... 118 Results ........................................................................................................... 119 Manipulation check .................................................................................. 119 Primary Analysis ...................................................................................... 119 Mediation analysis for negative affect consumers .................................... 120 Mediation analysis for positive affect consumers ..................................... 120 Discussion ...................................................................................................... 121 Experiment 2 ......................................................................................................... 122 Design and Procedure .................................................................................... 122 Design ...................................................................................................... 122 Stimuli ...................................................................................................... 122 Procedure ................................................................................................ 123 Results ........................................................................................................... 123 Mediation analysis for negative affect consumers and complex chips ..... 125 Mediation analysis for positive affect consumers and complex chips ...... 126 Ancillary mediation analyses .................................................................... 127 Discussion ...................................................................................................... 127 Experiment 3 ......................................................................................................... 128 Design and Procedure .................................................................................... 129 Design ...................................................................................................... 129 Stimuli ...................................................................................................... 129 Procedure ................................................................................................ 129 Results ........................................................................................................... 129
8 Primary analysis ....................................................................................... 129 Supplemental Analysis ............................................................................. 131 Mediation analysis for negative affect consumers experiencing a simple product .................................................................................................. 132 Mediation analysis for positive affect consumers experiencing a simple product .................................................................................................. 132 Mediation analysis for negative affect consumers experiencing a high complexity product ................................................................................ 132 Mediation analysis for positive affect consumers experiencing a high complexity product ................................................................................ 133 Discussion ...................................................................................................... 133 General Discussion ............................................................................................... 134 LIST OF REFERENCES ............................................................................................. 143 BIOGRAPHICAL SKETCH .......................................................................................... 159
9 LIST OF TABLES Table page 1 1 Consumer Phenomena/Conceptual Questions and Managerial Implications. .... 21 2 1 Timeline of Consumers on Earth. ....................................................................... 22 2 2 Fundamental Pleasures that appear to have sensory antecedents. ................. 43 2 3 Higher order Pleasures that appear to have semantic antecedents (in addition to sensory antecedents). ....................................................................... 43 2 4 Consumer feedback on semantic consumption, and implied neural response. .. 47 4 1 Opponent sensory categories. ............................................................................ 58 4 2 Opponent semantic categories ........................................................................... 64 5 1 The evaluation of skin lotion under different affective states. ............................. 84 5 2 The evaluation of different qualities of skin lotion under different affective states. ................................................................................................................. 88 5 3 The evaluation of skin lotion under different affectiv e states and visual channel blocked or unblocked (experiment 4). ................................................... 93 5 4 The evaluation of regular or warming skin lotion under different affective states (Experiment 5B). ...................................................................................... 99 6 1 The influence of affective state on the consumers ability to perceive flavors, palatability of the food, and the consumers willingness to pay. ....................... 120 6 2 The influence of affective state on the ability to perceive specific tastes and the willingness to pay. ...................................................................................... 125 6 3 The influence of affective state on the ability to perceive tastes, palatability, and willingness to pay. ..................................................................................... 131
10 LIST OF FIGURES Figure page 1 1 Aristippus of the Cyrenaic School of Physical Hedonism. .................................. 16 1 2 Zeno of Citium, champion of the Stoic view of happiness. ................................. 19 2 1 The Process of Visual processing. ..................................................................... 25 2 2 The visible portion of the electromagnetic spectrum. .......................................... 26 2 3 The Sensory Perceptual System as a Bayesian Probability calculator ............... 33 2 4 Pathways from the semantic system toward the pleasure system. ..................... 41
11 LIST OF DEFINITIONS Eudaimonia A life is welllived, embedded in meaningful values, together with a sense of engagement in that life. This is the cognitive or Aristotelian ingredient of happiness. We speculate here about the possibility that cortical systems of self and cognitive appraisal, such as the default network, make a eudaimonic contribution to happiness. At a mor e basic brain level, the motivational component of reward, mesolimbic wanting or incentive salience, might also contribute to eudaimonic engagement by adding attraction or zest to life; however, overly intense wanting leads to unhappiness and becomes m ore akin to addiction (Kringelbach and Berridge 2010). Hedonic Affect A pleasurable feeling generated by the pleasure circuits in the brain, usually starting from the orbitofrontal cortex, which sends forward activations to the ventral pallidum and nucleus accumbens. Loose colloquial equivalents to the term hedonic affect are pleasure and liking, although many usages of the terms pleasure and liking are confused with the other type of affect, state affect. Hedonia The conscious feelings of pleasant wellbeing or hedonic niceness that is the affective ingredient of happiness. We suggest here subjective hedonia occurs when orbitofrontal and related cortical brain systems elaborate core liking reactions (which are produced by hedonic hotspots and whic h need not be conscious) into conscious feelings of positive affect (Kringelbach and Berridge 2009). State Affect A feeling generated by the desire circuits in the brain, usually starting from the brain stem (substantia nigra pars compacta). Loose colloquial equivalents to the term state affect are desire, and wanting, although many usages of the terms desire and wanting are confused with the other type of affect, hedonic affect.
12 Abstract of Dissertation Presented to the Graduate School of the U niversity of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A TWO SYSTEM, TWO AFFECT FRAMEWORK OF CONSUMPTION EXPERIENCE By Dennis Danny Ong Uy May 2010 Chair: Chris Janiszewski Major: Marketing This dissertation uses a twosystem, two affect approach to explore various consumer phenomena. Emerging literature suggests that there are two distinct affective systems, as well as two distinct consumption systems, in the consumer brain. There are two d istinct affective systems in the mammalian brain: one for generating desire, and another for generating pleasure. The antecedents and consequence of consumer desire and consumer pleasure are hence distinct. A synthesis of a separate body of literature suggests that there are two distinct consumption systems: one that extracts pleasure from sensory stimulation, and another that extracts pleasure from semantic stimulation. The antecedents and consequences of sensory consumption and semantic consumption are hence also distinct. Using this two affect, two system framework, the author uses six chapters to theorize and explore interesting consumer phenomena. The first chapter uses philosophical foundations from Ancient Greece (e.g., Aristippus of Cyrene, Zeno of Citium, Aristotle) to motivate the theoretical framework of different sources of pleasure and happiness in consumption. The second chapter uses the literature on perception and semantic memory to motivate the processes of extracting sensory and semantic
13 stimuli from the environment. The third chapter uses the literature on orbitofrontal cortex, nucleus accumbens, and ventral pallidum to argue for two distinct affective systems that generate separate kinds of feel states that provide different kinds of inf ormation value to the consumer. The fourth chapter examines some of the operating properties of the twosystem, two affect framework. The fifth and sixth chapters are empirical chapters that test a portion of the predictions that flow out from the framewor k.
14 CHAPTER 1 A TWO AFFECT, TWO CONSUMPTION SYSTEM FRAMEWORK Introduction Consumers enjoy a wide variety of products and services soda, potato chips, lotions, hamburgers, movies, TV programs, books, art, and other objects of consumption. Although these consumption episodes are commonplace, the processes underlying their consumption are still shrouded in mystery. Why do consumers enjoy certain kinds of potato chips and not others? Why do consumers enjoy certain kinds of paintings but not others? What i s the difference with the way consumers extract pleasure from products such as movies and paintings, from the way consumers extract pleasure from products such as soda and ice cream? Why does the enjoyment of certain products (e.g., comfort food) depend on the consumers prevailing mood, such that certain kinds of foods are desired a lot (and avoided a lot) when consumers are sad? This dissertation proposes a parsimonious twoaffect, two consumption system that can account for a wide range of consumption phenomena. It provides theoretical background in the first four chapters, and then provides empirical support for how the framework can be used to account for various consumer phenomena. T heoretical F ramework The theoretical framework that organizes the var ious consumption phenomena that will be reported in this dissertation has two components: An affective system containing two subsystems (state affect system and hedonic affect system), and a consumption system containing two subsystems (sensory consumption system and semantic consumption system). Together, these two systems control the consumers desire and lack of desire for an object, and control the consumers generation of
15 pleasure and aversion for an object, based on the objects sensory or semantic qualities. The insights generated from this dissertation, and the subsequent research that it aims to stimulate, can help marketers understand which sensory and semantic aspects of the product should be improved (and which can be ignored), and which sensory and semantic aspects of advertising and other marketing communications should be used in order to extract the maximum amount of desire and pleasure generation from the consumer. Notions of sensory and semantic consumption as pillars of the good life have actually begun as early as the 4th Century B.C., when Aristippus of Cyrene, one of the students of Socrates, first championed the view of hedonism as linked to the pursuit of physical pleasure. This view lasted for a century, before the Stoics added mental aspects to the view of good life. The next two sections discuss the theoretical precursors of sensory and semantic consumption, founded on the philosophical schools of Hedonism and Stoicism. Hedonism and the Philosophical Precursors of Sensory Consumpti on The Cyrenaics (e.g., Aristippus of Cyrene, 435 to 356 BC), one of the earliest Socratic schools in Greek Philosophy, championed the first, and most fundamental, form of Hedonism. In the Cyrenaic school of hedonism, individual physical pleasures were the only intrinsic good, and physical pain was the only intrinsic evil. The Cyrenaics also believed that momentary, real time physical pleasures are more intense than memory or anticipation (of the physical pleasure). Memory or anticipation of the physical pleasure, which exists in the past and future, are deemed to be nonexistent. Hence, this view of consumption championed managing consumption experiences in such a
16 way that real time consumption of physical sensations is maximized. This management of consu mption would include actively shaping circumstances to oneself, which meant that decisions should be made to maximize the number of present consumptions. Figure 11 Aristippus of the Cyrenaic School of Physical Hedonism. [Reprinted from T. Tomisti, 2 006. This image is in the public domain because its copyright has expired.] The Cyrenaic view of hedonism foreshadows an important theoretical contribution in contemporary neur oscience. It posited that there are no meaningful qualitative distinctions between present consumptions of physical pleasure, a view that foreshadowed the now established neural view of a common neural currency for sensory pleasure (Rolls 2005; Kringelbach and Berridge 2010). In the Greek era dominated by the classical five senses distinction proposed by Aristotle, the Cyrenaic view implied that the ostensibly distinct consumption from different sensory channels are in some ways fungible, in so far as t he goal of consumption is the obtainment of physical pleasure.
17 Hedonism II The Epicurean View One century after Aristippus of Cyrene founded the original view of hedonism, Epicurus (circa 307 BC) said, I know not how to conceive the good, apart from the pleasures of taste, tactile pleasures [sexual pleasures], the pleasures of sound, and the pleasures of beautiful form. The Epicurean view thus expounded on pleasure as deriving from the five senses: pleasure from gustatory and olfactory channels, tact ile channel, auditory channel, and visual channel. One major addition of the Epicurean view of hedonism in addition to the Cyrenaic view is that absence of mental pain (ataraxia) and bodily pain (aponia) are considered pleasures as well, foreshadowing the emerging view that serotonergic activation medicating calmness can generate pleasure. In other words, a phenomenological feeling of calmness, in the absence of hot, more intense phenomenological feelings of pleasure (e.g., gustatory pleasure or sexual pl easure), can also be considered pleasurable in and of itself. For the Cyrenaics and the Epicureans, the goal of consumption, happiness, can be achieved by maximizing pleasure, and avoiding pain. However, Aristotle posed a question that challenged the hedonistic view of the good life. If it were true that pleasure is the only intrinsic good and pain the only intrinsic evil, why is it that nobody would desire to have the mind of a child forever, even if having the mind of a child would mean maximizing plea sures and avoiding pain? Similarly, given the choice to be a beast with the full allowance of pleasures versus a human being with a life of fewer pleasures than the beast, most people would still choose to be a human. Although this was not his explicit stand, Aristotle appeared to view hedonism (e.g., physical hedonism) as just one half of an overall portfolio of happiness, which he called eudaimonia. The Stoics,
18 led by Zeno of Citium, provided a conceptualization that would account for the other half o f happiness (but which the Stoics themselves viewed as the entirety of happiness i.e., physical pleasure is not even needed according to the Stoics). Stoicism and The Philosophical Precursors of Semantic Consumption A competing school of Hellenistic Phi losophy, Stoicism, argued that hedonism falls short in terms of defining what it means for people to have a good life, and what the goal of living, hence consumption, should be. Zeno of Citium (circa 301 BC) founded the school of Stoicism, which proposes that the possession (and exercise) of the virtues is both necessary and sufficient for happiness. The Stoics argue that the original impulse of ensouled creatures is toward what is appropriate for them, or aids in their self preservation, and not toward w hat is pleasurable, as the Epicureans and Cyrenaics contend. The Stoics propose that whatever is good must benefit its possessor under all circumstances. Hence, neither wealth nor health can be considered good, because wealth can be used to consume cannabis, and health can be used to commit atrocities. For Stoics, a person who has attained intellectual and moral perfection is immune to emotions such as fear and envy (a view that foreshadowed the contemporary neural finding of limbic inhibition from the prefrontal cortex), and has attained maximal happiness. Because the Stoics identify the moral virtues with knowledge, hence the perfection of our rational natures, stimuli that is genuinely good (virtue, intellect) is also most appropriate to consumers. So, if the moral and intellectual development goes as it should for consumers, they will progress from valuing food and warmth (as infants), to valuing social relations (as young children), to valuing moral virtue (as adults).
19 Figure 12 Zeno of Citium, ch ampion of the Stoic view of happiness [Reprinted from S. Shakko, 2008, Zeno of Citium with permission under the Creative Commons Attribution Share Alike 3.0 license. ] Stoicism thus lays some of the philosophical foundation for semantic consumption as the main source of pleasure. For Stoics, happiness comes from the practice of virtue and moralit y (which can be seen as the generation of pleasure toward stimulus specific patterns of virtuous or highly moral behavioral configuration), and the learning of knowledge (which can be seen as the building of ever more sophisticated semantic networks). In m odern contemporary society, some of the mental enjoyment we have from watching movies comes from viewing the virtues expressed in the movie. Although the Stoic view of consumption and happiness may at first glance appear extreme, many consumers, both his torical and contemporary, have been
20 observed to consume in a Stoic fashion. Galileo, for example, eschewed physical pleasures and risked the Spanish inquisition so that he could pursue knowledge and build ever more sophisticated semantic networks. In contemporary society, Warren Buffett eschewed physical pleasures, living in a small Omaha bungalow and donating over 95% of his wealth to the Bill and Melinda Gates Foundation, a consumer choice that implies that more pleasure is generated from semantic sources (e.g., virtue and intellect) than from physical sources (e.g., luxury of living in a large house, sensuous food). Charity, movies, sports, and book consumption appear to be some of the prototypical consumption where semantic networks are the main sources of pleasure. Sensory Antecedents of Cyrenaic Hedonism and Semantic Antecedents of Stoicism The goal of this dissertation is to take these philosophical precursors of sensory and semantic consumption and identify some of the sensory and semantic antecedents of consumption. In addition, this dissertation will focus on sensory and semantic antecedents over which the marketer has some control, so that the conceptual predictions have managerial implications. For example, studying the practice of virtue to generate pleasure in the consumer may have limited managerial implications for the marketer, but the communication of virtue in advertising, in a way that generates pleasure for the consumer (and in particular product categories in which virtue leads to pur chase), has broader managerial implications. Table 11 summarizes some of the conceptual questions, and managerial implications, of sensory and semantic antecedents of pleasure.
21 Table 11. Consumer Phenomena/Conceptual Questions and Managerial Implicatio ns. Consumer Phenomena/Conceptual Question Managerial Implication Why do stimuli with negative valence, such as vampires (e.g., Twilight), spiders (e.g., Spider Man), and bats (e.g., Dark Knight) fare better as protagonists compared to stimuli with positive valence? Conceptual Question: How does a negatively valenced stimulus interact with the consumers semantic network to produce positively valenced pleasure? Why do semantic scripts produce pleasure? Why is mere categorization pleasurable? Why does A causes P appear to be more persuasive than P causes A? (a phenomenon I will call Semantic Template Fluency) Why are comfort foods typically sweet and savory? Why are sad consumers more sensitive to touch? Why are certain color pairs more liked ? Why are certain semantic pairs more liked? What are other negative stimuli that movie producers can use to generate pleasure among consumers? (i.e., How does the marketer create a character that will be a financial success?) How can marketers integrate semantic scripts into ads to induce liking? How can marketers organize product lines in a way that is pleasurable to consumers? (e.g., in creating product lines for automobiles, or in creating toy collections) How can marketers choose words in a dvertising copy in a way that maximizes the believability of the claim in the ad? How should marketers optimally design comfort foods? How should marketers optimally design products for sick consumers? How should marketers design their branding materials? How should marketers construct brand associations?
22 CHAPTER 2 TWO CONSUMPTION SYST EMS System 1 The Sensory Consumption System Since the earliest life forms emerged on Earth about 3.8 billion years ago, consumptionthe act of extracting value from s timuli--has been taking place. The earliest documented form of consumption is arguably that of the singlecelled prokaryotes, which swim up a chemical gradient to obtain the nutrients for its metabolic and reproductive needs. From the time of the earliest documented consumption (3.8 billion years ago) until just before mammals emerged on earth (200 million years ago), all consumption has been sensory consumption. Table 21. Timeline of Consumers on Earth. Approximate Time Type of Consumer 4.5 Billion Yea rs Ago 3.8 Billion Years Ago 3.5 Billion Years Ago 3 Billion Years Ago 1.5 Billion Years Ago 1.2 Billion Years Ago 580 Million Years Ago 500 Million Years Ago 485 Million Years Ago 215 Million Years Ago 6.5 Million Years Ago 200,000 Years Ag o No known consumers Single celled Prokaryotes Prokaryotes that use photosynthesis (consumes light) Single celled Eukaryotes Multi celled Eukaryotes Multi celled Eukaryotes develop sexual reproduction Ediacaran biota Animal phyla during the Cambri an explosion First Vertebrates First Mammals First Hominins Modern Humans
23 The complexity of the consumption systems in complex nonmammal animals may be a lot higher than that of the consumption systems in single cell organisms, but the main (and excl usive) channel of the consumption system remains sensory, defined in the sense that the consumer picks up physical sensory qualities (e.g., brightness, sweetness) and uses its sensory channels to extract the necessary chemical compounds needed for metaboli sm and reproduction. The span of 3.6 billion years of sensory consumption, which represents nearly 95% of the timeline of life forms in the universe, is meant to highlight the ubiquity of exclusively senso ry consumption, and meant to foreshadow one of the key proposals in a later section: that contemporary human consumption has become no longer exclusively sensory, and even some human sensory consumption appears to be modified by many nonsensory influences. A Broader View of Sensory Consumption From the c onsumers lay perspective, there is a tendency to view consumption as physical intake of products into the body, such as the eating or drinking of food and beverage. From this lay perspective, it would seem as though the sensory consumption system is limit ed to the gustatory channel. But the other four classical sensory channels (as defined by Aristotle), such as the visual, auditory, tactile, and olfactory channels all have an increasing role in consumers lives, and contribute greatly to the physiological generation of pleasure. Even the littleknown sixth sensory channel, the vestibular channel (known colloquially as the balance system), can be considered a channel of consumption in that it also contributes greatly to the physiological generation of ple asure (e.g., from driving a fast car, from sports consumption, from dancing) Hence, in broader terms, the sensory consumption system should include all sensory channels that directly contribute to the generation of consumer pleasure, and this dissertation will
24 use that broader perspective. This broader perspective leads to a fuller conceptual framework that not only allows consumer researchers to capture more types of consumption phenomena (e.g., the enjoyment of music), it also allows consumers researchers to explore the interactions of different sensory channels (e.g., the interaction of the vestibular and auditory systems in generating pleasure from music consumption). Structure of the s ensory c onsumption s ystem The sensory consumption system consists of multiple stages of visual, auditory, somatosensory, gustatory, olfactory, and vestibular channels that can operate in parallel The use of the term channel is largely a conventionto indicate that channels are subsystems of the overall sensory percept ual system. The o pponent p rocess s tructure of s ensory c hannels A key organizing framework of this dissertation is that sensory channels, as well as semantic systems, appear to be organized in an opponent process structure. An opponent process perspective of consumption implies that each type of sensory or semantic stimulus that is processed has an antagonistic counterpart, and this antagonism has implications for the way the system generates pleasure. Each time the perceptual system activates the primary neural network as a result of perceiving a sensory quality (e.g., red), it has consequences for the antagonistic neural network associated with the bside stimulus (e.g., green). This dissertation will attempt to identify general stimuli and their bsides The next section discusses the opponent process structure of the visual sensory channel, and explores the implications of its opponency process on the systems generation of pleasure.
25 The o pponent p rocess s tructure of the v isual channel Sensory inputs i n the form of light energy enter the perceptual system through the cornea and lens and fall on the retina1. These light particles get transformed into electrical signals and travel to the lateral geniculate nucleus, where these sensory inputs are then modified by sideways operating rules in the mind that are largely not subject to individual differences (e.g., perspective cues). Figure 21 The p rocess of visual p rocessing. [ This faithful reproduction of a lithograph plate from Gray's Anatomy a two dimensional work of art, is not copyrightable in the U.S. as per Bridgeman Art Library v. Corel Corp. ; the same is also true in many other countries, including Germany. Unless stated otherwise, it is from the 20th U.S. edition of Gray's Anatomy of the Human Body originally published in 1918 and therefore lapsed into the public domain.] 1 I restrict the discussion to visual perception, where there is most evidence for opponent processing.
26 The visual channel operates using an opponent process system. Herings (1964) classic formulation of the opponent process theory of visual channel was later validated by Hurvich and Jameson (1967) as well as emerging neur ological studies. There are three opponent channels: redgreen, blueyellow, and black white. A physiological response to one color in an opponent channel is antagonistic to the physiological response to the other color. Figure 22 The visible portion of the electromagnetic spectrum. [ Reprinted with permission from National Aeronautics and Space Administration, 2007. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License Version 1.2. ] A consequence of this opponent process property is that the visual channel is unable to perceive a stimulus as bluish yellow or reddish green. Although yellow can be produced as a mixture of red and green, the brain does not perceive the stimulus as reddish green.
27 A second consequence of this opponent process theory is that when a blue stimulus is placed next to a yellow stimulus, each indivi dual stimulus appears brighter than usual. This will have consequences for the marketer, as will be explored in this dissertation Chapter 5 will make the argument that human memory is able to store hedonic sensitization and habituation. As a consequence of this ability to store hedonic sensitization and habituation, hedonic contrast can also occur across time. Stimuli containing a sensory quality that is antagonistic to a previously perceived stimulus (if it has reached habituation) should generate a greater sensory contrast, and hence, greater pleasure. The gustatory channel Gustatory perception is assumed to play a gatekeeping function by allowing the organism to select the types of food that should enter the digestive system (Breslin 2001). In general taste quality is a reliable indicator of the effect of the substance, because foods that have high caloric value taste sweet, and foods that can cause biological damage to the organism taste bitter. Poisons such as sodium cyanide and strychnine, for example, taste bitter. The next section discusses how the gustatory system transduce s and generates these perceptual experiences Bottom up pathways of gustatory processing The bottom up pathway of gustatory perception begins when the chemicals from the sensory stimuli make molecular contact with the receptors on the tip of the taste cell in the Papilla (see figure below). Because chemical molecules need to fit into the taste pore to enable contact with the taste cell, conditions such as temperature and sol idity of the substance influence the intensity of the taste perception. Transduction occurs when
28 electrical signals generated by the taste cells travel the ascending pathway via the chorda tympani nerve to the brain stem in the nucleus of the solitary tract, where signals travel to the thalamus and to insula and the frontal operculum cortex. There are two schools of thought in explaining the process of taste encoding: (1) Distributed encodi ng, and (2) Specific encoding. The first posits that the locus of taste resides in neural network spike codes, whereas the second posits that the locus of taste resides in specific neurons. Two schools of thought in gustatory encoding Distributed Taste Encoding. Researchers in the distributed encoding school believe t hat the encoding of taste depends on the pattern of firing from different taste receptors. Erickson (1963) first demonstrated that encoding of taste is a function of response patterns, such that different molecular compounds that have different response p atterns taste different, but different molecular compounds that have similar response patterns taste similar. For example, consider the gustatory perception of three different compounds, ammonium chloride (NH4Cl), potassium chloride (KCl), and sodium chlor ide (NaCl). Because these three compounds have different molecular properties, one would expect them to taste different. However, observe in Figure 20 below that ammonium chloride and potassium chloride have similar distributed encoding patterns (red and green patterns), whereas sodium chloride has a different pattern (white pattern). As a result, rats show similar behavioral responses to ammonium chloride and potassium chloride, but different behavioral responses to sodium chloride. R ats who were shoc ked by Robert Erickson (1963) whenever they ingested ammonium chloride avoided potassium chloride in subsequent encounters, but did not avoid sodium
29 chloride. Similarly, rats who were shocked whenever they ingested potassium chloride avoided ammonium chloride in subsequent encounters, but did not avoid sodium chloride. These experimental results suggest that stimuli that induce similar patterns of distributed encoding are perceptually similar, and stimuli that induce different patterns of distributed enc oding are perceptually different, regardless of their actual molecular properties. Specific Taste Encoding. Researchers in the specific encoding school believe that taste perception is a function of specific encoding from the taste receptor (Kim et al. 2003; Mueller et al. 2005). That is, receptors respond specifically to each taste perception. In order to demonstrate this specificity story, Mueller et al. (2005) identified taste receptors that are present in normal humans but not in normal mice. F or example, humans have a taste receptor that responds to the bitter taste in PTC, whereas mice lack this receptor. As a result, normal mice do not avoid PTC, even at very high concentrations in food (see flat response pattern in the figure below ). To te st the specific encoding hypothesis, Mueller et al. (2005) added the PTC receptor in mice through genetic cloning. As a result, these mice, which normally would not avoid PTC, now avoid PTC (see downwardsloping response patter n in Figure 21). This suggests that the encoding processes in specific receptors, by themselves, can contribute to a perception of bitter taste. Reconciling the two schools of thought On one hand, researchers in the Distributed Encoding school believe that taste perception is a function of response patterns across a wide range of receptors (Erickson 1963 ). On the other hand, researchers in the Specific Encoding school believe that
30 taste perception is a function of specific receptors (Kim et al. 2003, Mueller et al. 2005). Both camps are able to provide evidence that support their school of thought, and a direct strong inference test of theory is unfortunately not performed because neither camp used an experimental design that incorporates balanced replication. How do we reconc ile the two schools of thought? One observation is that both camps appear to use stimuli that differ on perceptual complexity Observe that researchers in the distributed encoding camp used experimental stimuli such as Potassium Chloride, substances that are perceptually complex. Potassium Chloride has a complex taste which can be best described as a mixture of saltiness, bitterness, and a metallic taste. When a stimulus is high in complexity, taste perception may be a function of the response pattern of different receptors, because each receptor is responding to a specific stimulus property. In contrast, observe that researchers in the specific encoding camp used experimental stimuli such as phenylthiocarbamide (PTC). Phenylthiocarbamide appears to be low in perceptual complexity, in that its taste can easily be described as simply bitter. Under conditions of low perceptual complexity, receptors should be able to influence taste perception directly, without needing other types of receptors to contr ibute to the perceptual experience. Note, however, that this perceptual complexity interpretation is probably imperfect, or at least incomplete, because the response pattern toward Sodium Chloride (NaCl), which appears to be perceptually simple (salty ), nevertheless contains a complex response pattern, in that there are receptors that respond intermediately to Sodium Chloride. This suggests either that there are taste dimensions that have yet to
31 be identified (i.e., Sodium Chloride may actually be per ceptually complex but we are unable to identify the other taste dimensions besides saltiness), or that certain taste dimensions are explained by distributed encoding (salty dimension), whereas other taste dimensions are explained by specificity encoding (e.g., bitter dimension). To date, this conceptual reconciliation has not been empirically tested in the taste perception literature. In Chapter 6, I will present a few experiments that suggest that affective states can influence the experience of perceptually complex products such as potato chips, wine, ice cream, and bubble gum. A perceptually complex gustatory stimulus implies that not all dimensions of the taste can be perceived at the same time (this goes back to Goldsteins (2007) idea of sensory exp losion, that as much as 10,000,000 bits of information can be available per second), and dopaminergic projections from positive affect can induce the consumer to activate the more subtle (i.e., lower resting potential) perceptual dimensions of the wine, whereas inhibitory GABA and Serotonin projections in an affective state of sadness can induce inhibition of these perceptual dimensions. Predictive property of sensory processing Although processing for each of the sensory channels is independent to some extent, the perceptual system (which one could view as the Gestalt processor that extracts information from the sensory channels) is a Bayesian system that allows concurrent probability properties to trigger activation in an unrelated system. For ex ample, the rubber hand phenomenon illustrates that the somatosensory channel can be activated whenever the co activation of somatosensory and visual information appears to be 100%. The rubber hand phenomenon is a classic demonstration showing Bayesian p roperties of perception. The subject sits on a chair in front of an opaque table (i.e., not
32 a glass table), and for five minutes, his/her right hand (under the opaque table) and a rubberized right hand (on top of the opaque table) will both receive simultaneous tactile stimulation from a confederate. After five minutes of continuous tactile stimulation of the same finger (both on the real and the rubber hand), the confederate will stop the tactile stimulation to both the real and the rubber hands. The conf ederate then takes a needle and pokes the rubber hand. Empirical demonstrations show that the subject will experience real pain from seeing the needle poke the rubber hand, even though the rubber hand is not part of the subjects body. This illustrates the Bayesian properties of the perceptual system it takes statistical information from the environment (such as the 100% correlation between the visual stimulation and tactile stimulation on the rubber hand) and generates a perception based on this statistica l information. It is as if the perceptual system decides that the rubber hand is now part of the subjects body, since the subject is experiencing tactile stimulation (of getting stroked on the real hand) and visual stimulation (of seeing the rubber hand getting stroked) that is tightly correlated. A fter the sensitization period of co activating somatosensory and visual information, somatosensory activation in the system remains even when the veridical object in the physical environment is removed. This leads to a phenomenon in which somatosensory activation can be achieved with visual stimulation alone. The implications for Bayesian properties in neural circuits is that any consistent co occurrence of any pair of sensory (and, I later would argue, semantic) stimuli will be encoded in the neural circuits as a single unit. This encoding as a single unit will have consequences for later retrieval during perception, and will result in the perception of a stimulus that is not objectively present in the physi cal environment.
33 Figure 2 3 The Sensory Perceptual System as a Bayesian Probability calculator. Concurrent activation of somatosensory and visual systems lead to automatic activation of somatosensory channel even when veridical somatosensory input no longer exists. Response properties of the sensory consumption system pleasure pathway In contrast to sensitization and habituation at the level of perception (i.e., whether one perceives a stimulation or not), the processes of sensitization and habitu ation at the level of pleasure (i.e., whether one perceives pleasure or not) is more shrouded in mystery. This is in part because the experimental measurement of perception has been established in the paradigm of psychophysics. In psychophysics, a perceiv ed or did not perceive response (however measured) from a human subject is easy to interpret a perceived response meant that the veridical physical stimulation entered the phenomenological consciousness of the subject, whereas a did not perceive res ponse from the subject meant that the veridical physical stimulation did not enter the phenomenological consciousness of the subject.
34 Studying the response properties of the sensory consumption system at the level of pleasure generation has been much more difficult, in part because the measures of pleasure are often confounded with the measures of perception. For example, self reports of experiences of pleasure arising from consuming an ice cream are often confounded with experiences of sweetness, even though the pleasure arising from sweetness is distinct from the experience of sweetness itself (Rolls 2005) This can result with the subject erroneously reporting some residual experience of pleasure, when it is merely perceptual sweetness that is being ex perienced. As a result, much more statistical noise exists in studies of response properties of the sensory consumption system at the level of pleasure generation. Reporting of experienced pleasure in studies of pleasure can be challenging for most subject s, who do not have expertise in distinguishing inputs of perception (this item tastes sweet) from inputs of pleasure (this item tastes good or I feel pleasure from tasting this item) Recently, however, imaging technology is beginning to allow researchers to more clearly delineate measures of pleasure from measures of perception. fMRI (functional Magnetic Resonance Imaging) technology allows neuroscience researchers to identify which neural structures are generating most activity through the BOLD (Bl ood Oxygen Level Dependent Measures). To the extent that separate brain structures generate perception of the sensory qualities of an object (e.g., sweetness) versus perception of the pleasure that is generated from sensory qualities of the object (e.g., pleasure from the sweetness), BOLD measures can help researchers identify response properties of the sensory consumption system at the level of pleasure generation.
35 A finding that has been gaining acceptance in the neuroscience literature is that repeated sensory stimulation can lead to the neural habituation of certain circuits in the pleasure system. Although there remains some disagreement within the field, most neurobiologists recognize that the orbitofrontal cortex is a vital part of the pleasure syste m, i.e., the system being the series of circuits involved in the generation of pleasure. A response property of the sensory consumption system is the propensity of the pleasure system to habituate after repeated stimulation of a specific sensory stimulus. Neuroscientists have found that habituation of certain neural circuits in the orbitofrontal cortex correspond with behavioral indicators of habituation of pleasure (e.g., the cessation of observable liking responses). Two types of sensitization and habituation implications for contrast and pleasure generation The two types of sensitization and habituation (perceptual habituation and pleasure habituation) have implications for consumer research. At the level of memory trace formation, sensory sensit ization appears to be the foundational antecedent of processing fluency processes processes whereby memory traces in the network accumulate activation level toward traces of a stimulus as repeated exposure is encountered. As the perceptual system encounter s the same stimulus repeatedly, it releases neurotransmitters (primarily glutamatergic) that increases the level of neural activation in the network and consolidates the memory trace. When the stimulus is again encountered in a future event, this highly c onsolidated memory trace is activated, and a feeling of familiarity (probably mediated by serotonergic and oxytocynergic reactions) is experienced by the perceiver, which leads to liking. Hence, repeated
36 exposure to a product design, for example, leads to consumer liking because the highly consolidated memory trace of a product appearance leads to a feeling of familiarity, which is attributed to liking. Much less attention has been paid to the process of sensory and pleasure habituation, especially in term s of how these two response properties result from pathways in interactions between sensory systems and pleasure systems. It has recently been found that repeated stimulation of a sensory quality (e.g., the sweet taste of ice cream among humans; and in mor e controlled settings, the sweet taste of sucrose among rats ) leads to habituation. This leads the consumer to reject the food product after consuming a sufficient portion. What has not been investigated, however, is how visual stimulation from consumption influences habituation, and whether this habituation is at the level of sensory, pleasure, or both. Recall that there are numerous pathways between sensory channels in the sensory consumption system, and that visual stimulation often predicts stimulation from other sensory channels. Visual quality in naturally occurring fruits and vegetables is correlated with gustatory quality (and hence correlated with intrinsic nutrients). It is not coincidental, therefore, that green stimuli taste similar to other gree n stimuli, and that green stimuli provide similar macro and micronutrients. Purple stimuli taste similar to other purple stimuli, and also provide similar macro and micronutrients (blueberries, raspberries, and pomegranates) Since the perceptual system is a Bayesian system with the goal of providing probabilistic prediction of meeting physiological needs in a balanced way, repeated visual stimulation of the same visual sensory quality (e.g., red in the example of a strawberry ice cream) may lead to habituation.
37 Sensory qualities are not perceptually organized by the neural system in a random way. As the Hering (1964) and YoungHelmholtz theories of color perception demonstrated, the sensory perceptual processes that lead to the perception of red stimuli is in a direct opponency process to the sensory processes that lead to the perception of green stimuli. Similarly, the sensory perceptual processes that lead to the perception of blue stimuli is in a direct opponency process to the sensory processes t hat lead to the perception of yellow stimuli. That is to say, the perceptual system can perceive a mixture of red and another color at the same time, but not a mixture of red and green. Similarly, the perceptual system can perceive a mixture of blue and another color at the same time, but not a mixture of blue and yellow. Adding green qualities to a red stimulus necessarily reduces the redness of the stimulus (while still remaining perceptually non redgreen), and adding yellow qualities to a blue stimulu s necessarily reduces the blueness of the stimulus (while still remaining perceptually non blueyellow). Thus, the maximum amount of perceptual contrast can be achieved by juxtaposing red against green stimuli, and the maximum amount of perceptual contrast can be achieved by juxtaposing blue against yellow stimuli. There are potentially other color pairs that generate significant contrast as well. Although the reasons for this type of opponent process perceptual organization are not clear, it appears that color perception is tied to the prediction of nutrients in food intake. Red foods appear to contain nutrients that are most (but not completely) orthogonal to those in green foods, and blue (strictly closer to purple) foods appear to contain nutrients that are most orthogonal to those in yellow foods. To the extent that consuming stimuli with a sensory quality that is in direct contrast with a previously
38 consumed stimuli enhances the overall portfolio of nutrients consumed, stimuli with a directly contrast ing sensory quality should be most ideal in slowing down habituation toward the consumption of a product, even in an ostensibly irrelevant sensory dimension (visual versus gustatory). These propositions can be summarized as follows: P1. Marketing stimu li (e.g., packaging, advertising) that is visually consistent with the product (e.g., candy) leads to a faster rate of habituation at the level of pleasure generation. P2. The rate of habituation to pleasure from sensory consumption can be slowed down by using marketing stimuli (e.g., packaging, advertising) that contains a sensory quality that is in direct contrast with the sensory quality of the product. System 2The Semantic Consumption System Although sensory consumption has dominated organismic cons umption in the first 3.8 billion years of history in which life has emerged on Earth, a different kind of consumption, semantic consumption, has developed in at least the last 2.5 million years, when hominids first appeared. In contrast to earlier life for ms, whose systems extracted stimuli from the environment based on physical sensory properties, hominids started to deve lop neural circuits that extracted value from semantic and symbolic information. These neural circuits allowed these early semantic consu mers to generate secondorder symbolic representations from first order perceptual representations. Using the same definition as that in sensory consumption, semantic consumption is defined as the process of semantic generation (some pattern of activation in semantic networks) that leads to the generation of pleasure (opioidergic activation in hedonic hotspots, e.g., Pecina et al. 2006) Semantic consumption does not have to be a
39 process that is exclusively semantic a notion that is impossible, because the generation of meaning necessarily requires an initial sensory processing. Instead, a particular consumption episode can be classified as predominantly semantic if the largest variance of amount of pleasure experienced by the consumer depends largely on ch anges in semantic content of the product or service (as opposed to changes in the sensory content of the product or service). One of the main types of pleasure generated in the semantic consumption system appears to be information. Since ancient Greek civ ilization, people have had the notion that information, in and of itself, is rewarding, and can be pursued without any other end, in other words, no other contingent reward or punishment (such as profit). Aristotle observed that men study science for intri nsic reasons and not for any utilitarian end (Posnock 1991). Today, Physics PhDs eschew higher paying salaries in the industry to pursue knowledge for more modest salaries, implying that knowledge can be an intrinsic reward, an unconditioned stimulus that compensates for opportunity cost. Harlow, Harlow, and Meyer (1950) demonstrated that monkeys would attempt to solve a puzzle (which can be assumed to be represented in a simple nonlanguage based semantic network), even when there was no external incentive for solving the puzzle. In fact, Harlow et al. showed that monkeys performed best in solving a puzzle when there was no external incentive. Human subjects who have not had the opportunity to read or listen to any semantic material for a long time will request to listen to otherwise uninteresting semantic material such as old stock market reports from decades earlier (Hebb 1958). These findings suggest that the mere activation of semantic network (in
40 certain types of configurationprecisely what this resear ch program seeks to identify) can generate pleasure. Structure of the Semantic Consumption System The main structures that take in semantic stimuli, and generate pleasure, are (1) Wernickes area and the Inferior Parietal Lobule, (2) Hippocampus, (3) Medi al Temporal Cortex, (4) Dorsolateral Prefrontal Cortex, and (5) Orbitofrontal Cortex (as well as Nucleus Accumbens/Ventral Pallidum). The Inferior Parietal Lobule is a multi modal brain area that receives visual input such as words and sentences (through the sensory thalamus), and auditory input such as speech sounds. As a mnemonic simplification, one can think of the Wernickes area and the Inferior Parietal Lobule as Language Areas (or, more specifically, Language Comprehension Areas, to distinguish t hem from the Language Production Areas). The Inferior Parietal Lobule, and Wernickes Area, receiv e visual symbols, and auditory speech, and then compares the inputs with the spreading activation network in the hippocampus and medial temporal cortex. This process helps the person generate meaning in response to those visual and auditory symbols. Most accounts of semantic processing stop here, that is, they stop at the point in which meaning is generated by the processing system. However, it is clear that t he semantic processing system is not a sterile, cold system that merely computes a meaning in response to incoming symbolic stimuli. Observing the lives of consumers, and indeed human beings in general, it is clear that pleasure and aversion are important outputs of the semantic processing system.
41 Figure 2 4 Pathways from the semantic system toward the pleasure system. Meaning alone, without any pleasurable sensory stimulation, can generate pleasure among consumers. One of the main objectives of thi s research program is to identify the various ways in which meaning can generate pleasure, and to break meaning down to its fundamental parts. When meaning is generated as a result of the match between incoming symbolic stimuli (e.g., words, sentences) and the medial temporal cortex, this meaning generation process appears to activate groups of neurons in the orbitofrontal cortex, which in turn activates neurons in the nucleus accumbens and ventral pallidum, leading to a phenomenological feeling of pleasu re.
42 A recent fMRI study implied the generation of pleasure toward semantic information was reported by Kang et al. (2008). Kang et al. showed that anticipatory circuits were activated in response to a question (almost the same set of anticipatory circuits that would be activated in response to the smell of food, as in Rolls 2005), and pleasuregenerating circuits were activated in response to the answer. Kang et al. (2008) stated that neurons in left PFC receive input from neurons in the substantia nigra via the dorsal striatum, which respond to primary rewards and rewardprediction. What are some implications from these response patterns? This implies that the curiosity corresponds to reward anticipation, whereas learning perhaps corresponds to reward c onsumption. For example, reading the question What instrument was invented to sound like a human singing? appears to activate anticipatory areas of the brain (this corresponds to expectation of consumption) whereas reading the answer Violin appears to activate the pleasuregenerating areas of the brain (this corresponds to the consumption itself) The two tables on the next page show a number of prototypical sensory pleasures (food, scent, sexual intromission) and their neural correlates, as well as a number of prototypical semantic pleasures (movies, books, paintings) and their neural correlates. The main insight here is that, in all of the sensory pleasures (also called fundamental pleasures in Rolls 2005) the pre frontal cortex and semantic areas are not needed to generate a feeling of pleasure. In contrast, in all of the semantic pleasures, the feeling of pleasure cannot be generated without activation of the prefrontal cortex and semantic areas. This response pattern suggests that the locus of se nsory and semantic consumption is quite different.
43 Table 22 Fundamental Pleasures that appear to have sensory antecedents. Fundamental Pleasures Requires activation of pre frontal cortex and semantic areas such as Geschwind module and inferior pariet al lobule? Gustatory stimulation (food) Olfactory pleasures (scent by itself, aroma in food) Tactile pleasures (sexual intromission) No (Kringelbach and Berridge 2010) No (Rolls 2005) No (Kringelbach and Berridge 2010; Rolls 2005) Table 23 High er order Pleasures that appear to have semantic antecedents (in addition to sensory antecedents). Higher order Pleasures Requires activation of prefrontal cortex and semantic areas such as Geschwind module or inferior parietal lobule? Art Transcendent pleasures Books Movies Collecting (e.g., stamp collecting) Altruism* Transcendent pleasures* *Altruism and transcendent pleasures can be contained within movies and books, of course. Yes (Skov 2010) Yes (Kringelbach and Berridge 2010) Yes (Krin gelbach and Berridge 2010) Yes (Andrade and Cohen 2007**) Yes (Panksepp 1998) Yes (Kringelbach and Berridge 2010) Yes (Kringelbach and Berridge 2010) **Assumes that verbal instruction regarding framing activated prefrontal cortex and semantic areas. Key Features of Semantic Consumption: Fecundity and Purity The semantic consumption system has some surprising advantages over the sensory consumption system. The philosopher D. H. Monro (1967), in reference to the English philosopher Jeremy Bentham (17481832), proposed that The higher pleasures
44 are those that afford a more lasting satisfaction than the lower pleasures, that enlarge our horizons and so open up new possibilities of pleasure (fecundity), and that are less likely to be followed by the pain of satiety and boredom (purity).These are the characteristics that distinguish intellectual activity from purely physical pleasure. Fecundity in Semantic Consumption A consumers acquisition of semantic consumption skills gives rise to fecundity (repr oductivity), which opens up many new possibilities for new semantic pleasures. In neural terms, fecundity refers to the ability of newly constructed semantic networks to produce yet newer semantic networks that can generate new sources of pleasure (new sem antic to pleasure pathways). People who consume semantic product A today are thus able to feel pleasure toward semantic product B tomorrow, an outcome that would otherwise not have been possible if the first consumption of semantic product A did not occur. Consider the semantic networks that are constructed as a result of consuming physics. Physics consumption generates new ways of generating semantic pleasure, such as Engineering. Engineering phenomena become pleasurable as a result of the semantic networks constructed from the earlier Physics consumption. In the same way, the semantic networks that result from the consumption of mathematics give rise to new semantic pleasures, such as Finance and Accounting. Finance and Accounting Phenomena, phenomena that were once unable to generate pleasurable for the consumer, now generates pleasure because of the semantic networks constructed as a result of first consuming mathematics. Fecundity is not just generated in the consumption of science, which may be c onsidered somewhat atypical consumption.
45 The next section shows some prototypical consumption domains that illustrate the property of fecundity in semantic consumption: The consumption of sports, movies, books, and advertising and other marketing material. Sports Consumption. Sports consumption qualifies as semantic consumption in the sense that the extraction of sensory qualities does not fully account for the pleasure that is generated. Sports consumption also illustrates the fecundity of semantic consu mption. The learning of one sport gives rise to new pleasures in response to the consumption of other sports. For example, the semantic network constructed as a result of consuming Tennis enables the generation of pleasure as a result of consuming Badminton. In other words, the construction of the Tennis semantic network gave rise to new possibilities of pleasure not otherwise available to the consumer. Movie Consumption. Movie consumption qualifies as semantic consumption in the sense that extraction of sensory qualities does not fully account for the pleasure that is generated. Movie consumption illustrates the fecundity of semantic consumption. The semantic network constructed as a result of consuming one movie (e.g., Trading Places) enables the generat ion of pleasure in response to consuming another movie (e.g., Erin Brokovich), in so far as the first consumption constructs new semantic structures, such as a the semantic script Underdog good overcomes powerful evil (Trading Places) that enables the se mantic consumption system to produce opioids in response to consuming the second movie (Erin Brokovich). These semantic scripts are nonconscious, but are recognized by the system as a result of repeated Bayesian occurrence. Purity in Semantic Consumption
46 What the philosophers call purity in intellectual consumption appear to correspond quite closely to conceptions of neural habituation of key areas in the pleasure circuit, such as the orbitofrontal cortex (e.g., Rolls 2005). Jeremy Bentham (Monro 1967) noted that intellectual pleasures have higher purity, that is, they are less prone to the pain of satiation than physical pleasures. Rolls (2005) has demonstrated that neurons in the orbitofrontal cortex habituate to a physical reward (e.g., sucrose) af ter the organism has had the opportunity to consume sucrose, but no such habituation has yet been demonstrated in terms of semantic consumption (although conditions can be set up to produce it, as will be discussed later). This appears to be attributable t o the structure of the semantic consumption system. Watching a second movie after having watched a first movie generates pleasure for the consumer, because the semantic networks that are activated from watching the second movie are different from the semantic networks that are activated from watching the first movie (as long as the movies are different). Eating a second gallon of ice cream after having eating the first gallon of ice cream, however, will not generate pleasure for the consumer (even if the ic e cream flavors are different), because the sensory networks responding to sweetness that are activated from eating the second gallon of ice cream are the same as the sensory networks that are activated from eating the first gallon of ice cream (i.e., sens ory specific satiationthe pleasure system stops firing toward gustatory system, but only in response to specific sensory categories, such as sweetness. Hence, if the organism consumes a different sensory category, the pleasure system resumes firing). In addition, the sensory consumption system appears to have built in homeostatic controls for preventing overeating. Panksepp (1998), for example, demonstrates that
47 the orbitofrontal cortex firing in the rat (responsible for generating a feeling of pleasure) can be vastly decreased by directly injecting sucrose into the digestive system of the rat. No such neural habituation, in contrast, appears to occur in the semantic consumption system. Neurons in the orbitofrontal cortex do not appear to habituate based on the semantic fullness of the system. If true, this suggests that semantic consumption has some advantages over sensory consumption in terms of less habituation. Conditions for Habituation toward Semantic Consumption Although semantic consumption of different products (e.g., different movies, different football games) appear to be fecund and gives rise to the possibility of other pleasures, there appear to be consumption phenomena which is consistent with neural habituation of the pleasuregenerating circu its following semantic consumption. For example, consider the following prototypical consumer comments on more semantically oriented products : Table 24 Consumer feedback on semantic consumption, and implied neural response. Consumer Feedback on Semantic Consumption Implied neural response I used to like it, but I am now sick of watching CSI (a particular TV show). I used to like him, but I am now tired of watching Robin Williams. I used to enjoy her songs, but I no longer want to listen to Lady Gaga s songs. Habituation of pleasu re circuits to semantic network activation Habituation of pleasure circuits to semantic network activation Habituation of pleasure circuits to semantic network activation Although these consumer comments all sound very co mmonplace, the processes that lead to the loss of pleasure from consuming products, and their antecedents, are not well understood. At the neural level, there is likely a decrease in the production of neurotransmitters that help generate pleasure or desire ( it is still unclear whether the
48 pleasure that is generated by these types of semantic consumption is really pleasure in the sense that neuroscientists talk about pleasure (opioidergic processes), or whether this pleasure is actually desire (dopaminergic processes) but is given the wrong label. Once again, like Ralph Hulls tuned deck, pleasure from semantic consumption may actually be two kinds of pleasure one that is truly generated as pleasure in the opioidergic sense, and another that is generated as desire in the dopaminergic sense. It may be the case that different semantic antecedents can lead to different kinds of pleasure in the process of consuming semantic products. One of the goals of the research to be generated from this dissertation is to identify the semantic antecedents to semantic habituation, and explore how the marketer of a semantic product can advertise or design a product in such a way as to prevent habituation.
49 CHAPTER 3 THE TWO AFFECTIVE SYSTEMS The consumer has two broad processing systems sensory and semantic systems, but mere perception is not sufficient in consumption. In order for the consumer to assign a common currency for valuing different stimuli, and to determine when the consumer should assign relatively high (or low currency) to these stimuli, the consumer has to rely on endogenous inputs from its two affective systems. The organism has two affective systems the hedonic affective system (sometimes loosely called the pleasure system), and the state affective syst em (sometimes loosely called the wanting system) (Barbano and Cador 2007; Berridge and Robinson 2003). One way to conceptually organize the two systems is that the first affective system (hedonic) evaluates the value of external stimuli, and the second aff ective system (state) evaluates internal stimuli. One of the main findings in this dissertation (e.g., the chapters Affect Gating and Affect Tuning) is that the output of the state affective system changes the output of the hedonic affective system, so that different external stimuli are valued under different internal states. The first affective system, the hedonic affective system, enables the consumer to hedonically discriminate among various external stimuli (Rolls 2005). This hedonic discriminati on is a heuristic for determining which stimuli are needed most by the consumer. The second affective system, the state affective system, enables the consumer to identify its own organismic status (Panksepp 1998). Organismic status signals the organisms l evel of fitness at that particular moment, and hence the rewardto risk appetite it should have in pursuing different external stimuli, which entail different levels of reward and risk.
50 Hedonic Affective System The hedonic affective system, with key circu its located in the orbitofrontal cortex (Rolls 2005) and the nucleus accumbens (Berridge and Robinson 1998), evolved in mammals to enable the mammal to determine which external stimuli it should approach or avoid. Because the hedonic affective system recei ves projections from the sensory channels, the amount of pleasure generated toward a specific stimulus depend on the stimulus fitness consequences. The more positive the stimulus fitness consequence is, the higher the pleasure response of the system (e.g., the more opioids are released). For example, more pleasure is generated toward highly concentrated sources of sugar, compared to moderately concentrated sources of sugar (Panksepp 1998). The more negative the stimulus fitness consequence is, the higher the aversion response of the system (e.g., the more acetylcholine is released; Hobel, Avena and Rada 2007). For example, the level of acetylcholine released is a function of the intensity of the bitterness. The levels of pleasure neurotransmitters releas ed, as well as the levels of aversion neurotransmitters released, can be approximated by behavioral measures. For example, the frequency and intensity of tongue licks by rats vary as a function of the pleasure they experience toward varying sweetness from different concentrations of sucrose, and the intensity of gapes by rats vary as a function of the aversion they experience toward varying bitterness from different concentrations of quinine. Natural selection of response properties in neural systems selectively retains hedonic systems that produce pleasure toward fitness giving stimuli, and pushes out hedonic systems in the ecology that do not produce pleasure toward fitness giving stimuli. Similarly, natural selection of response properties in neural sys tems selectively retains hedonic systems that produce aversion toward fitness reducing stimuli, and
51 pushes out hedonic systems in the ecology that do not produce aversion toward fitness reducing stimuli. State Affective System The consumers state affecti ve system, with key circuits in the brainstem, performs a function that is different from the hedonic affective system (Barbano and Cador 2007). Whereas the hedonic affective system evaluates external stimuli based on sensory (and, this dissertation argues, semantic) qualities, the state affective system evaluates internal stimuli, and assesses the organismic status of the consumer. When the consumer is in an organismically deficient state, such as when suffering physical injury, or when sick, the release of CRF (Corticotrophin Releasing Factor) slows down firings in the sensorimotor cortex, minimizing the risks taken by the consumer. The risk to reward appetite of the consumer under an organismically deficient state is low because risky stimuli are more l ikely to kill the organism than rewarding stimuli are likely to benefit the organism (i.e., the impact of the bad is higher than the impact of the good). Under organismically deficient states, the consumer will feel sluggish, and this discourages the consu mer from pursuing stimuli (or pursuing types of stimuli) that can be risky to an already vulnerable system. In contrast, when the consumer is in an organismically sufficient state, the release of dopamine accelerates the sensorimotor cortex, maximizing the pursuit of fundamental rewards (sources of food, water, and reproductive opportunities) even at the cost of greater risks. Panksepp (1998) reports that activating the organisms rewardseeking system increases forward locomotion in almost all mammals tested in the laboratory. Although not exactly equivalent, Sacks et al. (1972) suggest that
52 dopaminergic activation in humans, through the administration of Ldopa in dopaminedeficient patients, increases approach locomotion, increasing the frequency and sp eed of walking toward rewards such as food and the opposite sex. The risk to reward appetite is higher because (as a result of natural selection) the benefits of pursuing additional rewards during an organismically sufficient state outweigh the risks invol ved in pursuing these rewards (i.e., the impact of the good is higher than the impact of the bad). In the end, the organism is (nonconsciously) increasing its probability of passing down its genetic codes, and pursuit of reproductive opportunities (for ex ample) when in an organismically sufficient state increases its chances of passing down its codes, even when this pursuit is risky. The Heuristic Value of Sensory Qualities in the TwoAffect Organism The fitness giving, or fitness reducing, qualities of stimuli cannot be directly computed by the perceptual system. The natural selection of response properties takes time, and so the only physical qualities that neural systems are able to perceive are those that are evolutionarily persistent (i.e., physical qualities that have remained persistent over a long period of time, at least thousands of years). Because perceptual systems have to predict the fitness giving (or fitness reducing) status of a stimulus based on physical qualities that are evolutionarily persistent, the sensory system is able to hedonically adjust toward only broad sensory quality categories of stimuli. In gustatory stimuli, the classical sensory qualities of sweetness, sourness, saltiness, bitterness, savoriness, and spiciness are physical ly extractible by the system (Goldstein 2007; Rolls 2005). There are, however, problems with regards to a hedonic system that relies only on sensory qualities to predict its fitness enhancing or fitness reducing status. Because
53 each sensory quality is broad, each sensory quality can be predictive of both fitness enhancement and fitness reduction. For example, sourness can predict fitness reduction in the case of rotting food (the perceptual system generates sourness toward the presence of acids, such as lactic acid), or fitness enhancement in the case of citrus food (the perceptual system generates sourness toward ascorbic acid and citric acid). One way for the system to resolve this issue is to use the organisms state affect system to adjust its riskto reward ratio upward or downward, depending on whether the organisms state affect signals organismic vulnerability or organismic sufficiency. When the state affective system signals organismic vulnerability, it will modify the response of the hedonic affec tive system toward sour stimuli in order to produce aversion, and limit the organisms risk taking when in a vulnerable state (e.g., when physically injured, sick or separated from parent). This may be consistent with the comfort foods phenomenon, in whi ch foods consumed by people under negative affective states are rarely sour or spicy. In the universe of foods, comfort foods tend to be, on average, less sour and less spicy than the average food. In contrast, when the state affective system signals organismic sufficiency, it will modify the response of the hedonic affective system toward sour stimuli in order to produce pleasure (i.e., the sour sensation produces a feeling of pleasure in the organism) and increase the organisms risk taking when in a s ufficient state. This modification takes the form of heightened sensitization (increased perceptual intensity of sourness) and sensory interpretation (experience of the sensory quality as punishing as opposed to rewarding; e.g., Rolls 2005). These proposit ions can be summarized as follows:
54 P 3a. Consumers under negative affective states will experience aversive tastes as more intense, and more punishing, than consumers under neutral affective states (comfort foods phenomenon). P 3b. The increased intensity of sourness, and the interpretation of the aversive taste as punishing, will lead to consumers lower willingness to pay for sour foods. These propositions will be tested in the chapter on Affect Tuning.
55 CHAPTER 4 OPERATING PROPERTIES OF THE TWO SYSTEM CONSUMER Four Pathways Given the two consumption systems (Sensory and Semantic) and two affect systems (State and Hedonic), I will identify four pathways through which we can study various consumer phenomena. Each of these pathways contain processes that may eventually account for multiple consumer phenomena (sample phenomena in parentheses): (1) Sensory to Hedonic Affect Pathway (e.g., sensory contrast phenomenon), (2) Semantic to Hedonic Affect Pathway (e.g., semantic habituation phenomenon in branding), (3) State Affect to Hedonic Affect Pathway as mediated by Sensory Channels (e.g., comfort foods phenomenon; cold hands phenomenon), and (4) State Affect to Hedonic Affect Pathway as mediated by Semantic Channels (e.g., the tortured artist phenomenon). A discussion of each of these pathways will lead to a number of general and specific propositions, some of which would then be tested in laboratory experiments reported in chapters 5 and 6. First Pathway: Sensory to Hedonic Affect Pathway One of the most fundamental functions of sensory channels in the perceptual system is to provide perceptual discrimination. For example, visual discrimination of edible stimuli, such as purple blueberries from a background of green foliage, is widely considered to be the fundamental function of color perception in the visual channel ( Goldstein 2007). Without color perception, it would be extremely difficult for organisms to select the berries from the rest of the foliage, greatly reducing the probability of survival and passing down genetic code.
56 A key insight, however, in the study of color perception in the visual channel, is that the generation of color perception is not random; in other words, not purely for nonsystematic discrimination as is assumed in the visual perception literature (i.e., it is not a random label) For example, green stimuli (e.g., leafy vegetables) contain very different phytochemicals compared to red stimuli (e.g., tomatoes, strawberries). As a result, it is a convenient perceptual heuristic for organisms to consume stimuli of different colors to obtain all the phytochemicals and micronutrients that are needed for metabolism and growth (i.e., consuming food of different colors is more likely to provide adequate nutrition that consuming food of the same color) As discussed in the previous chapter regarding the two affective systems it is a universal property of hedonic systems to generate pleasure toward stimuli that are generally beneficial to the organism. The natural selection of protein codes in DNA that produce a physiological system capable of producing pleasure toward beneficial stimuli occurs in all ecological systems in which there is competition for finite resources (Darwin 1879). This leads to the following proposition: Hedonic habituation should occur toward continuous perception of the same sensory quality of a stimulus (e.g., red in a strawberry). It appears that the consumption does not even have to be gustatory, i.e., the stimulus does not have to enter the organisms digestive sys tem, as previously thought, before habituation could occur. These propositions can be summarized as the following: P 4 Repeated visual exposure to a sensory quality (e.g., color, texture) should lead to hedonic habituation and cause liking to decline for t hat sensory quality.
57 Perhaps more interestingly, the opponent process structure of the visual channel, along with the natural selection of hedonic properties (the tendency of systems to generate pleasure toward stimuli that are incrementally beneficial), lead to a proposition regarding what stimuli would be more preferred after hedonic habituation to a sensory stimulus occurs: P 5 Repeated visual exposure to a sensory quality (e.g., blue) should lead to a hedonic buildup for the sensory quality that is antagonistic to it (e.g., yellow). The aforementioned propositions have marketing implications. A marketer, such as Starburst (whose product and its packaging is yellow), can increase hedonic response toward its product by exposing consumers to its opponent sensory quality in the ad (by showing a bluedominant ad before presenting the product). Selecting antagonistic pairs of sensory qualities would enable the greatest release of hedonic pleasure at the point when the consumer perceives the product (usually towards the end of an advertising copy). These predictions may even extend toward products that are non gustatory. A marketer, such as Apple (formerly known as Apple Computer), can increase hedonic pleasure of consumers toward its product designs by increasing consumer exposure to an opposing sensory quality (e.g., blue, jagged, matte) just before consumers are exposed to its product with certain sensory qualities (e.g., yellow, smooth, glossy). This presentation sequence can be implemented either in adver tising or in the retail environment (i.e., its Apple store). Following this presentation sequence will ensure the greatest release of pleasure response among consumers, compared to a presentation
58 strategy that is random (or consistent with the product col or). This proposition can be summarized as follows: P 6 Presenting sensory qualities (e.g., blue, jagged, matte) that are antagonistic to a products sensory qualities, before presenting the product itself, will enhance visual pleasure toward the product, and enhance willingness to pay for the product. The following table summarizes the main opponent categories of sensory channel extraction, and consequently, pleasure generation: Table 41. Opponent sensory categories Sensory Category Sensory Percept Anta gonistic Percept Visible light (color) Visible light (color) Gustatory sensation (taste) Gustatory sensation (taste) Blue Red Savoriness (Umami in Rolls 2005) Yellow Green Pungency Second Pathway: Response Properties of the Semantic Consumption S ystem Semantic Generation Pathway Sensitization and habituation are two processes that have been documented for percept ion. A consequence of cells in the body, especially neurons (brain cells), is that repeated activation leads to neural habituation. As sociative links connecting memory traces carrying meaning (family, tradition) and memory traces carrying brands (Coca Cola) are activated by neurons that may also habituate. Marketers routinely spend millions of dollars in efforts to increase consumer expo sure toward their brands. A paradoxical consequence of increasing consumer exposure to brands is that brands may lose their meaningfulness, even if the consumer can still perceive or recall the brand (the memory trace containing the brand itself). To the extent that these brand associations carry positive valence that is then imputed into
59 an overall evaluation of brand value or brand equity, increasing exposures to a brand may actually hurt brand equity, at least under certain conditions. Brand knowledge is represented in an extensive memory network that contains memory traces in multiple dimensions, spanning category needs, descriptive features, personal value and meaning, imagery, and cognitive and affective responses (Aaker 1997; Keller 2003; Peter and Olson 2001). These memory traces become part of the brands semantic network because a consumers temporally simultaneous exposure to the brand and an entity (such as a person, place, or event) can establish associative links between the brand and the enti ty (Mitchell 1982). As a result, subsequent exposure to a brand can unconsciously and automatically activate these semantic associates. To the extent that these semantic associates are positive in valence, their automatic activation enhances the overall sy nthesized brand meaning and increases the consumers desire to consume the brand. Brand networks operate much like associative networks, such as the classic spreading activation model by Collins and Loftus (1975). The cognitive and neurological processes that underpin the operational properties of brand networks are thus similar to semantic networks for words. Each semantic association (e.g., firetruck) is connected to the word (e.g., red) via an associative link, and the activation strength of this link is a function of the number of co activational exposures to the semantic association and the brand. The more exposures involve the co activation of the word and its semantic association (e.g., the more red firetrucks are encountered), the stronger the acti vation strength of the associative link becomes. A consequence of a stronger associative link is that activation of the word is more likely to activate its semantic associate.
60 The neurobiology literature has demonstrated that neurons can habituate following repeated activation (Groves and Thompson 1970; Kandel et al. 2000; LeDoux 2002; Thompson and Spencer 1966) For example, neurons in the mammalian central nervous system first sensitizes toward a stimulus, showing increased activation after a few exposur es. But after repeated exposure to the stimulus, neurons begin to habituate, as its firing rate decreases until they completely cease to fire. Neural habituation appears to be a universal response property in the mammalian physiology, applying to sensory phenomena involving touch, gustatory, auditory, and visual systems. The conscious perception of a shirt, for example, is most obvious at the time that the wearer first wears it, and gradually becomes unconscious as neurons in the somatosensory cortex stop f iring in response to the tactile stimulus. Most demonstrations of sensory habituation have been performed in the animal literature, particularly in animals that have relatively few neurons, enabling researchers to isolate each cell as measure its response But among humans, sensory stimulation is not the only dimension of consumer experience. Semantic stimulation is a major aspect of consumer experience, particularly in adults with well developed neural circuits for semantic networks. As human beings grow, neural circuits that support semantic networks develop in complexity, enabling humans to predict and anticipate consequences from the activation of semantic associations in semantic networks. Semantic associations allow consumers to quickly evaluate different products in the environment without requiring effortful thinking. These semantic associations can be linked to traces that are either positively or negatively tagged (Bower 1981), allowing the consumer to make a quick evaluative judgment. Brands are one such heuristic that
61 allow consumers to make a quick evaluative judgment about a product or service. Because brands activate associative links containing semantic associations that are probabilistically predictive, brands help consumers choose among opti ons with minimal deliberative effort. Although habituation of sensory stimuli is fairly well established, habituation of semantic stimuli has not been demonstrated. Because brand associations are mediated by neuronal firings, this dissertation proposes t he same pattern of results for semantic habituation, in which the first few exposures sensitizes the associative links, facilitating the activation of brand associations. After repeated exposures, however, the neurons that mediate the activation of the brand association may start to habituate, causing a failure to activate some of its brand associations. As a result, a brand may become less meaningful and brand equity can, paradoxically, decline following repeated exposure. Because the brand network involves associative links to both the semantic associations (Collins and Loftus 1975) and to affective responses (Bower 1981; Keller 2003), this dissertation proposes that semantic habituation can occur at two levels. The first level occurs at the level of associative links to semantic associations. The associative link between the brand and its semantic associations can habituate, causing the brand to lose meaning. The second level at which semantic habituation can occur is at the linkage between brand as sociations and affective response, in which the affective nodes that are attached to the semantic associations attached to the brand can habituate. The associative link between the brands semantic associations and their corresponding affective responses c an habituate, causing the consumer to lose the experience of pleasure toward an activated meaning. Both levels of habituation have
62 important consequences for the brand, but they have different antecedents that marketers can control. Because the first lev el of habituation, at the level of associative link between the brand and its association, has a direct neuronal contact point with the brand, a consequence is that repeated exposure to the brand per se can cause habituation. In other words, merely exposing the consumer to the brand can cause habituation of its meaning associations, as long as the number of exposures is enough to cause habituation. However, because there is a phonological component to brand representations (Dell 1986), alternative brand names with distinct phonological patterns (e.g., Coca Cola and Coke, Reebok and Rbk) activate slightly different nodes through distinct associative links, and can thus slow down the process of habituation following repeated exposure. These predictions lead to the following propositions: P7. Repeated consumer exposure to the brand can decrease the automatic activation of brand associations, leading to a loss of brand meaning. P8. To the extent that brand meaning underlies consumer desire of the brand, loss o f brand meaning can decrease consumer desire and purchase intentions. P9. The use of alternative brand names can slow habituation. The second level of habituation occurs at the link between the brands semantic associations and their corresponding affect ive responses. A phenomenon in social psychology that may be consistent with the idea of contrasting semantic categories is the propensity of female humans to have heightened liking for male humans when the latter is perceived to possess a pair of semant ically
63 antagonistic qualities, such as physical athleticism and intellect (e.g., a varsity player who is at the same time a national merit scholar), or rugged and sensitive (e.g., a rock star who is emotionally attentive). The role of Bayesian prediction in pleasure This next discussion attempts to use Bayesian prediction in the context of Pavlovian conditioning as the possible underlying principle behind increased pleasure generation across sensory and semantic categories. The response properties of the h edonic affective system appear to be consistent with a predictive system that attempts to obtain the widest range of needs with the smallest amount of effort (Rolls 2005). In the cases of both sensory and semantic contrasts, it appears to be the case that each antagonistic (sensory or semantic) percept is predictive of an orthogonal portfolio of qualities. For example, red stimuli such as strawberries and tomatoes are predictive of Vitamin C and lycopene, whereas antagonistic green stimuli such as spinach a re predictive of a different set of vitamins and micronutrients. In the same way, the semantic category athleticism is predictive of an orthogonal portfolio of qualities, maximally independent of the portfolio of qualities predicted by the semantic category intelligent. The idea that these antagonistic percepts are maximally independent is a function of the Bayesian probabilistic distribution of concurrent pairs of semantic categories highly athletic individuals are usually not highly intellectual at the same time. In the same way, it is not probabilistically common for an organism to encounter both red and green stimuli, or both blue and yellow stimuli (which provide maximally orthogonal nutrients) during the course of grazing, so those stimulus packa ges are especially attractive, compared to stimulus packages that do not contain antagonistic sensory or semantic qualities (e.g., athlete or responsible intellectual).
64 Table 42. Opponent semantic categories Semantic Category Semantic Percept Antagonist ic Percept Physicality Refinement Telicity Intellectual Rugged For work Athletic Elegant For play Semantic representations such as intellectual and athletic can become semantically opposite as a result of Bayesian probabilistic distribution and the infrequent co activation of sensory motor representations attached to these two semantic constructs (Andrews, Vigliocco, and Vinson 2009). One goal of this research program would be to identify semantics that are perceived to be opposite, and to test whe ther a stimulus (e.g., a brand) that contains contrasting semantic representations is considered to be more valuable than a stimulus that does not.
65 CHAPTER 5 AFFECT GATING The twosystem, two affect framework was motivated and described in the first four chapters. The first four chapters contain a number of propositions that flow from the framework as theorized. In the current chapter (Chapter 5), I will present five empirical studies that test a portion of the propositions in the framework. The term Af fect Gating refers to the process by which the state affective system (i.e., loosely the dopaminergic/vassopressinergic system) changes the output of the hedonic affective system (i.e., loosely the opioid system) that is connected to different sensory channels, making different sensory aspects of the consumption experience dominant. Because of this affect gating, the consumer experience differs according to the state affect that is modulating neural transmission in the organism. How Affect States Influence Relevance of Sensory Channels Extant models of affect and cognition assume that affect can play three different roles in information processing (Cohen, Pham, and Andrade 2008; Patrick and MacInnis 2006; Shiv 2007). First, affect can be an input in decision making. Affect can exert a direct influence (a ffect as information, e.g., Schwarz and Clore 2007) or an indirect influence (e.g., mood congruency, Bower 1981) on a judgment. Second, affect can influence the quality of decision making. For example, pos itive affect leads to greater creativity (Greene and Noice 1988), b etter problem solving (Isen, Daubman, and Nowicki 1987), and greater risk taking (Kahn and Isen 1993), but can decrease judgment accuracy ( Keller, Lipkus and Rimer 2002). In contrast, negative affect decreases the ability to cope with feedback (Agrawal, Menon, and Aaker 2007), increases suboptimal choice switching (Ratner and Herbst 2005), increases a focus on
66 concrete attributes ( Labroo and Patrick 2009), and induces a bias toward positi vely framed information (Keller, Lipkus, and Rimer 2003; Trope and Pomerantz 1998). Third, affect can be the focus of decision making. People try to repair their negative moods (Patrick and MacInnis 2006; Zillmann 1988) and maintain their positive moods (A ndrade 2005; Clark and Isen 1982), and will try to attend to information that will regulate their longterm affect (Keller, Lipkus, and Rimer 2003). Common to all of these research streams is the assumption that cognitive processes dictate the use, influence, and management of affective experience. Neural models of affective processing argue that the affective system and the cognitive system collectively control information processing (Barbano and Cador 2007; Berridge 2009; Berridge and Robinson 1998). These models assume that affective processes manage perceptual experience in order to induce the organism toward stimuli that enhance the organisms longterm survival. One such model, the opioid mediation model, assumes that there are hedonic circuits that are the target of affective regulation (Pecina, Smith, and Berridge 2006). Affective regulation is accomplished by a gating process wherein the prevailing affective state gates the information processing systems access to the most survival relevant sensory information (Carr and Bak 1987; Panksepp and Moskal 2008). For example, experiments on animals show that negative affective states increase sensitivity to affiliative touch received by the tactile channel. Upon receiving tactile stimulation, the brain induces the release of beta opioids (chemicals that have morphinelike properties) that quell the negative affective state (Keverne, Martensz, and Tuite 1989; Martel et al. 1993; Panksepp 1998). This hedonic pleasure generated from beta opioids induces the animal to seek social affiliation when
67 it is organismically vulnerable (as signaled by the negative affective state). The sensory channel opened by affective gates is "privileged" in the sense that other sensory channels, such as the visual channel, cannot u nconditionally (i.e., without the need for associative learning) induce the release of the same opioids generated by social rewards (Panksepp 1998). There are evolutionary and developmental reasons for affective gating. For example, a negative affective s tate is a visceral indication that the perceiver is organismically vulnerable, such as when the seal pup is physically injured or sick, or when it has been separated from its parent for an extended period of time (Panksepp, Nelson, and Bekkedal 1997). Vulnerable organisms that return to their mothers when they experience negative affective states receive tactile stimulation. Tactile stimulation then induces beta opioid release, thus returning the juvenile to affective and organismic homeostasis (Keverne et al. 1989). Juveniles that are more sensitive to tactile stimulation when organismically vulnerable receive physical support from their mother and are more likely to survive. Survival then allows these juveniles to sexually mature and pass on their gene codes (expressing the response pattern of high tactile sensitivity in negative affective states) to offspring. Over hundreds of thousands of years, the process of natural selection results in privileged (i.e., gated) sensory channels in the affective system In contrast, a positive affective state is a visceral indication that the juvenile is organismically self sufficient and should explore its environment. Juveniles that make visual explorations during positive affective states develop better predatory and foraging abilities, hence become better hunters or food gatherers. In the end, natural selection dictates that the mammal has to assign fluctuating value to sensory
68 channel stimulation (i.e., it has to develop a tendency to be more perceptually and hedoni cally sensitive to different sensory channels in different affective states) in order to survive, reproduce, and pass on its genes. Affective gating helps the organism assign fluctuating value to sensory channel information without requiring the participat ion of higher order cognition (i.e., without requiring the animal to think). If current affect states gate the perception of stimulation from competing sensory channels, then this gating process should have implications for product experience and the amount of hedonic response. First, products that are experienced under a positive or negative affective state should be perceived along the sensory channel dimensions that are favorably gated by the state. Second, people should be more sensitive to physical ch anges in the product if these changes occur along the gated sensory channels. Third, any consumer behavior that amplifies the primary sensory channel used for consumption, or inhibits competing sensory channels, should increase the intensity of the sensory perception and influence the amount of pleasure generated from that sensory perception. This article provides evidence for these three processes using a common consumer product (hand lotion). T wo T ypes of A ffect Two distinct types of affect are relevant t o consumer judgment and decision making: state affect and hedonic affect (Barbano and Cador 2007). State affect refers to the nonspecific, global affective state of the consumer (e.g., a mood). It is nonspecific in the sense that perceivers are typically unable to accurately attribute the source of the affect. For example, consumers in a positive state affect (henceforth positive affect) experience a light, airy feeling that makes them feel alive and upbeat, although they do not necessarily know why they feel good (Ciompi and Panksepp 2005; Panksepp 1998;
69 Sacks et al. 1972). Consumers in a negative state affect (henceforth negative affect) experience a heavy feeling that makes them feel downcast and lethargic, a global state that is sometimes called anergia to indicate the lack of psychic and physical energy (Salamone, Cousins, and Bucher 1994). Because the source of state affect is difficult to accurately identify, its diffuse influence on different neural networks can exert an assimilative influence on the evaluation of products ( Gardner 1985), attitudes toward an ad ( Goldberg and Gorn 1987), decisions about future consumption (Pham 1998), and judgments of life satisfaction ( Schwarz and Clore 1983). State affect exerts an influence on judgments by making v alence congruent information more accessible (Isen et al. 1978), by directly serving as information (Schwarz and Clore 1983), and by affect transfer (Morales and Fitzsimons 2007). In organisms with highly developed prefrontal cortices, difficulty from mak ing a decision (Luce, Payne, and Bettman 1999) can also lead to negative affect, which can lead to choosing options that minimize the negative affect generated by the choice difficulty (Drolet and Luce 2004). Neurologically, this choice difficulty is generated by CRF activation, so it still falls under state affect In contrast, hedonic affect refers to the affective experience (e.g., hedonic response) that is generated as a direct function of a product exposure or consumption experience. Positive hedonic affect is what consumers experience as pleasure, a positive hedonic feeling that is correctly attributed to sensory stimulation from an object that is consumed (Barbano and Cador 2007; Berridge and Robinson 1998). Negative hedonic affect is what consumers experience as aversion (Hoebel, Avena, and Rada 2008; Morales and Fitzsimons 2007). The lack of positive hedonic affect is generally referred to as satiation or loss of pleasure, sometimes called anhedonia to indicate
70 the lack of hedonic pleasure (Panksepp 1998; Rolls 2005; Salamone, Cousins, and Bucher 1994). There is psychological and neurological evidence showing that state affect and hedonic affect are separate forms of affect. State affect is a nonspecific state that can influence a variety of unrelated judgments in an assimilative direction, as long as the source of the affective state is not salient (Schwarz and Clore 1983). In contrast, hedonic affect is a stimulus specific, rapid response that is more likely to be integrated into an evaluative response when it is salient (Berridge and Robinson 1998). Neurological evidence shows that destruction of the dopaminergic pathways that mediate state affect does not prevent the generation of hedonic affect (Berridge and Robinson 1998; Panksepp 1998; Rolls 2005). That is, an organism without intact dopaminergic pathways can become very sluggish, and not want to pursue food or sex, but once it receives tactile stimulation from intromission, it will show clear evidence of a positive hedonic response (anergia without anhedonia). Second, deactivating the opioidergic pathways that mediate hedonic affect does not prevent the generation of state affect (Berridge and Robinson 1998; Fantino, Hosotte, and Apfelbaum 1986; Panksepp 1998; Pecina et al. 2003). Subj ects in this second situation report wanting to have sex or wanting to self stimulate but not actually enjoying the tactile stimulation once it is received (anhedonia without anergia). A ffective Gating Affective gating is a process whereby an organisms affective state changes the kind of sensory input that is privileged to enter perception, subsequently influencing the ability to experience positive or negative hedonic affect (Carr and Bak 1987; Carr, Bonnet, and Simon 1982; Panksepp and Moskal 2008). This process can be described
71 as gating because the various state affect circuits, through different neuromodulators (chemical neurotransmitters that increase or decrease the activation of sensory neurotransmitters), regulate information coming in fr om the sensory channels. More specifically, mammals in a negative or positive state affective state are differentially sensitive to opioids activated by specific sensory channels (i.e., the affective state has created sensory specific needs that can only be fulfilled by the appropriate sensory stimulation). In turn, this increased hedonic sensitivity makes the animal more responsive to information from the particular sensory channel that is gated. Negative State Affect Negative state affect (henceforth, negative affect for brevity) emerges from limbic circuits that respond to physical or social psychological pain, indicating loss of physical health or loss of social support (Eisenberger, Lieberman, and Williams 2003; MacDonald and Kingsbury 2006; Panksep p 1998; Panksepp 2003; Zubieta et al. 2001). This visceral loss signal induces the organism to seek stimuli that restore physical or social resources (e.g, its mother and other sources of physical or social support). These restorative stimuli have been s hown to be detected as an unconditioned stimulus1 most reliably through tactile information, in the form of affiliative touch (McGlone et al. 2007; Panksepp et al. 1980; Panksepp 1998). In negative affective states such as pain or sadness, receptors beco me highly sensitized and disproportionately available (Zubieta et al. 2003). receptors have a molecular structure that is opened only by endorphins, the primary opioid released during tactile stimulation. When endorphins 1 The literature has shown that only affiliative touch can generate pleasure without requiring associative processes (i.e., even new born infants show a pleasure response to affiliative touch, even if they show no affective pleasure in response to the sight of their mother) (Panksepp 1998).
72 bond to the receptor, there is a subjective alleviation of the negative affective state (i.e., pain or sadness is reduced or eliminated) and a subsequent experience of hedonic pleasure is generated (Nelson and Panksepp 1998). In support of this hypothesis, Martel et al. (1993) showed that prolonging the negative affective state increases the motivation of rhesus monkeys to receive tactile stimulation from mothers or other conspecifics. In addition, the administration of morphine, which removes negative affect, significantly reduces t he ability of tactile stimulation to generate hedonic pleasure (Martel et al. 1993; Panksepp, Najam, and Soares 1979). Negative affective states thus have the surprising influence of increasing positive hedonic affective responses toward tactile stimulatio n. Positive State Affect Positive state affect (henceforth, positive affect) emerges from limbic circuits that induce the nonspecific searching for rewards and the learning of reward cues (Barbano and Cador 2007; Berridge and Robinson 1998; Panksepp 1998). In mammals, food and sexual rewards are spatially located and identified using the olfactory system (lower mammals) or visual system (higher mammals, including humans). The visual system helps the organism search for food, for example, by identifying colorful berries against a backdrop of foliage, and by identifying prey movement against a backdrop of other objects (visual motion detection and visual temporal resolution). Additionally, the visual system helps the organism safely perform seeking forward l ocomotion, one of the universal behavioral responses among mammals in a positive affective state (Panksepps Search for nuts and knowledge hypothesis, see Panksepp 1998; Ikemoto and Panksepp 1999). Positive affective states cause pupil dilation and an enlargement of the visual field (Partala and Surakka 2003), increase visuospatial breadth (Rowe,
73 Hirsh, and Anderson 2007), modulates visual temporal resolution (MoraFerrer and Gangluff 2002), and aids in visual motion detection (MoraFerrer and Gangluff 2000). The effects of positive affect on visual intake are generally adaptive, although in extreme cases, overproduction of dopamine (the main mediator of positive affect) can increase visual intake excessively, causing a failure of the affective perceptual system to perform gating against irrelevant visual stimuli ( Mansbach, Geyer, and Braff 1988). Positive affective states are instrumental i n the development of visual perception (Imanaka et al. 2008; MoraFerrer and Gangluff 2000; Rolls 2005). Positive affect indicates that the organism is self sufficient and should search its environment for rewards and reward cues. As a result, visual intak e is enhanced through autonomic and nonautonomic processes. First, eye tracking studies show that subjects in a positive affective state make more eye saccades than subjects in a neutral or negative affective state (Wadlinger and Isaacowitz 2006) and are better at visual motion detection (MoraFerrer and Gangluff 2000). Second, limited visual exploration inhibits perceptual development. For example, kittens that are prevented from visual exploration become deficient in judgments of visual depth and line or ientation (Hubel and Wiesel 1963; Wiesel and Hubel 1963a; Wiesel and Hubel 1963b). Mammals that do not perform visual exploration are more likely to become developmentally challenged, and become less able to judge color contrast, stimulus depth, and stimul us movement (Findlay and Gilchrist 2003; Held and Hein 1963). Taken together, these results are consistent with the idea that organisms in a positive affective state are hardwired to search and explore, and this exploration mode is induced by a greater se nsitivity to visual sensory input.
74 Evidence for Affective Gating Although the affective gating hypothesis has not been previously tested in humans, there are mammalian results that are consistent with the hypothesis. For example, Harlow (1958) demonstrated that a rhesus monkey will attach to a terry cloth mother when it is in a negative state, despite the terry cloth mother's absence of food (and the presence of food in a steel wire mother in an alternate location). With the benefit of neurological hindsight, it could be argued that Harlows results illustrated the privileged access of affiliative tactile channel information in a negative affective state. Similarly, Panksepp et al. (1980) demonstrated that chicks in a negative affective state (as me asured by the number and intensity of separation distress calls) stop their separation distress calls and exhibit hedonic eye closure when caressed by human hands. Hofer (1987) shows that even tactile stimulation from a flat strip of synthetic fur on the f loor reduces distress vocalizations among distressed rats. The fact that tactile stimulation from human hands (as opposed to the chicks mother) or flat synthetic fur was able to quell the animals negative affective state is consistent with the idea that nonspecific (i.e., unconditioned) tactile stimulation can alleviate a negative affective state (i.e., visual or olfactory recognition of the mother is not necessary). Lewis (2000) makes an argument from a developmental perspective. Lewis posits that the response dynamics of affective perceptual circuits iteratively wire together in response to life events. It can be argued that the hedonic response to tactile stimulation in a negative affective state is an adaptation that promotes social bonding and cohesion, whereas the lack of a hedonic response (or aversion) to tactile stimulation in a positive affective state is an adaptation that promotes environmental exploration. Without the alternating increase and decrease of hedonic experience from
75 tactile stim ulation, social cohesion and exploration cannot be optimized to advance organismic survival. The organism needs to feel pleasure from tactile contact some of the time, and not feel pleasure from tactile contact at other times, in order to achieve both the social attachment and exploration needs that promote well rounded development. Hypotheses If the affect gating premises are true, then the experience of sensory stimulation should not simply be a function of the objective quality of stimulation (e.g., vel vet is experienced as more pleasant than wood), but should also be a function of the specific fit between the current affective state and the sensory channel. In negative affective states, stimulation via tactile channels releases endorphins that bind wi th receptors to generate hedonic pleasure. Hence, a consumer in a negative affective state will be more sensitive to tactile stimulation and should, paradoxically, experience a greater hedonic affective response from tactile stimulation compared to consu mers in a positive or neutral affective state. In addition, the pleasurable hedonic experience will cause the consumer to impute higher perceived quality to the source of the tactile stimulation. H1 Consumers in a negative affective state (as compared to consumers in a neutral or positive affective state) will a. experience a product in a more tactile way; b. rate tactile product benefits more positively; c. be more sensitive to changes in tactile product benefits; d. experience an increased hedonic affective response to tactile stimulation.
76 In contrast, a consumer in a positive affective state should be more sensitive to visual product benefits. Because autonomic increases in visual sensory intake occur during positive affect states, the consumer will be more sensitive to visual product attributes (generally positive in valence for most products), and will thus rate visual product attributes more positively. Yet, visual stimulation in a positive affective state does not result in an increased hedonic response toward tactile stimulation because visual channels do not have pathway access to brain opioids that activate receptors. H 2. Consumers in a positive affective state (as compared to consumers in a neutral or negative affective state) will a. experience a product in a more visual way; b. rate visual product benefits more positively; c. be more sensitive to changes in visual product benefits; d. not experience an increased hedonic affective response to tactile stimulation. To the extent th at negative and positive states induce a person to be more sensitive to tactile or visual information, respectively, and to the extent that this sensitivity leads to a more positive evaluation of product attributes, the consumer should value the product more. When in a negative affective state, this increased valuation of the product should be driven by the hedonic response that results from the tactile stimulation (i.e., mediated by hedonic affect). When in a positive affective state, this increased valuat ion of the product should be driven by the affective state itself (i.e., mediated by state affect). H 3. Consumers in a negative affective state (as compared to consumers in a neutral or positive affective state) will pay more for a product that is
77 experi enced tactilely. The increased product valuation will be mediated by the hedonic affective response to the tactile stimulation (hedonic affect). H 4. Consumers in a positive affective state (as compared to consumers in a neutral or negative affective st ate) will pay more for a product experienced visually. The increased product valuation will be a direct consequence of the affective state of the consumer (state affect). Experiment 1 Experiment 1 test ed the hypothesis that consumers affect ive state s can influence their sensitivity to information from competing sensory channels (hypotheses H1a and H2a) Participants were asked to experience a product (s kin lotion) while in a negative neutral, or positive affective state. Afterwards, participants were aske d to describe the product experience in their own words. It was expected that participants in a negative affective state would be more sensitive to the tactile product experience, whereas participants in a positive affective state would be more sensitive t o the visual product experience. Design and Procedure Design Forty eight undergraduate students participated in the study for course credit. The study used a betweensubject manipulation of the affective state (negative, neutral, positive). Participants w ere induced into an affective state, tried a product (hand lotion), and recorded their perceptions of the product. Procedure Participants were invited into a behavioral lab and were told that the study investigated how consumers experience products under different emotional states. It
78 was explained that people in certain professions (health care, military) often experienced temporary and/or prolonged emotional states and that these moods could influence their appreciation for products. Participants were told, your task is to help us design products by sincerely trying your best to feel the emotion that is suggested by the statements you are about to read. Participants were then induced into a negative, neutral, or positive affective state using the V elten (1968) affect induction procedure. The Velten procedure involves reading a series of statements that get progressively more negative (negative affect condition), progressively more positive (positive affect condition), or remain consistently neutral (control condition). Participants were told to read the statements into a microphone, and to take their time imagining how each statement applied to their lives. For example, the negative affective state manipulation starts with mildly negative statements such as I feel a little bit low today and ends with strongly negative statements such as All of the unhappiness of my past life is taking possession of me. The positive affective state manipulation starts with mildly positive statements such as I feel light hearted and ends with strongly positive statements such as God, I feel great! The neutral affective state manipulation starts with affectively neutral statements such as Oklahoma City is the largest city in the world in area, with 631.166 square miles and ends with affectively neutral statements such as At low tide the hulk of the old ship could be seen. This Velten procedure has been found to reliably induce physiological changes corresponding to positive and negative affect, both in terms of neural activation and psychopharmacology (Brown et al. 1993; Gadea et al. 2005).
79 After the affective state induction, subjects were asked to try a hand lotion. The lotion was Crystal Waters. The lotion was light blue in color and had a tropical lily sce nt, and was placed in an 8ounce, clear plastic bottle with a pump top. Each participant had a bottle of lotion in their personal cubicle. After trying the lotion, participants were asked to describe the product using any of the five senses. Product descr iption coding The product descriptions were coded by two judges who were blind to experimental condition, with disagreements resolved through negotiation. Judges coded perceptions in reference to tactile, visual, and olfactory dimensions. Perceptions relat ed to the tactile dimension included feels smooth and silky, moisturizing, and cool to the touch. Perceptions related to the visual dimension included has a nice blue color, looks calming, and looks like the ocean. Perceptions related to the ol factory dimension included smells good, floral scented, and smells sweet. Velten procedure manipulation check The effectiveness of the Velten (1968) procedure in inducing affective states was confirmed using the Positive Affect Negative Affect Sca le (PANAS) first developed by Watson, Clark, and Tellegen (1988) in the psychology literature and subsequently adopted for marketing by Aaker, Drolet, and Griffin (2008). Eighty three participants were induced into a negative, neutral, or positive affectiv e state using the Velten procedure. Participants then used 11point scales (0 = not at all and 10 = extremely ) to report how unhappy, disappointed, depressed, bad, unfavorable, dissatisfied, happy, elated, upbeat, good, favorable, and satisfied they felt i n real time. The six negative affect items were combined to form the negative af the six positive affect items were combined to form the positive affect index (Cronbachs
80 affect (F (2, 80) = 41.64, p < .01) and positive affect ( F (2,80) = 35.71, p < .01) indices. Participants in the negative affect condition felt more negative ( M = 5.57) than the participants in the neutral (M = 2.60, F (1, 80) = 30.92, p < .01) or positive ( M = 1.13, F (1, 80) = 67.72, p < .01 ) affect conditions. Participants in the positive affect condition felt more positive ( M = 7.54) than the participants in the neutral ( M = 3.96, F (1, 80) = 39.41, p < .01) or negative ( M = 2.58, F (1, 80) = 79.69, p < .01) affect conditions. Results The dat a were analyzed using a repeatedmeasure ANOVA with the number of tactile and visual product descriptions as the repeated measure and affective state as the betweensubjects factor. The affective state by sensory channel perceptions interaction was signifi cant ( F (2, 45) = 14.30, p < .05). The number of tactile perceptions ( F (2, 45) = 6.43, p < 0 1 ) and visual perceptions ( F (2, 45) = 12.57, p < 0 1 ) varied by affective condition. Participants experiencing a negative affect ive state generated more tactile per ceptions ( M = 2.07) than participants experiencing a neutral ( M = 1. 19; F (1, 45) = 10.19, p < .01 ) or positive ( M = 1.24 ; F (1, 45) = 9.93, p < .01) affect ive state. Participants experiencing a positive affect ive state generated more visual perceptions ( M = 1 .0 6 ) than participants experiencing a neutral ( M = 31; F (1, 45) = 16.17, p < .01 ) or negative ( M = .2 0; F (1, 45) = 20.70, p < .01 ) affect ive state. A supplemental analysis showed that olfactory perceptions did not vary by affective condition ( MNegative = 1. 07, MNeutral = 1.31, MPositive = 1.29; F (2, 45) = .82, p > .1 0). Discussion Experiment 1 demonstrates that consumers in a negative affective state generated more tactile perceptions. In contrast, consumers in a positive affective state
81 generated more visual perceptions. The affective statespecific response patterns provide preliminary evidence in support of hypotheses 1a and 2a. The competing pattern of results on the tactile and visual measures, and the null effect on the olfactory measure, suggest that the results are not a consequence of participants being more motivated to process information in certain affective states. For the remaining studies, it is important to distinguish between two types of tactile perception (McGlone et al. 2007; Pecina, Smith and Berridge 2006). Discriminative touch is mediated by fast conducting sensory pathways (McGlone et al. 2007). Discriminative touch corresponds to what is the tactile quality of this stimulus? In contrast, affective touch is mediated by slow co nducting sensory affective pathways (Vallbo, Olausson, and Wessberg 1999; Rolls 2005). Affective touch corresponds to phenomenological experiences such as how good does this stimulus make you feel (McGlone et al. 2007)? Hence, perceived tactile quality, and sensation in general, is independent from affective experience and can be separately measured and tested. Experiment 2 Experiment 2 test ed the hypothesis that consumers affective states (state affect ) can influence their hedonic response ( hedonic aff ect ) toward information from competing sensory channels. Consumers were asked to experience a product (skin lotion) while in a negative neutral, or positive affective state. Afterwards, participants were asked to rate the tactile, visual, and olfactory quality of the product, the hedonic experience of using the product, and the amount they would pay for the product. It was expected that participants in a negative (positive) affective state would rate the tactile (visual) quality of the product more favorably (Hypothesis 1b and 2b) Additionally, it was expected that participants in a negative affective state should have a better hedonic
82 experience from using the product (Hypothesis 1d) and should be willing to pay more for the product because of this hedoni c experience (Hypothesis 3). Participants in a positive affective state should not have a greater hedonic experience from using the product, but should be willing to pay more for the product because their state positive affective state generates a mood congruent evaluative perception (Hypothesis 4). Design and Procedure Seventy undergraduate students participated in the study for course credit. The study used a betweensubject manipulation of the affective state (negative, neutral, positive). The procedu re was identical to the procedure used in experiment 1 except for the dependent measures. First, participants indicated their hedonic affective response by rating how pleasurable it was to use the product from 0 (Does not feel pleasurable at all) to 10 (Feels very pleasurable). Then, participants rated the tactile (silky feel), visual (visually appealing), and olfactory (sweet smelling) quality of the product from 0 (not silky, not visually appealing, not sweet smelling) to 10 (very silky, ve ry visually appealing, very sweet smelling). Finally, participants stated the dollar amount they would be willing to pay for a 10ounce bottle of the lotion. Results Primary a nalysis The mean responses are presented in table 1. The data were initially analyzed using a repeatedmeasure ANOVA with the tactile and visual product ratings as repeated measures and affective state as the betweensubjects factor. The affective state by product rating interaction was significant ( F (2, 67) = 4.65, p < .05). The t actile ratings ( F (1, 67) = 4.78, p < 05) and visual ratings ( F (1, 67) = 7.05, p < 05) varied by
83 affective condition.2 Participants in a negative affective state rated the tactile quality of the lotion more favorably ( M = 6.91) than participants in the neutral affective state ( M = 5.42; F (1, 67) = 6.28, p < .05), whereas the ratings of participants in the positive affective state ( M = 6.05) did not vary from the ratings of participants in the neutral affective state ( M = 5.42; F (1, 67) = 1.33, p > .05). Participants in a positive affective state rated the visual appeal of the lotion more favorably ( M = 7.35) than participants in the neutral affective state ( M = 5.62; F (1, 67) = 6.46, p < .05), whereas the ratings of participants in the negative affective st ate ( M = 5.86) did not vary from those of participants in the neutral affective state ( M = 5.62; F (1, 67) = .12, p > .05). Olfactory quality ratings did not vary by affective state ( MNegative = 7 24, MNeutral = 7.42, MPositive = 7 70; F (2, 67) = .23, p > 05). The next analysis focused on the hedonic response to the lotion and the amount the participant was willing to pay to buy the lotion. The hedonic response ( F (1, 67) = 4.78, p < 05) and willingness to pay ( F (2, 67) = 3.22, p < 05) varied by the affec tive state condition. Participants in a negative affective state indicated that the lotion made them feel better ( M = 6.71) relative to participants in the neutral affective state ( M = 5.12; F (1, 67) = 6.47, p < .05), whereas the ratings of participants in the positive affective state ( M = 5.87) did not vary from the ratings of participants in the neutral affective state ( M = 5.12; F (1, 67) = 1.51, p > .05). Participants in a negative affective 2 Theory predicts that the tactile quality ratings should be higher in the negative affective state than in the neutr al or positive affective states. The affective state variable was contrast coded (e.g., negative = 1, neutral = .5, positive = .5) to be consistent with this prediction (see Keppel and Wickens 2004, pp. 8083). Contrast coding was used for tactile quali ty ratings, visual quality ratings, and hedonic responses. Contrast coding was retained for analyses performed in subsequent experiments and will not be discussed further.
84 state ( M = $4.62) and a positive affective state ( M = $4.78) were willing to pay more for the lotion than participants in the neutral affective state ( M = $3.21; F (1, 67) = 4.09, p < .05; F (1, 67) = 5.32, p < .05). T able 51. The evaluation of skin lotion under different affective states. Dependent Measure Affect ive State Tactile Visual Olfactory Hedonic Value ($) Negative 6.91 a 5.86 b 7.24 6.71 a 4.62 a Neutral 5.42 b 5.62 b 7.42 5.12 b 3.21 b Positive 6.05 b 7.35 a 7.70 5.87 b 4.78 a NOTE Column means with different superscripts are different at p < .05. Mediation analysis for negative affective state participants The Baron and Kenny (1986) threepart procedure was used to assess whether tactile quality rating and hedonic affective res ponse mediated the influence of the affective state on the amount the participants were willing to pay for the lotion (i.e., sensory mediation). The analysis used participants in the negative and neutral affective states. First, the affective state was a s 1.41, SE = .58, t (45) = 2.45, p = .02). Second, the affective state was a significant 1.48, SE = .55, t (45) = 2.68, p = .01) and the 1.60, SE = .65, t (45) = 2.46, p = .02). Third, when the willingness to pay measure was regressed on the affective state and the tactile quality, the affective state coefficient became non.81, SE = .58, t (44) = 1.40, p = .17) and the tactile SE = .15, t (44) = 2.78, p < .01; Sobel z = 2.13, p = .03). Likewise, when the willingness to pay measure was regressed on the affective state and the hedonic response, the affective state coefficient beca me non.62, SE = .52, t (44) = 1.21, p = .23) and the SE = .11, t (44) = 4.42, p <
85 .01; Sobel z = 2.22, p = .03). These results indicate that tactile quality and hedonic response mediated the relationship between affect and willingness to pay among negative affect subjects. Lack of mediation for positive affective state participants The Baron and Kenny (1986) threepart procedure was used to establish that the visual quality rati ng and hedonic response did not mediate the influence of the affective state on the amount the participants were willing to pay for the lotion (i.e., no sensory mediation). The analysis used participants in the positive and neutral affective states. First, SE = .68, t (47) = 2.29, p = .03). Second, the affective state was a significant predictor of SE = .68, t (47) = 2.55, p = .01) but not of the hedonic SE = .59, t (47) = 1.27, p = .21). Third, when the willingness to pay measure was regressed on the affective state and the visual quality measures, the SE = .72, t (46) = 1.64, p = .11) and t he visual quality SE = .15, t (46) = 1.54, p = .13) both became nonsignificant. The nonsignificance of the hedonic state coefficient in part two of the analysis and visual quality coefficient in part three of the analysis indicates t hat neither process is a mediator among positive affect subjects. Discussion Experiment 2 provides additional evidence for the affective gating hypothesis. Participants in a negative (positive) affective state were more sensitive to tactile (visual) senso ry information. People in the negative affective state experienced a positive hedonic response (hedonic affect) from the tactile stimulation, which in turn resulted in a greater willingness to pay for the product. People in the positive affective state were
86 more sensitive to visual stimulation, but this visual stimulation did not drive the valuation of the product. Instead, the affective state was responsible for the valuation of the product. This mood congruency effect is consistent with many state affect findings in the psychology and consumer behavior literature (Forgas 1995; Gardner 1985). If different affective states indeed make the consumer sensitive to different sensory channels, and if the tactile sensory channel indeed has privileged access to hedonic affect under negative affective states (through the interaction of endorphins with receptors), then consumers should generate less endorphins, and subsequently less liking, for the product when tactile quality is reduced, because a lower tacti le thickness induces less neural activation in the tactile channel. However, for consumers in a positive affective state, the slight decrease in objective tactile quality should go unnoticed. These predictions were investigated in experiment 3. Experiment 3 Experiment 3 test ed the hypothesis that consumers affect ive state s ( state affect ) can influence their sensitivity to changes in the objective quality of sensory stimulation (Hypotheses 1c) Consumers were asked to experience a high or low quality3 product (skin lotion) while in a negative neutral, or positive affective state. Afterwards, participants were asked to rate the tactile, visual, and olfactory quality of the product, their hedonic experience from using the product, and the amount they would pay for the product. It was expected that participants in a negative affective state would be more sensitive to changes in the tactile quality of lotion compared to 3 To be precise, the terms high and low quality really correspond to tactilely viscous and tactilely less viscous the terms high and low quality are used for brevity. In a separate pretest, consumers lay theory of a high quality lotion was a tactilely viscous (creamy) lotion, with very few participants preferring a less v iscous (more runny) lotion.
87 participants in a neutral or positive affective state. The differential sensitivity to the tactile quality was also expected to be reflected in the consumers hedonic experience, and willingness to pay, for the product. Design and Procedure Design One hundred eighty eight undergraduate students participated in the study for course credit. The study used a 3 x 2 betweensubjects design, in which there was a betweensubject manipulation of the affective state (negative, neutral, positive) of the participant and a betweensubject manipulation of the objective quality (low quality, high quality) of the product. Procedure The procedure was identical to that used in experiment 2. The low quality lotion was created by mixing 8 mL of water into every 100 mL of the lotion. Each bottle of lotion was shaken 20 seconds before the start of each experimental session to ensure product consistency. The high quality lotion was the lotion used in previous experiment s. Results Primary analysis The mean responses are presented in table 2. The key predictions involved the tactile quality, hedonic response, and w illingness to pay dependent measures. The affective state by objective product quality interaction was significant for the tactile quality rating ( F (1, 184) = 6.01, p < .05).
88 Consumers in a negative affective state rated the tactile quality of the lotion to be lower when the lotion had a low ( M = 5.38) as opposed to high ( M = 6.93) objective quality ( F (1, 182) = 8.80, p < .05) whereas neutral ( MLow Quality = 5.57, MHigh Quality = 5.34; F (1, 182) = .18, p > .05) and positive ( MLow Quality = 5.55, MHigh Quality = 5.74; F (1, 182) = .12, p > .05) affective state consumers rated the two lotions to be equally good. It was predicted that the sensitivity to tactile quality exhibited by the consumers in the negative affective state would influence their hedonic res ponse to the lotion and the amount they were willing to pay for the lotion. Consumers in a negative affective state had a more negative hedonic response to the low ( M = 5.59) as opposed to high ( M = 6.97) objective quality lotion ( F (1, 182) = 6.32, p < .0 5) whereas neutral ( MLow Quality = 5.17, MHigh Quality = 5.47; F (1, 182) = .30, p > .05) and positive ( MLow Quality = 5.59, MHigh Quality = 5.77; F (1, 182) = .11, p > .05) affective state consumers had similar hedonic responses. Consumers in a negative af fective state wanted to pay less for the low ( M = $3.29) as opposed to high ( M = $4.71) objective quality lotion ( F (1, 182) = 7.86, p < .05) whereas neutral ( MLow Quality = $2.98, MHigh Quality = $3.62; F (1, 182) = 1.58, p > .05) and positive ( MLow Quality = $3.79, MHigh Quality = $4.40; F (1, 182) = 1.44, p > .05) affective state consumers wanted to pay similar amounts. Table 52. The evaluation of different qualities of skin lotion under different affective states. Dependent Measure Affective State Tac tile Visual Olfactory Hedonic Value ($) Negative Low Quality 5.38 5.29 7.24 5.59 $3.29 High Quality 6.93 5.17 7.48 6.97 $4.71 Neutral
89 Table 52. Continued Dependent Measure Affective State Tactile Visual Olfactory Hedonic Value ($) Low Quality 5.57 5.83 7.70 5.17 $2.98 High Quality 5.34 5.75 6.88 5.47 $3.62 Positive Low Quality 5.55 6.74 6.70 5.59 $3.79 High Quality 5.74 6.62 7.24 5.77 $4.40 Mediation a nalysis for negative affective state participants The Baron and Kenny (1986) threepart procedure was used to assess whether the tactile quality rating and hedonic response mediated the influence of the objective quality of the lotion on the amount that the participants were willing to pay for the lotion. The analysis used participants in the negative affective state. 1.41, SE = .50, t (61) = 2.85, p < .01). 1.55, SE = .50, t (61) = 3.13, p < .01) and th SE = .50, t (61) = 2.80, p < .01). Third, when willingness to pay was regressed on the objective quality and the tactile quality ratings, the objective quality coefficient became nonSE = .49, t (60) = 1.56, p .42, SE = .12, t (60) = 3.60, p < .01; Sobel z = 2.56, p = .01). Likewise, when the willingness to pay was regressed on the objective quality and the hedonic response, the objective qual ity coefficient became non SE = .49, t (60) = 1.75, p SE = .12, t (60) = 3.46, p < .01; Sobel z = 2.33, p = .02).
90 Discussion Experiment 3 demonstrates that consumers in a negative affective state were more sensitive to changes in the objective tactile quality of a lotion, leading to changes in the amount that they were willing to pay for the product In contrast, consumers in neutral and positive affective states were relatively insensitive to changes in the objective quality of the lotion, leading to no significant differences in the amount that they would pay for the lotion. A simple way to describe the results thus far is to claim that people in a negative state are sensitive to tactile information, whereas people in a positive state are sensitive to visual information. This explanation is a bit too simplistic because it suggests that the perceptual sensitivity differences between people in a negative and positive affec tive state is a function of directed attention. This is not so. Neurophysiological findings suggest that people in a negative state are differentially receptive to information from the tactile channel (i.e., there is a gating effect). That is, if visual c hannel information were to be reduced, the emphasis on the tactile channel information would increase and the hedonic experience would become more intense. However, people in a positive state are not affect gated, hence, are not affectively receptive to the tactile channel information. Thus, even if the visual channel information were to be reduced, the person would still be unable to affectively appreciate the tactile channel information. In effect, the person is the positive state is affectively numb (in relative terms) to the influence of tactile information, in the same way that a person who just ate a large amount of ice cream is affectively num b to the sweetness of ice cream (that is, the person tastes the ice cream but does not experience pleasure fr om it).
91 Experiment 4 Experiment 4 test ed the hypothesis that blocking the visual channel could increase a negative state consumers sensitivity to the activation in the tactile sensory pathways. In the experience of tactile stimulation, blocking a competing sensory channel (visual channel) transfers neural activation to the remaining, unblocked channel, increasing subsequent affective sensitivity. In turn, this increased sensitivity should be experienced hedonically and result in an increased valuation of the product. In contrast, blocking the visual channel should not influence consumers in a neutral or positive affective state, because these consumers are not affectively receptive (in relative terms) to tactile information. Design and Procedure Design Two hundred and fifty one undergraduate students participated in the study for course credit. The study used a 3 x 2 betweensubjects design, with affective state (negative, neutral, positive) and visual channel blocking (block, control) as the betweensubjects factor s. Participants were induced into an affective state, and then either experienced the lotion while closing their eyes (block visual channel condition) or while keeping their eyes open (control condition). The participants then rated their hedonic experience, the products perceived tactile, olfactory, and visual qualities, and indicated their willingness to pay in dollars and cents. One participant did not follow the blocking instructions and two respondents listed prices were more than three standard deviations from the mean. These data points were excluded from the analyses.
92 Procedure The procedure w as identical to experiments 2 and 3 with the exception of the sensory channel blocking condition. Participants in the blocking condition were told, Please close your eyes and try the lotion. Remember to keep your eyes closed (for 20 seconds or so) while you are applying the lotion. In the control condition, participants were not given any instructions. The experiment administrator and assistant confirmed that participants in the control condition kept their eyes open during the application of the product. Results Primary analysis The mean responses are presented in Table 3. The key predictions involved the tactile quality, hedonic response, and w illingness to pay dependent measures. The affective state by blocking interaction was significant for the tactile quality measure ( F (1, 244 ) = 3.95 p < .05). Consumers in a negative affective state rated the tactile quality of the lotion higher when their eyes were closed during product application ( M = 7.30) as opposed to when their eyes were open ( M = 6.30; F (1, 242 ) = 4.3 0 p < .05) whereas neutral ( MClosed = 5.27, MOpen = 5.41; F (1, 242) = .09, p > .05) and positive ( MClosed = 6.19, MOpen = 6.36; F (1, 242) = .14, p > .05) affective state consumers were insensitive to the visual channel blocking manipulation. It was predicted that the sensitivity to blocking exhibited by the consumers in the negative affective state would influence their hedonic respons e to the lotion and the amount they were willing to pay for the lotion. Consumers in a negative affective state had a more positive hedonic response with their eyes closed ( M = 7.50) as opposed to when their eyes were open ( M = 6.37; F (1, 242 ) = 5.06, p < .05) whereas neutral ( MClosed
93 = 5.51, MOpen = 5.36; F (1, 242) = .09, p > .05) and positive ( MClosed = 6.05, MOpen = 6.13; F (1, 242) = .03, p > .05) affective state consumers had similar hedonic responses. Consumers in a negative affective state wanted to pay more for the lotion when their eyes were closed ( M = $5.13) as opposed to open ( M = $3.98; F (1, 242 ) = 7.22 p < .05) w hereas neutral ( MClosed = $3.30, MOpen = $3.37; F (1, 242) = .02, p > .05) and positive ( MClosed = $4.25, MOpen = $4.88; F (1, 242) = 1.54, p > .05) affective state consumers wanted to pay similar amounts. T able 53. The evaluation of skin lotion under different affective states and visual channel blocked or unblocked ( experiment 4) Dependent Measure Affective State Tactile Visual O lfactory Hedonic Value ($) Negative Eyes Closed 7.30 4.95 7.10 7.50 $5.14 Eyes Open 6.30 6.03 6.68 6.37 $3.98 Neutral Eyes Closed 5.27 5.81 6.90 5.56 $3.30 Eyes Open 5.41 5.87 7.00 5.36 $3.37 Positive Eyes Closed 6.19 6.35 7.98 6.05 $4.25 Eyes Open 6.36 6.71 7.47 6.13 $4.88 Mediation analysis for negative affective state participants The Baron and Kenny (1986) threepart procedure was used to assess whether the tactile quality rating and hedonic response mediated the influence of the visual channel blocking on the amount that the participants were willing to pay for the lotion. The analysis used participants in the negative affective state. First, visual channel blocking was a significant predictor of the willingness to pay ( = 1. 16, SE = .5 6 t ( 78) = 2. 0 8, p < .0 5 ). Second, visual channel blocking was a significant predictor of the tactile quality ratings ( = 1, SE = 4 5, t ( 78) = 2.22 p < .0 5 ) and the hedonic response ( = 1. 1 3, SE = 49, t ( 78) = 2. 31, p < .0 5 ). Third, when the willingness to pay was regressed
94 on visual channel blocking and the tactile quality ratings, the visual channel blocking coefficient became nonsignificant ( = .35, SE = .4 4 t ( 78) = 80, p = 4 3) and the tactile quality rating coeffici ent remained significant ( = 81, SE = .1 1 t ( 78) = 7.51, p < .01; Sobel z = 2. 14, p = .0 3 ). Likewise, when the willingness to pay was regressed on visual channel blocking and the hedonic response, the visual channel blocking coefficient became nonsigni ficant ( = .28, SE = .4 3 t ( 78) = 16, p = 51) and the hedonic response coefficient remained significant ( = 78, SE = .1 0 t ( 78) = 8 10, p < .01; Sobel z = 2. 23, p = .0 3 ). Discussion Experiment four showed that blocking the visual channel increased perceptions of the tactile quality of a product, but only when a person was in a negative affective state. This result is consistent with the affect gating hypothesis. Blocking the visual channel increased sensitivity to the activation in the remaining, unblocked tactile channel. Blocking the visual channel was especially effective because the visual cortices are the most energy consuming cortices in the human brain, consuming about 60% of the total oxygen distributed to all five sensory cortical groups when a person is awake (Kringelbach 2009). Thus, blocking information acquisition by the visual channel can increase sensitivity to information presented in the remaining dominant sensory channel used in consuming a product. When people were in a negative af fective state, the activation of the tactile channel led to a greater release of endorphins that provide a pleasant feeling of hedonic affective pleasure. Thus, blocking the sensory channel that did not have access to affective gates helped enhance the pleasure of the consumption experience.
95 The neuroscience literature offers further insights on how to improve the quality of tactile stimulation in affect specific ways. Panksepp (1998) proposes that negative affect (mediated by CRF, or Corticotrophin Rel easing Factor), is accompanied by thermoregulatory changes in the dermis and epidermis (skin). Animal studies show that sad rhesus monkeys that were separated from their mothers had a lower surface temperature compared to control monkeys (Kalin, Shelton, and Barksdale 1989). Panksepp, Nelson, and Bekkedal (1997) suggest that this decrease in temperature is an adaptive response meant to induce rhesus monkeys to look for their mothers (who provide tactile warmth). It is evolutionarily advantageous for young r hesus monkeys to be in contact with their mothers, their primary source of nourishment and protection, when the young monkeys are vulnerable. Otherwise, a rhesus monkey that does not experience a negative affect mediated decrease in skin temperature may not seek its mother (the adaptive action), and this may lead to organismic damage. This is one of the interesting ways in which negative affect can have positive consequences for the organism. Once the distressed rhesus monkey is reconciled with the parent it appears that two sources of stimulation restore the monkey toward affective equilibrium: (1) Tactile stimulation from the mother (mediated by Merkel receptors on the skin), and (2) Thermal stimulation from the mother (mediated by unmyelinated C fibers for perceiving warmth). There is interaction between tactile and thermal stimulation (see Nelson and Panksepp 1998) because tactile stimulation, via rubbing or caressing the skin, can stimulate unmyelinated C fibers and generate a perception of warmth. This thermotactile stimulation leads to beta opioid release, resulting in pleasurable feelings and
96 lowered blood pressure (Uvnas Moberg 1998). This stimulation also leads to a decrease in levels of stress hormones, alleviating negative affect (Uvnas Moberg 1997). The idea that there are two approaches to increasing the intensity of tactile stimulation (rubbing and direct heat transduction) has conceptual and marketing implications. The first source of tactile stimulation (rubbing) was tested in experiment s 1, 2, and 3. These experiments suggest that tactile stimulation in human consumers influences product perception and has a significant impact on the consumers hedonic affective response to the product (see Experiment 2). The second source of tactile sti mulation (direct heat transduction) will be tested in experiments 5A and 5B: Experiment 5A tests whether negative affect causes a physiological response of coldness among humans, and experiment 5B tests whether consumers under negative affect, who feel col der, would experience greater pleasure from warming products. Experim ent 5a Experiment 5A tested the hypothesis that a negative affective state causes a physiological response of coldness among consumers. Participants were induced into a negative or neut ral affective state. Afterwards, participants were asked to rate how cold their hands felt. It was predicted that participants under negative affect would rate their hands to be colder than participants under neutral affect. Design and Procedure Design Forty two undergraduate students participated in the study for additional course credit at a southeastern university in the United States of America The study used a betweensubjects design with affective state (negative, neutral) as the between subjects fa ctor.
97 Procedure The procedure used the same Velten (1968) affect induction procedure used in the previous experiments. After the subjects were induced into their respective affective state conditions, they read the question How cold do your hands feel ri ght now? (on a scale from 0 to 10, with 10 being coldest) and indicated their answer on a blank next to the question. Results The main effect of affective state is significant ( F (1,40) = 4.16, p < .05). The participants in a negative affective state ( M = 4.23) felt colder than the participants experiencing a neutral affect ive state ( M = 2.80; F (1, 40 ) = 4. 16, p < .05). Discussion Participants under a negative affective state felt colder than participants under a neutral affective state, a response pat tern that is consistent with the idea that negative affect induces a physiological response of coldness to make it more likely for the organism to seek physical or social support when it is organismically vulnerable. If a physiological response of coldnes s indeed results from negative affective states, then we should expect that a sensory stimulation of tactile warmth should be perceived as more pleasurable for consumers under negative affect (but not for consumers under neutral affect). Experiment 5B test s this idea in the context of the consumption of warming versus nonwarming products. Experiment 5b Experiment 5B tested the hypothesis that increasing the tactile warmth of a product can enhance a consumers hedonic response to the product when the consu mer is in a negative affective state. Participants were induced into a negative or
98 neutral affective state. Afterwards, participants were asked to try either a regular or warming lotion, and then indicate how the lotion made them feel as well as rate the perceptual qualities of the lotion. It was expected that participants in a negative affective state would show a greater hedonic response, rate the tactile qualities of the lotion more favorably, and indicate a higher willingness to pay for the warming ver sion of the product Design and Procedure De sign One hundred sixty nine undergraduate students participated in the study for course credit at a southwestern university in the United States of America The study used a 2 x 2 betweensubjects design with a ffect ive state (negative, neutral ) and tactile warmth (warming, control) as betweensubjects factor s Procedure and Stimuli The procedure was identical to experiments 2 and 3. The control lotion was the unadulterated lotion and the warming lotion was the same lotion infused with eucalyptus oil, a wellestablished warming agent. Exactly 2 mL of eucalyptus oil was infused into 140 mL of lotion using a latex free syringe and subsequently shaken. Each bottle of lotion was shaken for twenty seconds prior to each experimental session to ensure that the warming ingredient was consistently distributed in the lotion. Results Primary analysis The key predictions involved the tactile quality, hedonic response, and willingness to pay dependent measures. The affectiv e state by tactile warmth interaction was significant for the tactile quality measure ( F (1, 165) = 4. 13, p < .05). Consumers in a
99 negative affective state rated the tactile quality of the lotion higher when it contained the warming ingredient ( M = 7.64) as opposed to when it did not ( M = 6.21; F (1, 165) = 9.98, p < .05), whereas neutral affective state consumers were insensitive to the presence of the warming ingredient ( MWarming = 5.59, MControl = 5.46; F (1, 165) = .09, p > .05). It was expected that the sensitivity to the warming ingredient by the consumers in the negative affective state would influence their hedonic response to the lotion and the amount they were willing to pay for the lotion. Consumers in a negative affective state had a more positive hedonic response to the lotion containing the warming ingredient ( M = 7.38) as opposed to the control lotion ( M = 6.18; F (1, 165) = 5.36, p < .05), whereas neutral affective state consumers had similar hedonic responses ( MWarming = 5.44, MControl = 5.59; F (1, 165 ) = .0 9 p > .05) Consumers in a negative affective state wanted to pay more for the lotion containing the warming ingredient ( M = $6.57) as opposed to the control lotion ( M = $ 5.21 ; F (1, 165) = 5. 40 p < .05) wh ereas neutral affective state consumers wanted to pay similar amounts ( MWarming = $4.06, MControl = $4.06; F (1, 165) = .00, p > .05). T able 54. The evaluation of regular or warming skin lotion under different affective states (Experiment 5B). Dependent Measure Affective State Ta ctile Visual Olfactory Hedonic Value ($) Negative Control 6.21 5.39 7.44 6.18 $5.21 Warming 7.64 5.91 6.38 7.38 $6.57 Neutral Control 5.46 5.50 6.74 5.59 $4.06 Warming 5.59 6.08 6.82 5.44 $4.06
100 M ediation analysis for negative af fective state participants The Baron and Kenny (1986) threepart procedure was used to assess whether the tactile quality rating and hedonic response mediated the influence of the warming agent on the amount that the participants were willing to pay for t he lotion. The analysis used participants in the negative affective state. First, the warming ingredient was a marginally significant predictor of the willingness to pay ( = 1. 36, SE = 72, t ( 82) = 1 .8 9 p = .0 6 ). Second, the warming ingredient was a significant predictor of the tactile ratings ( = 1 .43, SE = 42, t ( 82) = 3.44, p < .0 5 ) and the hedonic response ( = 1. 20, SE = 44, t ( 82) = 2. 71, p < .0 5 ). Third, when the willingness to pay was regressed on the warming ingredient and the tactile quality ratings, the warming ingredient coefficient became nonsignificant ( = .23, SE = 59, t ( 82) = 39, p = 70) and the tactile quality rating coefficient remained significant ( = 1 11 SE = .1 5 t ( 82) = 7.56, p < .01; Sobel z = 3.13, p < .0 1 ). Likewise, whe n the will ingness to pay was regressed on the warming ingredient and the hedonic response, the warming ingredient coefficient became nonsignificant ( = .20, SE = 61, t ( 82) = 31, p = 74) and the hedonic response coefficient remained significant ( = 9 7 SE = .1 5 t ( 82) = 6.69 p < .01; Sobel z = 2. 55, p = .0 1 ). Discussion Experiment 5B shows that consumers in a negative affective state are sensitive to tactile stimulation as well as the heat generated by the warming ingredient. We argue that the tactil e contact stimulated the Merkel receptors on the skin and the thermal ingredient stimulated the unmyelinated C fibers (i.e., tactile stimulation and tactile warmth pathways). Both types of stimulation activated endorphins that molecularly
101 interacted with receptors, generating the hedonic response. As a consequence, the product was more valued. General Discussion Taken together, the five experiments suggest that the affective perceptual system has gating properties, in which certain sens ory stimulation are intensified or inhibited depending on the affective state. Consumers appear to be more perceptually sensitive, and hedonically responsive to tactile stimulation when in a negative affect ive state than when in a neutral or positive affective state (experiment 1) The negative state induced sensitivity to tactile information results in a greater hedonic affective response (experiment 2), a greater appreciation for differences in tactile product quality (experiment 3), a greater responsive ness to increases in relative tactile stimulation through blocking a competing channel (experiment 4), and a preference for products that have been engineered to have additional tactile benefits (experiment 5B). All of the negative state induced responses to tactile information result in a higher willingness to pay for the product. Moreover, there is evidence that tactile perception and hedonic experience mediate the relationship between product attributes and product valuation. There is considerable anecdotal evidence that is consistent with the experimental results. First, consumer pursuit of tactile stimulation is a common response to negative affect (experiment 2). Children who are sad hug their stuffed animal or snuggle with their blanket, and adults who are sad often desire a massage or a hug. Second, consumers often close their eyes when they kiss or when they experience a massage, especially after a hard day (negative affect). This is consistent with sensory channel blocking effects, in which blocking a competing sensory channel transfers neural activation to the unblocked sensory channel, and intensifies the pleasure, especially
102 under negative affective states (experiment 4). Third, tactile stimulation and warmth are an effective combination for allev iating negative affective states (experiment 5B). A warm shower, which combines tactile stimulation from the dribbling water with warmth, appears to alleviate a negative affective state better than simply wearing a thick sweater (which provides warmth but not tactile stimulation). It is also interesting to note that some products are engineered to enhance tactile stimulation (e.g., multiplejet shower heads), warmth (cough & cold syrups with warming agents, e.g., Theraflu Warming), or both (e.g., BenGay Warming Capsaicin Rub). To the extent that these products are used in negative affective states (e.g., medication when sick, showering after a tough day at work), their consumption should be more pleasurable. There are four conceptual issues related to the different forms of affect and sensory experience. The first issue concerns the influence of sensory stimulation during affective states. The evidence from the studies suggests that tactile experience influences hedonic affect when people are in a negative affective state. The subsequent conceptual quest ion is How does this change in hedonic affect influence the consumers original state affect? Does tactile stimulation provide a temporary hedonic experience with no underlying influence on the affective state (e.g., the person feels good temporarily, but remains in a negative affective state: Andrade 2005; Vohs, Baumeister, and Lowenstein 2007) or does the hedonic affect from tactile stimulation turn a negative affective state into a neutral or positive affective state? The neurobiology literature suggests that the influence of tactile stimulation during a negative affective state depends on whether the organism perceives an opportunity to search its environment. If there is no clear opportunity to search the
103 environment, tactile stimulation turns a negative affective state into a neutral state. For example, isolationinduced vocalizations in rats are reduced or eliminated when they receive tactile stimulation (Hofer 1987; Kuhn, Pauk, and Schanberg 1990). However, if rats are presented with the opportuni ty to search or interact with littermates, tactile stimulation can turn a negative state into a positive state. Imanaka et al. (2008) showed that tactile stimulation not only alleviates negative affect, but also induces positive affect and makes the rats s how enhanced exploration and locomotor activity in an opensearch task. This appears to be due to the ability of the beta opioid system to indirectly (and partially) activate the dopaminergic system (responsible for inducing positive state affect) (Burgdor f and Panksepp 2006; Smith and Berridge 2007; Spanagel, Herz, and Shippenberg 1990), as long as the environment appears to be conducive for searching or playing (Panksepp 1998), with many novel and unpredictable stimuli (Schultz 2006). Thus, given an oppor tunity for consumers to subsequently explore their world or interact with other people, tactile stimulation may not only alleviate a negative affective state, but also indirectly activate a positive affective state. A second conceptual issue involves the t wo sources (types) of affect, and their implication on goal pursuit and means valuation. Although the literatures on affect (e.g., Shiv Fedorikhin and Nowlis 200 5 ; Johnson and Stewart 2004) and goal pursuit (Baumgartner and Pieters 2008; Dijksterhuis Chartrand, and Aarts 2007) are not well integrated, emerging neuroscience findings suggest that goal states are mediated by many of the same neuropeptides that mediate state and hedonic affect Positive state affect appear s to be mediated by dopaminergic and vassopressinergic activation, whereas positive hedonic affective responses appear to be mediated by opioidergic and
104 oxytocinergic activatio n. The former initiates search processes a general sense of excitement and exploration ( stimulus hypervaluation during goal pursuit), whereas the latter activates satisfaction, neural habituation, and satiation (stimulus devaluation following goal fulfil lment ). Given that the two types of positive affect have different underlying processes, the novelty unpredictability, and variety of products versus the sensory hedonic qualities of products, sh ould be differentially important in influencing goal pursui t. For example, it may be more important for product novelty (e.g., new pr oduct, new packaging) unpredictability (e.g., sales promos, product related trivia), and product line variety (e.g., number of product variations) to activate search processes stimu lated by positive state affect In contrast, the objective sensory quality of the perceptual stimulation (e.g., tactile quality thermo tactile quality ) may be more important for influencing the hedonic experience slowing down habituation, and limiting ex posureinduced satiation. Consumers under positive state affect may be more likely to search, and be attracted to, products and product lines with high novelty, unpredictability, and variety, whereas consumers under negative state affect may be more like ly use their remembered hedonic experience as a result of the sensory hedonic qualities (how good it felt, e.g., Cowley 2007; MacInnis, Patrick, and Park 2006; Pham and Avnet 2009) as the main driver for repeat purchasing. Given that consumers under positi ve state affect are relatively insensitive to veridical product information (Experiment 3), marketers should spend more effort on releasing new products, increasing unpredictability, or enhancing the variety of the product line. In contrast, given that consumers in a negative affective state are more sensitive to objective sensory product qualities (Experiment 3),
105 marketers should spend more effort on improving product quality and the sensory usage experience (e.g., suggest closing eyes during product appli cation) that will directly act on the consumers hedonic system. The correct strategy would depend on whether the product is marketed toward consumers under positive affect (celebratory products) or consumers under negative affect (comfort products, e.g., comfort foods). A third conceptual issue is the relationship between higher order and lower order processes. The consumer behavior literature shows that consumers under negative affect perform more systematic cognition (Batra and Stayman 1986; Forgas 1995). The five experiments in the present article suggest that the influence of a negative affective state is not limited to higher order thought processes such as the evaluation of the strength of message arguments, but can also come in the form of increased reliance on lower order sensory perceptual processes (bottom up processing, i.e., relying on objective sensory information as opposed to topdown, brandinduced inferences). The relationship between these two forms of affectively moderated processes is a potential area of investigation. An important social issue for consumers is how the experience of hedonic affect influences compulsive consumption. The beta opioids generated from hedonic affect appear to contribute to biologically mediated compulsive behavior, a form of aberrant goal pursuit. For example, consider sexually compulsive behavior. It is estimated that between 17 million and 37 million Americans exhibit uncontrollable sexual behaviors that result in negative consequences for the person and rel evant others ( Hagedorn and Juhnke 2005). Although public displays of socially aberrant sexual behaviors receive significant media attention, roughly 70% of sexually compulsive people perform these
106 behaviors in private (Kafka 2001). One contribution to the compulsion may be that the tactile induced hedonic pleasure is not subsequently accompanied by readily available, nonsexual opportunities to search, play, or interact. In nonenriched environments, repeated hedonic pleasure can lead to hypersensitization to object specific incentives related to the pleasuregenerating stimulation and subsequent devaluation of other stimuli (Tindell et al. 2005). Repeated hedonic pleasure and hypersensitization can increase the seeking of one specific stimulation (as opposed to nonspecific seeking of general rewards in a healthy, positive affective state) and the seeking of cues predictive of that specific source of hedonic affect (hedonic pleasure). This can cause the consumer to spiral into a cycle of hedonic self stimul ation. Thus, the lack of access to alternative means of enrichment appears to be a key moderator that tilts hedonic consumption toward compulsion, although there are undoubtedly other variables that contribute to the phenomenon. Finally, this article is an effort to bring attention to the under researched area of consumer experience and perception. Despite the undeniable importance of studying the actual consumer experience and real time hedonics in consumption, the literature in consumer behavior has hit herto focused disproportionately on high er order cognitive processes that explore how the consumer predicts the consumption experience from advertising messages branding, and other linguistic and symbolic stimuli, rather than how the consumer experiences the lower order sensory consumption of the product or service itself. The cognitive revoluti on sparked by Miller (1956) and the linguistic revolution sparked by Chomsky (1955) has served the field of consumer behavior well for five decades, but the last decade has seen the affective and consciousness
107 revolution provide equally important insights. Certainly, consumer prediction based on linguistic symbols is important in the consumer literature, but exploring the processes that generate hedonic experience in real time is equally important in understanding the actual process of consumption and repeat consumption (Pham et al. 2001) Given that Nisbett and Wilson (1977) admonishes that consumers are generally unable to accurately predict their affective preferences through symbolic cognition (e.g., lay theories) studying affective hedonics and non thinking mediated preferences generated from consumption experiences in real time would be an important complement to the symbol based prediction we have inherited f rom the cognitive and linguistic revolution.
108 CHAPTER 6 AFFECT TUNING The twosystem, two affect framework was described in the first four chapters. The first four chapters contain a number of propositions that flow from the framework as theorized. In the current chapter (Chapter 6), I will present five empirical studies that test a portion of the propositions in the framework. The term Affect Tuning refers to the process by which the state affective system changes, in cross over fashion, between the pos itive hedonic affective system (i.e., loosely the opioidergic system) and the negative hedonic affective system (i.e., loosely the acetylcholine system). Affective Influences o n Interpreting Sensory Qualities As Rewarding Or Punishing People in the United States consume approximately 3770 calories per person per day, far in excess of the 2000 to 2500 calories per day recommended by the U.S. Food and Drug Administration (FAO Statistical Yearbook 2005/2006; FDA 2007). Although there are many reasons for exce ss calorie consumption, one prominent cause of excess eating is the appealing taste of food (Mitchell et al. 1998; Sorensen et al. 2003). In fact, consumers identify taste as the most important factor in food choice (Glanz et al. 1998). The importance of t aste is not lost on the worlds $3.2 trillion ($1 trillion in U.S.) processed food industry. In 2008, the U.S. processed food industry spent approximately $15 billion on research to refine the taste of food and $8.1 billion on artificial flavorings. For this and other reasons, it may not be surprising that tasty products comprise a large portion of U.S. food consumption. For example, it has been estimated that onethird of the calories consumed in the United States come from junk food, food that is low in nutritional value but high in taste palatability (Block 2004).
109 The importance of taste in the consumption of food naturally leads to the fundamental question of what makes food taste good. The generally accepted wisdom is that certain foods taste good because of taste engineering (direct effects, i.e., bottom up processes) and consumption contexts or marketing actions (indirect effects, i.e., topdown processes). Whereas the research and development departments of processed food manufacturers have focu sed on perfecting the sensory quality of foods, the marketing departments of these manufacturers have focused on ways of making consumption experiences more enjoyable. Marketers have found that a taste experience can be enhanced by manipulating expectations (e.g., Allen, Gupta, and Monnier 2008), serving containers (e.g., Krishna and Morrin 2008), and consumption settings (e.g., Germov and Williams 1999). In general, these indirect effects on taste perception show how information that is independent of the food product itself can bias the experience. We contend that there are additional factors that can bias taste perception. One such class of factors, individual state factors (e.g., affective states, visceral states, motivational states), are part of the autonomic nervous system responsible for monitoring the internal state of the body. These state factors represent innate systems that have helped humans survive and evolve. A consequence of this evolutionary significance may be that specific states can infl uence the perceptual experience derived from specific sensory systems (Panksepp 1998; Rolls 2005). This influence could range from a simple change in the sensitivity to information from a sensory system (nonspecific sensory amplification) to an increased v igilance for sensations indicating threats or rewards (specific sensory amplification).
110 We use three studies to explore how affective states influence nonspecific and specific sensory amplification of taste experiences. Experiment 1 shows that consumers i n negative and positive affective states are more sensitive to the flavors in food (i.e., can taste more flavors). This nonspecific amplification of a taste experience increases the appreciation for the food and the amount the person is willing to pay for the food. Experiment 2 shows that affective states make consumers more sensitive to specific tastes. Consumers in a negative affective state are more s ensitiv e to spiciness and sourness ( tastes that predict potential danger), whereas consumers in a positiv e affective state are more sensitiv e to savoriness and sweetness ( tastes that predict potential reward). Again, this sensitivity influences the evaluation of the taste experience and the consumers valuation of the product. Experiment 3 shows that affectiv e states create the potential for nonspecific and specific amplification, but that specific amplification takes precedence Taste Perception In his classic work on the taste system, Titchener (1896) explained that most taste sensations can be classified into four basic categories: saltiness sweetness sourness and bitterness When chemicals from food such as salt, sucrose, acids, or quinine come into contact with receptor sites on the tongue, electrical signals generated by the taste cell are transmitted through the chorda tympani, hypoglossal, and glossopharyngeal nerves toward the primary taste cortex, the insula (Goldstein 2007; Pritchard, Macaluso, and Eslinger 1999). An expanded view of taste processing later included umami ( savor iness ), a brothy tast e found in glutamate and protein rich foods (McCabe and Rolls 2007), and spiciness a sensation that is carried by the trigeminal nerves to the brain stem, before reaching the insula (Kadohisa, Rolls, and Verhagen
111 2005; Lim and Green 2007). Lim and Green ( 2007) presented evidence that, despite being mediated by different nerve fibers, bitter and spicy are qualitatively similar. Thus, the five basic taste sensations that are considered fundamental in taste experience are salty, sweet, sour, bitter/spicy, and umami (savory).1 There is a long held assumption that taste perceptions are substance invariant and context dependent. In other words, a substance that is presented in the same context across two occasions should result in an identical taste experience, whereas a substance presented in two different contexts (e.g., substance X after tasting a sweet versus a sour food) should result in a varied taste experience. The substance invariance assumption has, in large part, been based on knowledge about neural pathways between taste receptors and the insula. These neural pathways are direct, fast, and myelinated, as might be expected for a sensory system that is necessary for survival (Goldstein 2007). Myelin is a conductive sheathing that insulates the axon of a neuron. The myelination of neurons in excitatory and inhibitory taste pathways increases temporal resolution, so that the taste sensation is quickly activated and deactivated with stimulus onset and offset, a benefit for identifying potentially rewarding and threatening food. Recent advances in the neurological mapping of sensory experience challenge the invariance assumption for taste perception. A key insight was the discovery of the role of the orbitofrontal cortex, a strip of brain tissue above the ey e sockets, in 1 Technically, spiciness is not a taste but a gustatory or somatosensory response in the oral cavity. Most experts in the field now consider spiciness to be a basic taste response, because it is clearly integrated int o the perception of the foods overall flavor and consumption experience (Rolls 2005), and has a strong influence on food liking or disliking, despite its technical status as a gustatory response. For the sake of brevity, we will use taste to stand for oral gustatory and somatosensory perception in the rest of the article.
112 regulating consumption by changing the taste experience (see Rolls 2008 for review). Pritchard, Schwartz, and Scott (2007) demonstrated that the orbitofrontal cortex, unlike the insula, has neurons that fluctuate in their response to taste se nsations. The orbitofrontal cortex has sensationspecific neurons that represent pleasant sensory experiences, including gratifying taste experiences. When a persons bodily level of a specific nutrient (e.g., energy, fatty acids) becomes sufficient, these sensationspecific neurons become less responsive and, in effect, alter the pleasure of the perceived taste sensation associated with consuming particular foods (i.e., mute the hedonic taste experience). This regulatory process is the autonomic systems guard against over consumption. More importantly, this finding presents evidence that an autonomic state factor (e.g., organismic sufficiency) can influence sensory experience (e.g., the pleasure and taste of food). If the orbitofrontal cortex has satiation based processes that can influence taste perception, then every autonomic state represented in the orbitofrontal cortex has the potential to influence sensory experience. Emerging research in neurobiology (Panksepp 1998) has demonstrated that the subcort ical affective system, responsible for generating emotions such as fear (LaBar and LeDoux 2006), anxiety (Nelson and Panksepp 1998), and positive engagement (Panksepp 1998), projects activations toward the prefrontal cortex, including the orbitofrontal cor tex. Because projections from the affective system are diffuse (Ciompi and Panksepp 2005), and involve neuromodulators that modulate all types of neurons (Panksepp 2008), they should influence the neural responses of the taste representations in the orbitofrontal cortex as
113 well. Hence, we suggest that the consumers affective state should have an influence on taste perception, and should influence it in statedependent ways. Affective System The subcortical affective system is a network of brain tissue (brainstem/basal forebrain, amygdala, and anterior cingulate) that generates affective programs responding to environmental situations in ways that help an animal survive. The most fundamental forms of affect are positive and negative affective states.2 A po sitive affective state is an upwardregulating stateenergetic program that is activated when an animal is free of physical pain, has adequate metabolic or regulatory energy, and has abundant social support (Panksepp 1998). In humans, positive affective st ates generate a subjective feeling of well being and excitement, which in turn activates forward locomotion in the organism (e.g., exploration) (Barbano and Cador 2007; Panksepp 1998; Sacks et al. 1972). A negative affective state is a downwardregulating state energetic program that is activated when an animal is physically injured, has depleted its metabolic or regulatory energy stores, or has lost social support (Nettle, forthcoming). In humans, a negative affective state generates a subjective feeling of distress, which in turn activates fleeing locomotion in the organism (e.g., escape) or freezing (Panksepp 2008). Evolutionary theories of affective system emergence (e.g., Panksepp, Nelson, and Bekkedal 1997; Rolls 2005) propose that affective circuits developed because they are beneficial to individual level and species level survival. A common theme across these 2 Although the prefrontal cortex (responsible for higher order cognition) can modify affective programs into more nuanced emotions such as guilt and other types of discrete emoti ons (Panksepp 1998), the emerging consensus is that positive and negative engagement are neurally fundamental or what Lang and Bradley (2008) call neural substrates (Panksepp uses the term blue ribbon affective programs to refer to affective programs that does not require mediation from higher order cortices).
114 theories is that affective states serve the two oftenconflicting functions of ensuring survival of the individual animal and expanding the reach of the species (Fredrickson 2001; Rolls 2005; Schiller et al. 2008). Of particular interest to our investigation is the role of affective states in directing attention and controlling information acquisition and synthesis (i.e., perception). For exampl e, negative affective states are instrumental in the development of tactile perception. Negative affective states emerged from more primitive limbic circuits that respond to physical pain, such as body tissue injury (Eisenberger, Lieberman, and Williams 2003; MacDonald and Kingsbury 2006; Panksepp 2003). An understanding of the relationship between tactile sensation, pain, and threat is fundamental to the survival of an animal. In contrast, positive affective states are instrumental in the development of visual perception (Imanaka et al. 2008; Mora Ferrer and Gangluff 2000; Rolls 2005). Positive affect indicates that the organism is self sufficient and should search its environment for rewards and reward cues. The relationship between positive affective stat es and vision exists because evolving animals needed to expand their territory in order to have sufficient natural resources (e.g., food, water) for population expansion. An evolutionary perspective suggests that affect is also involved in the perception of incentive salience, such as taste perception and other reward cues (Pecina, Schulkin, and Berridge 2006). Panksepp (1998) showed that organisms experiencing positive affect are also more sensitive to positive reward cues. Pecina et al. (2006) showed t hat organisms experiencing negative affect are, paradoxically, also more sensitive to positive reward cues. Systems under duress (negative affective state) or anticipating exploration (positive affective state) require more nutrition than a system in hiber nation
115 or homeostasis (neutral affective state). Thus, the subcortical affective system, in concert with orbitofrontal cortex, should have a mechanism for inducing more food consumption in an affective state than would be necessary in a neutral state. In effect, the orbitofrontal cortex has to be able to do the opposite of the taste muting that accompanies satiation and neural habituation (cf. Rolls, Sienkiewicz, and Yaxley 1989). One way for an autonomic system to encourage consumption would be to enhance the taste perception of food when the animal is in an affective state. If the autonomic system could make edible foods taste better, the animal should consume more and be more successful at recovering from duress or succeeding at exploration. This leads to the hypothesis that an affective state amplifies a taste experience: H1 Consumers experiencing a negative or positive affective state should have a heightened state of taste perception. This heightened state of taste perception should a allow the consumer to taste more flavors; b enhance the palatability of the food; c. increase the desirability (value) of the food. Affective states must do more than simply encourage the consumption of any food. Affective states must encourage the consumption of appropriate f oods and discourage the consumption of inappropriate foods. These priorities depend on the state of the system. When an animal is in a negative affective state, it should prioritize the avoidance of foods that can cause (further) injury or death, outcomes that are more easily reached when an animal is already under duress (Nettle, forthcoming). In this state, the animal needs to avoid food that might cause further harm or death. Negative affective states should increase sensitivity to the threat indicators of bitter, spicy and
116 sour because these flavors probabilistically predict poisonous vegetation (bitter or spicy) and foods with fungal and/or bacterial growth (sour). When an animal is in a positive affective state, its body is preparing for exploration and territorial expansion (Fredrickson 2004). In this state, the animal needs to consume foods that are high in calories and protein to store energy reserves needed for broadening and building its resources (Fredrickson 2001). Thus, positive affective states should increase sensitivity to the rewarding flavors of sweet and savory because these flavors predict calories and protein. In summary, negative and positive states should selectively amplify flavors that support statedependent survival goals. H 2 Consu mers experiencing a negative affective state should a be more sensitive to bitter/spicy and sour flavors; b dislike foods that contain bitter/spicy and sour flavors; c avoid (devalue) foods that contain bitter/spicy and sour flavors. H 3 Consumers experiencing a positive affective state should a be more sensitive to sweet and savory flavors; b appreciate foods that contain sweet and savory flavors; c desire (hypervalue) foods that contain sweet and savory flavors. Experiment 1 Experiment 1 tested the idea that af fective states amplify the taste intensities of food and allow consumers to taste a greater variety of flavors (hypotheses H1). Consumers were asked to experience a product ( potato chips ) while in a negative neutral, or positive affective state. Afterward s, participants were asked to report and value their product experience. It was expected that participants in positive and negative
117 (but not neutral) affective states would perceive more flavors in the potato chips, appreciate the product to a greater extent, and, consequently, value the product more Design and Procedure Design One hundred eighteen undergraduate students participated in the study for extra credit. The study used a betweensubject manipulation of the affective state (negative, neutral, pos itive). Participants were put into an affective state, tried a product (potato chips), recorded their perceptions of the product, assessed the palatability of the product, and indicated the amount they were willing to pay for the product. Procedure Part icipants were invited into a behavioral lab and told that the study investigated, how consumers experience products under different emotional states. It was explained that people in certain professions (health care, military) often experienced temporary and/or prolonged affective states and that these moods could influence their appreciation for products. Participants were told, your task is to help us design products by sincerely trying your best to feel the emotion that is suggested by the statements y ou are about to read. Participants were then put into a negative, neutral, or positive affective state using the Velten (1968) affect induction procedure. The Velten procedure involves reading a series of statements that get progressively more negative (negative affect condition), progressively more positive (positive affect condition), or remain consistently neutral (control condition). Participants were told to read the statements into a microphone, and to take their time imagining how each statement applied to their lives. For example, the negative affect manipulation starts with mildly negative statements
118 such as I feel a little bit low today and ends with strongly negative statements such as All of the unhappiness of my past life is taking posse ssion of me. The positive affect manipulation starts with mildly positive statements such as I feel light hearted and ends with strongly positive statements such as God, I feel great! The control condition starts with affectively neutral statements su ch as Oklahoma City is the largest city in the world in area, with 631.166 square miles and ends with affectively neutral statements such as At low tide the hulk of the old ship could be seen. After the affect induction, subjects were asked to taste a potato chip. The potato chips were Pringles regular potato chips made by the Procter & Gamble Company (Cincinnati, OH). A potato chip was placed in an 8ounce dish, covered with a table napkin, and positioned on the upper left corner of the cubicle befor e the participants arrived in the laboratory. Each bowl contained one potato chip. After trying the potato chips, participants were asked to try to identify as many flavors/ingredients as you can in the chips and were provided with a sheet of paper containing 20 lines. Participants recorded their impressions of the potato chip on the lines provided on the form. Next, participants were asked to rate the taste of the chips (e.g., 0 = not flavorful, 10 = flavorful). Finally, participants indicated thei r willingness to pay for the chips using the question, Assuming that you need it for a big party, what is a reasonable price you would be willing to pay for a large, party size package (10.5 ounces) of the product? Coding The responses to the openended flavor were coded by counting every flavor that was listed.
119 Results Manipulation check The Velten (1968) procedure has been extensively validated. For example, Brown et al. (1993) showed that the positive and negative affect induction procedures induced physiological changes (hormone levels and other endocrine concomitants, heart rate) consistent with positive and negative states. Similarly, Gadea et al. (2005) showed that the negative affect induction procedure increased cortisol and CRF levels that medi ate negative affect. P rimary Analysis The mean responses are presented in table 1. The key predictions involved the perceived number of flavors, palatability ratings, and willingness to pay dependent measures. The influence of affective state on the number of perceived number of flavors was significant ( F (2, 115) = 3.15, p < .05). Participants experiencing negative affect ( M = 1.42; F (1, 115) = 5.20, p < .0 5) and positive affect ( M = 1.38; F (1, 115) = 4.17, p < .0 5) listed more flavors than participants ex periencing neutral affect ( M = 1.02) The influence of affective state on the perceived palatability rating was significant ( F (2, 115) = 5.98, p < .05). Participants experiencing negative affect ( M = 6.68; F (1, 115) = 6.82, p < .05) and positive affect ( M = 7.03; F (1, 115) = 10.64, p < .05) perceived the chips to be more flavorful than participants experiencing neutral affect ( M = 5.23). The influence of affective state on the amount the participant was willing to pay for the chips was significant ( F (2, 115) = 4.67, p < .05). Participants experiencing negative affect ( M = $3.57; F (1, 115) = 5.12, p < .0 5) and positive affect (M = $3.78; F (1, 115) = 8.43, p < .0 5) were willing to pay more for the product than participants experiencing neutral affect (M = $2.77)
120 Table 61. The influence of affective state on the consumers ability to perceive flavors, palatability of the food, and the consumers willingness to pay. Dependent Measure Affective State # Flavors Palatability WTP ($) Negative 1 42 a 6.68 a $ 3.57 a Neutral 1 02 b 5.23 b $ 2.77 b Positive 1 38 a 7.03 a $ 3.78 a significantly different at p < .05. Mediation analysis for negative affect consumers The Baron and Kenny (1986) threestep procedure was used to assess whether the palatability ratings mediated the i nfluence of negative affect on the willingness to pay for the product. The analysis used participants in the negative and neutral affective states. First, the affective state was a significant predictor of the participants willingness to pay for the product ( = .80, SE =.34, t (76) = 2.34, p < .05). Second, affective state was a significant predictor of the palatability of the product ( = 1.46, SE =.60, t (76) = 2.44, p < .05). Third, when the willingness to pay measure was regressed on the affective s tate and the palatability of the product the affective state coefficient became nonsignificant, ( = .281, SE = .281, t ( 75) = 1.0, p = 32) and the palatability coefficient remained significant ( = .36, SE = .05, t ( 75) = 6.88, p < .01; Sobel z = 2.32 p = .02). Mediation analysis for posi tive affect consumers The Baron and Kenny (1986) threestep procedure was used to assess whether the palatability of the product mediated the influence of positive affect on the amount participants were willing to pay for the product. The analysis used participants in the positive and neutral affective states. First, the affective state was a significant predictor of the participants willingness to pay for the product ( = 1.02, SE =.34, t (78) = 3.02, p < .05). Second, affective state was a significant predictor of the palatability of the
121 product ( = 1.80, SE = .49, t (78) = 3.65, p < .05). Third, when the willingness to pay measure was regressed on the affective state and the palatability of the product the affective state coefficient became nonsignificant, ( = .32 SE = 30 t ( 77) = 1.06, p = .29) and the perceived number of flavors coefficient remained significant ( = 39, SE = 06, t ( 77) = 6.07, p < .0 5 ; Sobel z = 3.25, p = 001). Discussion Experiment 1 demonstrated that consumers in positive and negat ive affective states generated more complex perceptual experiences than consumers in a neutral affective state. These perceptions led to a greater appreciation for the flavor of the product. In turn, these experiences influenced the amount consumers were w illing to pay for the product. This response pattern provides preliminary evidence in support of hypothesis 1. Hypotheses 2 and 3 propose that affective states not only increase the signal to noise ratio in sensory cortices, but that affective states ampli fy qualitatively different kinds of sensation ( selective amplification hypothesis ). That is, people in an affective state may be more sensitive to specific taste sensations. As argued earlier, natural selection (Darwin 1859) assumes that positive and negat ive affect states are adaptive (i.e., affective states emerged because they support survival) (Lang and Bradley 2008). Animals experiencing positive affect, an indicator of a state of sufficiency, are induced to build physical resources for the future (Fredrickson 2004) and, thus, will emphasize the storage of calories. In contrast, animals experiencing negative affect, an indicator of state illbeing (e.g., weakened immune system), should emphasize perceptual vigilance against vegetative poisons, spoiled f ood, and caustic substances.
122 Experiment 2 Experiment 2 assessed whether affective states amplif y specific taste perceptions. Consumers were asked to taste a simple product or a complex product. The simple product was the same brand of potato chip from ex periment 1. The complex product was the same chip with additional spicy, sour, sweet, savory, and salty flavors. Ground red chili peppers/c apsaicin powder and ground black pepper were added to increase spiciness, sour cream powder was added to increase sourness, buffalo wild wings powder, and monosodium glutamate were added to increase sweetness and savoriness. If affective states result in specific sensory amplification, taste perceptions should be guided by statedependent flavor responses. That is, parti cipants in a negative (positive) affective state should taste more of the punishing (rewarding) favors and respond with lower (higher) palatability ratings and lower (higher) product valuations. Design and Procedure Design Two hundred and forty one undergraduate students participated in the study for course credit. The study used a 3 x 2 betweensubjects design, in which there was a betweensubject manipulation of the affective state (negative, neutral, positive) and a betweensubject manipulation of stimulus complexity (simple, complex) of the product. Stimuli The simple potato chips were Pringles regular potato chips. The complex potato chip was created by sprinkling approximately 250 milligrams of capsaicin powder, pepper, sour cream powder buffalo w ild wings powder, and monosodium glutamate on each chip. Rolls (2005) identified spiciness and sourness as potentially
123 punishing flavors. In support of this claim, Verhagen, Kadohisa, and Rolls (2004) showed that capsaicin, the active ingredient in chili peppers that originally evolved to prevent birds from eating the fruit, causes a burning sensation (see also Rolls, Verhagen, and Kadohisa 2003). Sour tastes warn against spoiled foods and unripe fruits (Lindemann 2001; Meyerhof 2008). Panksepp (1998) and M cCabe and Rolls (2007) have identified savory and sweet as rewarding tastes. Savory sensations predict foods high in glutamate and protein, necessary for tissue growth and repair, whereas sweet sensations predict foods high in energy, necessary for metabol ism. Procedure The procedure was identical to that used in experiment 1 except for the dependent measures. Instead of listing flavors, participants judged how spicy, sour, sweet, savory, and salty the product was on an 11point scale (e.g., 0 = not spicy at all and 10 = very spicy). The participants then estimated how much they would be willing to pay for a large, party size package of the product. Results The mean responses are presented in table 2. The key predictions involved the perceived spiciness, sourness, sweetness, savoriness, and the willingness to pay measures. The first analysis focused on the punishing flavors. The affective state by stimulus complexity interaction was significant for spiciness ( F (2, 235) = 5.52, p < .01). Ratings of spic iness did not significantly vary for the simple stimulus ( F (2, 235) = .89, p > .10), but did significantly vary for the complex stimulus ( F (2, 235) = 18.17, p < .01). For complex stimuli, spiciness ratings were higher in the negative affect condition ( M = 7.27) than in the neutral ( M = 5.49; F (1, 237) = 18.06, p < .01) or positive ( M = 4.84; F (1, 237) = 34.68, p < .01) affect conditions. The affective state by stimulus complexity
124 interaction was significant for sourness ( F (2, 235) = 5.56, p < .01). Ratings of sourness did not significantly vary for the simple stimulus ( F (2, 235) = 1.75, p > .10), but did significantly vary for the complex stimulus ( F (2, 235) = 10.90, p < .01). For complex stimuli, sourness ratings were higher in the negative affect condition ( M = 5.57) than in the neutral ( M = 3.61; F (1, 237) = 15.02, p < .01) or positive ( M = 3.46; F (1, 237) = 18.07, p < .01) affect conditions. The second analysis focused on the rewarding flavors. The affective state by stimulus complexity interaction was s ignificant for savoriness ( F (2, 235) = 4.10, p < .05). Ratings of savoriness did not significantly vary for the simple stimulus ( F (2, 235) = 2.82, p > .05), but did significantly vary for the complex stimulus ( F (2, 235) = 16.49, p < .01). For complex stimuli, savoriness ratings were higher in the positive affect condition ( M = 7.68) than in the neutral ( M = 5.29; F (1, 237) = 15.37, p < .01) or negative ( M = 4.24; F (1, 237) = 30.16, p < .01) affect conditions. The affective state by stimulus complexity inter action was not significant for sweetness ( F (2, 235) = .63, p > .10). Ratings of sweetness did vary significantly by level of affect ( F (2, 235) = 12.58, p > .10). Sweetness ratings were higher in the positive affect condition ( M = 4.16) than in the neutral ( M = 2.97; F (1, 237) = 11.40, p < .01) or positive ( M = 2.31; F (1, 237) = 23.82, p < .01) affect conditions. The third analysis focused on the amount participants were willing to pay for the product. The affective state by stimulus complexity interaction was significant for the willingness to pay measure ( F (2, 235) = 15.39, p < .05). Replicating experiment 1, willingness to pay significantly varied for the simple stimuli ( F (2, 235) = 5.38, p < .05). Willingness to pay was higher in the negative ( M = $3.46; F (1, 237) = 6.00, p < .05) and
125 positive ( M = $3.49; F (1, 237) = 6.93, p < .01) affect conditions than in the neutral affect condition ( M = $2.71). Willingness to pay also significantly varied for the complex stimuli ( F (2, 235) = 27.85, p < .05). Willingness to pay was lower in the negative ( M = $1.90; F (1, 237) = 13.11, p < .05) and higher in the positive ( M = $3.93; F (1, 237) = 14.00, p < .01) affect conditions than in the neutral affect condition ( M = $2.92). A final analysis confirmed that perceptions of product saltiness did not exhibit an affective state by stimulus complexity interaction ( F (2, 235) = .78, p > .10). Table 62. The influence of affective state on the ability to perceive specific tastes and the willingness to pay. Dependent Measure Chip Type Spicy Sour Savory Sweet Salty WTP ($) Simple Chip Negative State 1.17 1.31 5.58 2.14 7.11 $3.46 Neutral State 0.68 1.69 5.05 3.17 6.67 $2.71 Positive State 0.68 0.78 6.49 4.12 6.59 $3.49 Complex Chip Negative State 7.27 5.57 4.24 2.49 7.05 $1.90 Neutral State 5.49 3.61 5.29 2.68 5.76 $2.92 Positive State 4.84 3.46 7.68 4.21 6.14 $3.93 Mediation analysis for negative affect consumers and complex chips The Baron and Kenny (1986) threestep procedure was used to asse ss whether the ratings of spiciness and sourness mediated the influence of negative affect on consumers willingness to pay for the complex chips. The analysis used participants who tasted high complexity chips in the negative and neutral affective states. First, the affective state was a significant predictor of the willingness to pay for the chips ( =1.02, SE = .26, t (76) = 3.98, p < .01). Second, the affective state was a significant predictor of the spiciness ( = 1.78, SE = .45, t (76) = 3.93, p < .0 1) and sourness ( = 1.96, SE =
126 .62, t (76) = 3.15, p < .01) of the chip. Third, when the willingness to pay measure was regressed on the affective state and the spicy taste ratings the affective state coefficient became less significant, ( = .79, SE = 28, t ( 75) = 2.86, p > .05) and the spicy taste rating coefficient remained significant ( = 13 SE = 06 t ( 75) = 2.07, p < .0 5 ; Sobel z = 2.56, p = .0 1 ). Similarly when the willingness to pay measure was regressed on the affective state and the sour taste ratings the affective state coefficient became less significant, ( = .82, SE = 278, t ( 75) = 3.15 p > .05) and the spicy taste rating coefficient remained significant ( = 09 SE = 05 t ( 75) = 1.97, p = .0 5 ; Sobel z = 2.19, p = .0 3 ). These res ults indicate that perceptions of spiciness and sourness each partially mediated the lower willingness to pay in the negative affect condition. Mediation analysis for positive affect consumers and complex chips The Baron and Kenny (1986) threestep procedure was used to assess whether the ratings of savoriness and sweetness mediated the influence of positive affect on consumers willingness to pay for the chips. The analysis used participants in the positive and neutral affective states. First, the affectiv e state was a significant predictor of the taste rating of the chips ( = 1.01, SE = .26, t (83) = 3.86, p < .001). Second, the affective state was a significant predictor of the savoriness ( = 2.39, SE = .58, t (83) = 4.09, p < .01) and sourness ( = 1.52 SE = .48, t (83) = 3.20, p < .01) of the chip. Third, when the willingness to pay measure was regressed on the affective state and the savory taste ratings the affective state coefficient became less significant, ( = .72, SE = 28, t ( 82) = 2.59, p < .05 ) and the savory taste rating coefficient remained significant ( = 12, SE = 05, t ( 82 ) = 2.57, p < .0 5 ; Sobel z = 2.80, p = .0 1 ). Similarly when the willingness to pay measure was regressed on the affective state and the sweet taste ratings the affective state coefficient became less significant, ( = .71, SE = 26, t ( 82) =
127 2.74, p < .05 ) and the spicy taste rating coefficient remained significant ( = 20 SE = 06, t ( 82) = 3.49, p < .0 1 ; Sobel z = 2.60, p = .0 1 ) These results indicate that perceptions of savoriness and sweetness each partially mediated the higher willingness to pay in the positive affect condition. Ancillary mediation analyses Additional mediation analyses were performed to confirm that (1) savo riness and sweetness did not mediate the influence of affect on willingness to pay for the complex chip in the negative versus neutral affect condition, (2) spiciness and sourness did not mediate the influence of affect on willingness to pay for the comple x chip in the positive versus neutral affect condition, and that (3) spiciness, sourness, savoriness, and sweetness did not mediate the influence of affect on willingness to pay for the simple chip in the negative or positive versus neutral affect condition. In all of these mediation analyses, affect failed to influence the mediator, the mediator failed to influence the willingness to pay, or the Sobel test failed to reach significance. Discussion Experiment 2 showed that affective states have qualitati vely different influences on taste perception: negative affective states amplify perceptions of the punishment cues of sourness and spiciness (even in a safe food), whereas positive affective states amplify perceptions of the reward cues of savoriness and sweetness. These results are consistent with the selective amplification hypothesis (hypotheses 2 and 3). Consumers in a negative affective state (as a result of illness or ill being) should be more sensitive to punishing taste experiences because their undermined organismic status cannot afford to take risks (Nettle, forthcoming). In contrast, consumers in a positive affective state should be more sensitive to rewarding taste experiences because their self -
128 sufficient organismic status induces maximizing energy storage in preparation for exploration (Fredrickson 2001). E xperiment 3 Experiments 1 and 2 provide an interesting set of results. Experiment 1 showed that negative and positive affective states amplify sensitivity to taste in general, whereas ex periment 2 showed that this amplification can be specific to certain types of taste experiences. In other words, affective states encourage both nonspecific and specific amplification of taste perceptions. In experiment 3, we assessed whether nonspecific a nd specific amplification of taste could occur in tandem. The experimental design included a simple potato chip and a complex potato chip that was spicy and sour. When the chip was simple, we expected nonspecific amplification, as was illustrated in the palatability rating and willingness to pay judgment of experiment 1. When the chip was spicy and sour, we expected specific amplification in the negative state condition, but nonspecific amplification in the positive state condition. The negative state should amplify the spicy and sour flavors, which should result in less palatability and a lower willingness to pay for the product. In contrast, t he positive state should amplify flavor in general, which should result in greater palatability and a higher wil lingness to pay for the product. Thus, when there is an opportunity for nonspecific and specific amplification, specific amplification takes precedence. We note that a savory/sweet chip condition was not included in the experimental design because the negative state should show a beneficial effect of nonspecific amplification and the positive state should show a beneficial effect of specific amplification, hence, the manipulation was nondiagnostic.
129 Design and Procedure Design One hundred and seventy two undergraduate students participated in the study for extra credit. The study used a 3 x 2 betweensubjects design, in which there was a betweensubject manipulation of the affective state (negative, neutral, positive) and a betweensubject manipulation of st imulus complexity. One potato chip was simple and the second potato chip was complex as a consequence of adding potentially punishing flavors. Stimuli The punishing flavors were the same that were used in experiment 2. Thus, the complex punishing chip was created by sprinkling approximately 250 milligrams of capsaicin powder, pepper, and sour cream powder on each chip. Procedure The procedure was identical to the procedure used in experiment 1. The experiment 1 procedure was used because it asks participants to list flavors, as opposed to rate specific flavors (as in experiment 2). It was felt that the nonspecific amplification of taste perception would be easier to observe if participants were not focused on specific flavors in the complex stimulus conditi on. Results Primary analysis The mean responses are presented in table 3. The key predictions involved the perceived number of flavors, the palatability rating of the chips, and willingness to pay dependent measures. The affective state by stimulus complex ity interaction was significant for the perceived number of flavors ( F (2, 166) = 3.71, p < .05). In the simple
130 stimulus condition, the affect manipulation was significant ( F (2, 166) = 2.93, p = .05). Negative state participants ( M = 3.24; F (1, 168) = 3.96, p = .05) and positive state participants ( M = 3.12; F (1, 168) = 3.17, p = .08) listed more flavors than neutral state participants ( M = 2.61). In the complex stimulus condition, the affect manipulation was significant ( F (2, 166) = 20.11, p < .05). Negati ve state participants ( M = 4.40; F (1, 168) = 28.89, p < .01) and positive state participants ( M = 4.49; F (1, 168) = 30.22, p < .01) listed more flavors than neutral state participants ( M = 2.86). The affective state by stimulus complexity interaction was s ignificant for the palatability ratings ( F (2, 166) = 6.70, p < .01). In the simple stimulus condition, the affect manipulation was significant ( F (2, 166) = 5.85, p < .01). Negative state participants ( M = 6.76; F (1, 168) = 6.45, p < .05) and positive state participants ( M = 7.00; F (1, 168) = 8.84, p < .01) rated the chips more palatable than neutral state participants ( M = 5.50). In the complex stimulus condition, the affect manipulation was significant ( F (2, 166) = 12.21, p < .01). Negative state participants ( M = 5.07) rated the chips less palatable than neutral state participants ( M = 5.93; F (1, 168) = 3.38, p = .07) and positive state participants ( M = 7.31) rated the chips more palatable than neutral state participants ( M = 5.93; F (1, 168) = 8.45, p < .01). The affective state by stimulus complexity interaction was significant for the willingness to pay judgments ( F (2, 166) = 6.98, p < .01). In the simple stimulus condition, the affect manipulation was significant ( F (2, 166) = 7.05, p < .01). Negative state participants ( M = $3.35; F (1, 168) = 6.26, p < .05) and positive state participants
131 ( M = $3.61; F (1, 168) = 11.13, p < .01) judged the chips to be more valuable than neutral state participants ( M = $2.62). In the complex stimulus condition, the affect manipulation was significant ( F (2, 166) = 15.16, p < .01). Negative state participants ( M = $2.14) valued the chips less than neutral state participants ( M = $2.67; F (1, 168) = 3.69, p = .06) and positive state participants ( M = $3.59) valued the chips more than neutral state participants ( M = $2.67; F (1, 168) = 10.98, p < .01). Table 63. The influence of affective state on the ability to perceive tastes, palatability, and willingness to pay. Dependent Measure Affective State Total Flavors Spicy/Sour Flavors Palatability WTP ($) Simple Chip Negative 3.24 .069 6.76 $ 3.35 Neutral 2.61 .036 5.50 $ 2.62 Positive 3.12 .037 7.00 $ 3.61 Complex Chip (Spicy/Sour) Negative 4.40 2.03 5.07 $ 2.14 Neutral 2.86 1.24 5.93 $ 2.67 Positive 4.49 1.31 7.31 $ 3.59 Supplemental Analysis It was hypothesized that people in a negative affective state should be sensitive to the punishing flavors in the complex stimulus, but that people in the neutral and positive affective state should not. The flavor listing was coded for mentions of flavors associated with the taste dimension of spicy and sour (e.g., spicy, sour, pepper, sour cream, chili). The mean results are shown in table 3. In the complex stimulus condition, the affect manipulation significantly influenced the report of these flavors ( F (2, 166) = 11.41, p < .01). Negative state participants ( M = 2.03) listed more spicy/sour flavors than neutral state participants ( M = 1.24; F (1, 168) = 21.33, p < .05) or positive state participants ( M = 1.31; F (1, 168) = 17.32, p < .05).
132 Mediation analysis for negative affect consumers experiencing a simple product The mediation analysis used participants who tasted simple chips in the negative and neutral affective states. First, the affective state was a significant predictor of the participants willingness to pay for the product ( = .73, SE = .28, t (55) = 2.63, p < .05). Second, the affective state was a significant predictor of the palatability rating of the product ( = 1.26, SE = .53, t (55) = 2. 38, p < .05). Third, when the willingness to pay measure was regressed on the affective state and the palatability ratings the affective state coefficient became nonsignificant, ( = .37, SE = 25, t ( 54) = 1.50 p > .05) and the palatability rating coefficient remained significant ( = 29, SE = 06, t ( 54) = 4.72, p < .0 5 ; Sobel z = 2.18 p = .0 3 ). These results replicate experiment 1. Mediation analysis for positive affect consumers experiencing a simple product The mediation analysis used participants who tasted simple chips in the positive and neutral affective states. First, affective state was a significant predictor of the participants willingness to pay for the product ( = .994, SE = .306, t (53) = 3.24, p < .05). Second, affective state was a significant predictor of the palatability rating of the product ( = 1.50, SE = .496, t (53) = 3.03, p < .05). Third, when the willingness to pay measure was regressed on the affectiv e state and the palatability ratings the affective state coefficient became less significant, ( = .753, SE = 32, t ( 52) = 2.32, p < .05 ) and the palatability rating coefficient remained significant ( = 161, SE = 08, t ( 52) = 1.98 p = .0 5 ; Sobel z = 2. 11, p = .0 4 ). These results replicate experiment 1. Mediation analysis for negative affect consumers experiencing a high complexity product The mediation analysis used participants who tasted the complex punishing chips in the negative and neutral affectiv e states. First, affective state was a significant
133 predictor of the participants willingness to pay for the product ( = .53, SE = .26, t (57) = 2.49, p < .05). Second, affective state was a significant predictor of the palatability rating of the product ( = .86, SE = .45, t (57) = 1.91, p < .06). Third, when the willingness to pay measure was regressed on the affective state and the palatability ratings the affective state coefficient became nonsignificant, ( = .29, SE = 23, t ( 56) = 1.22, p > .05) and the palatability rating coefficient remained significant ( = 28, SE = 07, t ( 56) = 4.25, p < .0 5 ; Sobel z = 1.77, p = .0 8 ). These results indicate full mediation. Mediation analysis for positive affect consumers experiencing a high complexity product The mediation analysis used participants who tasted the complex punishing chips in the positive and neutral affective states. First, affective state was a significant predictor of the participants willingness to pay for the product ( = .92, SE = .28, t (56) = 3.36, p < .05). Second, affective state was a significant predictor of the taste rating of the product ( = 1.37, SE = .40, t (56) = 3.47, p < .05). Third, when the willingness to pay measure was regressed on the affective state and the palatability ratings the affective state coefficient became less significant, ( = .66, SE = 29, t ( 55) = 2.24, p > .05) and the taste rating coefficient remained significant ( = 19, SE = 09, t ( 55) = 2.14, p < .0 5 ; Sobel z = 2.39 p = .0 2 ). These results indicate partial mediation. Discussion Experiment 3 provides evidence that affective states can encourage nonspecific and specific amplification of taste perceptions. Replicating experiment 1, simple stimuli were perceived to have more flavors when a participant was in a negative or positive affective state. This is evidence that affective states encourage nonspecific amplification of taste perceptions. The data also show that participants in a negative state were
134 sensitive to the punishing flavors in a more complex stim ulus (evidence for specific amplification), whereas were participants in a positive state were insensitive to the punishing flavors, but more sensitive to flavors in general (evidence for nonspecific amplification). Thus, specific amplification takes prece dence over nonspecific amplification, as was the case in the negative affective state condition, but that nonspecific amplification can still exert an influence if flavors relevant to other affective states are present, as was the case in the positive affective state condition. G eneral D iscussion The three experiments show that affective states induce nonspecific and specific amplification of taste experiences. First, taste experience was enhanced when participants were in a negative or positive affective state (experiment 1). Participants perceived a taste experience to be more complex, more satisfying, and more valuable when in an affective state. Second, affective states made participants more sensitive to specific taste experiences. Negative affect made participants more sensitive to sourness and spiciness, whereas positive affect made participants more sensitive to sweetness and savoriness (experiment 2). This increase in sensitivity to punishers and rewards influenced the valuation of the product in a perceptually assimilative direction. Third, when nonspecific sensory amplification and specific sensory amplification competed to influence taste perception, as was the case when punishing flavors were experienced in a negative affective state, the specif ic amplification takes precedence (experiment 3). This finding is consistent with the claim that specific amplification exists to heighten vigilance against the consumption of life threatening foods in a negative affective state.
135 The results provide insights to a paradoxical preference reversal among consumers. Consumers appear to like spicy and sour tastes when in a positive affective state, but appear to dislike the same sensory perceptions when in a negative affective state (experiment 3). This implies that consumers are eager to experience sensations and learn (both rewarding and punishing) when their affective systems indicate organismic self sufficiency, but are unwilling to experience punishing sensations when their affective systems indicate organis mic vulnerability. A marketing implication is that marketers of comfort food, which consumers buy to soothe themselves, should have more success with basic flavors that are sweet or savoury. In contrast, marketers of celebratory food (snacks that tend to be bought for parties), which consumers buy for perceptual stimulation, can be more willing to explore new, punishing flavors (e.g., Pringles Extreme Pickle Dill or Doritos Blazin Chips). At the fundamental level, this research also provides sugges tive insights on the interesting phenomenon of negative consumption, in which consumers subject themselves to stimuli that, at first glance, appear intrinsically negative or unpleasant (Andrade and Cohen 2007). For example, consumers appear to like horror movies or watch TV show s like Fear Factor, a program that shows aversive stimulation such as people eating worms and the like. The current explanation in the consumer literature is that these negative consumptions are typically accompanied by a detached protective frame, a cognitive state of mind in which the consumer experiences negative feelings but is well aware that he or she is completely protected from the stimuli (Andrade and Cohen 2007, 28687). This article contributes to the literature by showing conditions under which the consumers affective state can itself be a protective frame by increasing
136 the organisms perceived ability to deal with the negative stimulus. The response patterns show that consumer response to negative stimulation such as spiciness and sourness can become preferred when the consumer is in a positive affective state, provided there are other flavors concurrently available. Sensory stimulation becomes pleasurable when the consumer perceives organismic sufficiency, because the pleasure from nonspecifi c information and learning from complexity may outweigh any discomfort from the burning or biting sensation. This pattern may also apply to the consumption of nonfood products such as movies and TV programs, in which consumers would prefer to watch horror movies or disgusting shows when they are in a celebratory, positive affective state, but would not watch the same programs when they are having a bad day. Hence, consumer preference or aversion toward a negative stimulus is not just a function of the valence of the stimulus, but is also a function of the consumers perceived organismic sufficiency (How strong am I right now?), whether this sufficiency is instantiated through an affective variable (affect state) or a cognitive variable (protective frame). The results also provide a novel refinement to our understanding of mood repair. Previous research suggests that mood repair is a function of Cyrenaic/Epicurean hedonistic goals, in which a negative mood leads the consumer to eat foods in general i n order to bring the mood to an acceptable level (Labroo and Mukhopadhyay 2009). This article, however, shows that stimulus hypervaluation occurs in negative affective states for stimuli that are sensory predictive of calories or protein (and sensory nonpr edictive of acids and capsaicin), perhaps because these sensations indicate a stimulus that could restore the organisms damaged state (experiment 1), such that
137 consumers under negative affective state display hypervaluation toward these products (through higher willingness to pay). Emerging neuroscience findings suggest that the affective circuits responsible for negative affective states is also linked to the reward seeking system, such that activation of the negative affective system actually induces consumption (Pecina, Schulkin and Berridge 2006). This article goes a step further by showing that stimulus devaluation can also occur for products that are seen as punishing (experiment 2 and 3). Thus, in order to understand mood repair, the sensory nature of the stimulus is crucial for predicting whether the consumer will approach or avoid the product. The idea that consumption can be rewarding or punishing depending on the consumers affective state and the nature of stimulus may have implications for other types of consumer products. Emerging findings in neuroscience suggest that higher level cognitive experiences are affectively rooted in lower level perceptual experiences (Eisenberger, Lieberman, and Williams 2003; Panksepp 2003; Rolls 2005), and preliminary speculations indicate that pleasure can be generated from information or meaning alone (Loewenstein 2006). The process of generating meaning may have become pleasurable over time because human evolution required the use of various systems of meaning to predict outcomes in an uncertain environment. Humans that could process more complex semantic meaning predicted outcomes more reliably, survived, and passed on their genes. In support of this idea, recent neurobiological findings show that phasic bursts of pleasure in a monkey increase as meaning is generated from a situation with higher (as
138 opposed to lower) uncertainty, even if the resulting reward is equally desirable in absolute magnitude (Fiorillo, Tobler, and Schultz 2003). The relative influence of the two types of pleasure (semantic versus sensory) may depend on the expertise of the consumer. For example, a novice consumer (e.g., a child or a teen) may find semantic rich but sensory poor movies and music nonpleasurable, because the novice consum ers semantic networks do not have the complex connections that are projected toward opioidergic pathways. This novice consumer may instead prefer to watch a high definition movie with rich colors and sound. In contrast, an expert consumer (e.g., a history buff or movie connoisseurs in general) may prefer to watch old, meaningful films (e.g., Disraeli or The Life of Emile Zola) rather than new, less meaningful movies with high sensory quality (e.g., Fast and Furious or Monsters vs Aliens 3D). The nonintuitive finding that a higher level of perceptual complexity results in lower liking among consumers in negative affect (experiments 2 and 3) may shed light into the enjoyment of semantic products. Consider the consumers choice between two movies in Wang, Novemsky, Dhar, and Baumeister (2009): Schindlers List (high semantic complexity) and Bad Santa (low semantic complexity). When consumers had high self regulatory resources (analogous, but not exactly the same as, being in a positive affective state), they chose complex products ( Schindlers List ) over simple products ( Bad Santa). In contrast, when consumers self regulatory resources are spent (viscerally analogous, but not exactly the same as, being in a negative affective state), consumers avoided S chindlers List a movie that is high in semantic complexity, even though the movie was something they themselves regarded as the better movie (based
139 on participant self reports). An interesting question hitherto not raised is: Is it mere complexity, or th e presence of punishers, that contribute to consumer aversion? (i.e., Which process has the bigger influence, nonspecific amplification leading to greater semantic complexity, or specific amplification leading to semantic generation of punishers?) The m ere complexity hypothesis is intriguing it suggests that consumer working memory may have a lower threshold in negative affective states (or when self regulatory resources are spent), regardless of the valence of the individual bits of information. There i s suggestive evidence in the literature showing that negative affective states decrease the organisms ability to process complex information, because negative affective states cause a tunnel vision effect, in which many bits of information are disregarded (Baddeley 2007). Neurological research supports the idea that negative affective states may reduce the organisms mere capacity for information processing, and so consumers who had a bad day will avoid semantically complex products (e.g., complex movies or music ) because their semantic systems cannot handle the incoming information, regardless of the quality or valence. In contrast, positive affective states may increase the organisms mere capacity for information processing (e.g., Fredricksons Broaden and Build Theory, Fredrickson 2004; see also Ashby, Isen, and Turken 1999 for a consistent but different process explanation), and so consumers who had a good day will prefer semantically complex products because the increase in processing capacity (whet her this is mediated by an absolute increase in capacity, mediated by chunking processes, or mediated by the activation of nodes with lower resting activation levels, e.g., Ashby et al. 1999) enables
140 the consumer to process more complex products. Since com plexity increases the potential for learning (Berlyne 1974), consumers will prefer more complex products as long as their affect dependent capacity can handle the workload. The second explanation is the punisher hypothesis i.e., the presence of punisher s that are invariably included in highly complex products threatens consumers in negative affective states. This article provides supportive evidence for the punisher hypothesis among consumers in negative affective states with access to the spicy chips, and a similar process may also operate for semantic products such as movies and music. In contrast to the mere complexity hypothesis, the punisher hypothesis posits that the information processing system is selectively vigilant against potential punishers i n negative affective states, and so the consumer would avoid (or consume less) of the product. For semantic products, punishers may come in the form meanings that predict a threatening outcome (e.g., Coldplays Grammy awardwinning song Viva la Vida [Death and All His Friends ] ). The meaning that is generated by the song, movie, or play is threatening to consumers under negativ e affect, hence the consumer avoids it. Under this explanation, the consumer would not mind highbrow, semantically complex products (such as the Broadway play Les Misrables), as long as it does not contain semantic punishers. This approach would enrich our understanding of consumer self regulation, by exploring what conditions make consumers more sensitive to cues of increment/decrement (e.g., punishers and rewards), and what conditions make consumers more sensitive to limits on capacity for learning (e. g., mere complexity).
141 Finally, our finding that consumers may find punishers rewarding in positive affective states raises an interesting question. What contributes to longterm consumer happiness is it the absence of punishers in life events, or the co mplexity of life events collectively? John Stuart Mill (1861) appears to side with complexity, and he is well known for his paradoxical axiom a dissatisfied Socrates is not only better than a satisfied fool, but happier, too in his classic work, Utilitar ianism The implication of Mills view is that punishers add complexity to life and by doing so, adds learning (which generates pleasure), an idea that dovetails with Berlynes (1960, 1970, 1974) views of complexity and learning. Indeed, empirical research has shown that people who encounter a moderate number of negative events (1 negative event per 2.9 positive events) are happier than people who encounter a low number of negative events (less than 1 negative event per 11.6 positive events) (Fredrickson and Losada 2005; Losada 1999; Losada and Heaphy 2004). Since Losadas data was generated from mathematical modeling, he did not give a theoretical argument as to why too few negative percepts would be worse than a moderate amount of negative percepts. It wou ld be interesting to explore what about negative events makes consumers happier. Is it the fact that encountering events that include punishers causes learning that increases the value of the other stimuli that are about to be experienced (e.g., hypervalua tion of incoming stimulus resulting from perceptual contrast)? Or is it the fact that learning about a punisher increases knowledge about the relationships between that punisher and other rewards (e.g., configurative learning)? If so, it would be interesti ng to investigate whether consumers who experience a moderate number of punishing consumption episodes are happier over time, compared to consumers who
142 experience a small number of punishing consumption episodes. If life is like a box of chocolates, it may serve the consumer well to include a number of chocolates that are punishing.
143 LIST OF REFERENCES Aaker, Jennifer, Aimee Drolet, and Dale Griffin (2008), Recalling Mixed Emotions, Journal of Consumer Research 35 (August), 268278. Aaker, Jennifer L ( 1997), Dimensions of Brand Personality Journal of Marketing Research; 34 (August), 343 47. Adaval Rashmi (2003), How Good Gets Better and Bad Gets Worse: Understanding the Impact of Affect on Evaluations of Known Brands, Journal of Consumer Research 30 (December), 35267. Agrawal, Nidhi, Geeta Menon, and Jennifer L. Aaker (2007), Getting Emotional About Health, Journal of Marketing Research 44 (February), 100113. Allen, Michael W., Richa Gupta, and Arnaud Monnier (2008), The Interactive Effect of Cultural Symbols and Human Values on Taste Evaluation, Journal of Consumer Research 35 (August), 294308. Allen, Chris T., Karen A. Machleit, and Susan S. Kleine (1992), A Comparison of Attitudes and Emotions as Predictors of Behavior at Diverse Levels of Behavioral Experience, Journal of Consumer Research 18 (March), 493504. Andrade, Eduardo B. (2005), Behavioral Consequences of Affect: Combining Evaluative and Regulatory Mechanisms, Journal of Consumer Research 32 (December), 35562. Andrade, Eduardo B and Joel B. Cohen (2007), "On the Consumption of Negative Feelings," Journal of Consumer Research 34 (October), 283300 Andrews, Mark, Gabriella Vigliocco, and David Vinson (2009), Integrating Experiential and Distributional Data to Learn Semantic Representations, Psychological Review, 116 (3), 46398. Ashby, F. Gregory, Alice M. Isen, and And U. Turken (1999), A Neuropsychological Theory of Positive Affect and its Influenc e on Cognition, Psychological Review 106 (July), 52950. Baddeley, Alan (2007), Working Memory, Thought, and Action, New York, NY, US: Oxford Uni versity Press. Barbano, Flavia and Martine Cador (2007), Opioids for Hedonic Experience and Dopamine to get Ready for it, Psychopharmacology, 191 (April), 497506. Baron, Reuben M. and David Kenny (1986), The Moderator mediator Variable Distinction in Social Psychological Research: C onceptual, Strategic, and Statistical Considerations, Journal of Personality and Social Psychology 51 (December), 117382.
144 Barone, Michael J. (2005), The Interactive Effects of Mood and Involvement on Brand Extension Evaluations, Journal of Consumer Ps ychology 15 (3), 26370. Batra, Rajeev and Michael L. Ray (1986), Affective Responses Mediating Acceptance of Advertising, Journal of Consumer Research 13 (September), 23449. Batra, Rajeev and Douglas M. Stayman (1990), The Role of Mood in Advertising Effectiveness, Journal of Consumer Research 17 (September), 20314. Baumgartner, Hans, and Rik Pieters (2008), Goal directed Consumer Behavior, In C. Haugtvedt, P. Herr, & F. Kardes (Eds.), Handbook of Consumer Psychology Mahwah: Psychology Press. Berlyne, Donald E. (1960), Conflict, Arousal, and Curiosity New York, NY, US: McGraw Hill. ____________ (1970), Novelty, Complexity, and Hedonic Value, Perception and Psychophysics, 8 (November), 279 86. ____________ (1974), Studies in the New Experimental Aesthetics: Steps toward an Objective Psychology of Aesthetic Appreciation, Oxford, England: Hemisphere. Berridge, Kent C. (2009), Liking and Wanting Food Rewards: Brain Substrates and Roles in Eating Disorders, Physiology & Behavior 97 (July), 53755. Berridge, Kent C. and Ter ry Robinson (1998), What is the Role of Dopamine in Reward: Hedonic Impact, Reward Learning, or Incentive Salience? Brain Research Reviews 28 (December), 30969. Block, G ary (2004), Foods Contributing to Energy Intake in the U.S., Journal of Food Comp osition and Analysis 17 (3), 443947. Breslin, Paul (2001), Human Gustation and Flavor, Flavour and Fragrance Journal 16 (6), 43956. Breslin, Paul A. S. and A.C. Spector (2008) Mammalian T aste P erception, Current Biology 18, R148R155. Bodenhausen, Galen (1993), Emotions, Arousal, and Stereotypic Judgments: A Heuristic Model of Affect and Stereotyping, in Affect, Cognition, and Stereotyping: Interactive Processes in Group Perception, ed. Diane Mackie and David Hamilton, San Diego, CA: Academic. B odur, Onur, David Brinberg, and Elose Coupey (2000), Belief, Affect, and Attitude: Alternative Models of the Determinants of Attitude, Journal of Consumer Psychology 9 (1), 17 28.
145 Bower Gordon H. (1981), Mood and Memory, American Psychologist 36 (February), 12948. Brown Walter A. Alan D. Sirota, Raymond Niaura and Tilmer Engebretson (1993), Endocrine Correlates of Sadness and Elation, Psychosomatic Medicine 55 (Sep tember Oct ober ), 45867. Carr, Kenneth D., and Tina H. Bak (1987), Later al Hypothalamic Stimulation Produced Analgesia: Inferred Refractory Period of Directly Stimulated Neurons and Resistance to Pimozide Antagonism, Physiology & Behavior 41(4), 37178. Carr, Kenneth D., Kenneth A. Bonnet, and Eric J. Simon (1982), Medial Thalamic Lesions Reduce the AversionGating Action of Lateral Hypothalamic Stimulation, Brain Research, 246 (August), 34246. Chomsky Noam (1955), Logical Syntax and Semantics; Their Linguistic Relevance, Language, 31 (January to March), 3645. Ciompi, Luc and Jaak Panksepp (2005), Energetic Effects of Emotions on Cognitions, in Consciousness & Emotion, ed. Ralph D. Ellis and Natika Newton Amsterdam, Netherlands: John Benjamins Publishing Company, 23 55. Clark, Margaret and Alice Isen (1982), Toward Understanding the Relationship Between Feeling States and Social Beh avior, in Cognitive Social Psychology ed. Albert Hastorf and Alice Isen, New York, NY: Elsevier/North Holland, 73108. Cohen, Joel B., Michel T. Pham and Eduardo B. Andrade (2008), "The Nature and Role of Affect in Consumer Behavior," in Handbook of Consumer Psychology ed. Curt Haugtvedt, Frank Kardes, and Paul Herr, Mahwah, NJ: Erlbaum, 297348. Collins, Alan M. & Loftus, Elizabeth F. (1975) A S preading A ctivation T heory of S emantic P rocessing, Psychological Review 82, 407428. Cowley Elizabeth (2007), How Enjoyable Was It? Remembering an Affective Reaction to A Previous Consumption Experience, Journal of Consumer Research 34 (December), 494505. Darwin, Charles M.A. (1859), On the Origin of Species by Means of Natural Selec tion, or the Preservation of Favoured Races in the Struggle for Life, Fellow of the Royal, Geological, Linnan, etc. societies; Author of Journal of researches during H. M. S. Beagle's Voyage round the world, London, England: John Murray. De Houwer, Jan, S arah Thomas, and Frank Baeyens (2001), Associative Learning of Likes and Dislikes: A Review of 25 Years of Research on Human Evaluative Conditioning, Psychological Bulletin 127 (November), 85369. Dell, G ary S. (1986) A S preadingactivation T heory of R etrieval in S entence P roduction, Psychological Review 93, 283321.
146 DeMello, Gustavo, Debbie J. MacInnis, and Dave W. Stewart (2007), Threats to Hope: Effects on Reasoning about Product Information, Journal of Consumer Research 34 (October), 153161. D ijksterhuis, Ap, Tanya L. Chartrand, and Henk Aarts (2007), Effects of Priming and Perception on Social Behavior and Goal Pursuit, in Social Psychology and the Unconscious: The Automaticity of Higher Mental Processes ed. John A. Bargh, New York, NY: Psy chology Press, 51131. Drolet, Aimee and Mary Frances Luce (2004) The Rationalizing Effects of Cognitive Load on EmotionBased Tradeoff Avoidance Journal of Consumer Research 31 (1), 63 77. Ebert, Jane E.J. Daniel T. Gilbert, and Timothy D. Wilson (forthcoming), "Forecasting and Backcasting: Predicting the Impact of Events on the Future," Journal of Consumer Research (forthcoming). Eisenberger, Naomi I., Matthew D. Lieberman, and Kipling D. Williams (2003), Does Rejection Hurt? An fMRI Study of Social Exclusion, Science 302 (October), 29092. Erickson, Robert (1963), Sensory Neural Patterns and Gustation, in Y. Zotterman (Ed.), Olfaction and Taste, 1, 205213. Fantino, Marc, John Hosotte, and Mark Apfelbaum (1986), An Opioid Antagonist, Naltrexone, Reduces Preference for Sucrose in Humans, American Journal of Physiology: Regulatory, Integrative, and Comparative Physiology 251 (1), 9196. Fedorikhin, Alexander, C. Whan Park and Matthew Thomson (forthcoming), Beyond Fit and Attitude: The Effect of Emotional Attachment on Consumer Responses to Brand Extensions, Journal of Consumer Psychology Fiorillo, Christopher D. Philippe N. Tobler, and Wolfram Schultz (2003), Discrete Coding of Reward Probability and Uncertainty by Dopamine neurons, Science 299 (March), 1 8981902. Findlay, John M. and Ian D. Gilchrist (2003), Active Vision: The Psychology of Looking and Seeing Oxford, UK: Oxford University Press. Fishbach, Ayelet and Aparna A. Labroo (2007), Be Better or Be Merry? How Mood Influences Self Control, Journal of Personality and Social Psychology 93(2), 15873. Food and Drug Administration (2007), Nutrition Facts to Help Consumers Eat Smart. US Food and Drug Administration, Washington, DC (March 21, 2009; http://www.fda.gov/FDAC/special/foodlabel/facts.html ).
147 Forgas, Joseph P. (1995), Mood and Judgment: The Affect Infusion Model (AIM), Psychological Bulletin 117, (January), 3966. Fredrickson, Barbara L. (2001), The Role of Positive Emotions in Positive Psychology: The Broadenand Build Theory of Positive Emotions, American Psychologist 56 (March), 21 8 26. ____________ (2004), The BroadenandBuild Theory of Positive Emotions, Philosophical Transactions: Biological Sciences 359 (September), 136777. Fredrickson, Barbara L. and Marcial F. Losada (2005), Positive Affect and the Complex Dynamics of Hu man Flourishing, American Psychologist 60 (July), 67886. Gadea, Marien Cristina Gmez, Esperanza Gonzlez Bono, Raul Espert and Alicia Salvador (2005), Increased Cortisol and Decreased Right Ear Advantage (REA) in Dichotic Listening Following a Negative Mood Induction, Psychoneuroendocrinology 30 (February), 12938. Gardner, Meryl (1985), Mood States and Consumer Behavior: A Critica l Review, The Journal of Consumer Research 12 (December), 281300. Germov, John and Lauren Williams (1999), A Sociology of Food and Nutrition: The Social Appetite Melbourne FL: Oxford University Press. Glanz, Karen, Michael Basil, Edward Maibach, Jeanne Goldberg, and Dan Snyder (1998), Why Americans Eat What They Do: Taste, Nutrition, Cost, Convenience, and Weight Control Concerns as Influences on Food Consumption, Journal of the American Dietetic Association 98 (October), 111826. Goldberg, Marvin E. and Gerald J. Gorn (1987), Happy and Sad TV Programs: How They Affect Reactions to Commercials, Journal of Consumer Research 14 (December), 387403. Goldstein, Bruce (2007), Sensation and Perception, Wadsworth Publishing. Gray, John (1918), Grays A natomy th e figure represents a faithful reproduction of a lithograph plate from Gray's Anatomy a two dimensional work of art, and is not copyrightable in the U.S. as per Bridgeman Art Library v. Corel Corp. ; the same is also true in many other countries, including Germany. Unless stated otherwise, i t is from the 20th U.S. edition of Gray's Anatomy of the Human Body originally published in 1918 and therefore lapsed into the public domain. Greene, Terry R. and Helga Noice (1988), Influence of Positive Affect Upon Creative Thinking and Problem Solving in Children, Psychological Reports 63 (December), 89598. Groves, P hilip M and R ichard F Thompson (1970), Habituation: A Dual process Theory, Psychological Review, 77, 41950.
148 Hagedorn, Bryce and Gerald A. Juhnke (2005), Treating the Sexually Addicted Client: Establishing a Need for Increased Counselor Awareness, Journal of Addictions & Offender Counseling, 25 (April), 66 86. Harlow, Harry (1958), The Nature of Love, American Psychologist 13 (August) 67385. Harlow, Harry F., Margaret Harlow, Meyer Kuenne, and D onald Meyer (1950), Learning motivated by a manipulation drive, Journal of Experimental Psychology Vol 40(2), Apr 1950, 228234. Hebb, Donald O. (1958), The motivating effects of exteroceptive stimulation, American Psychologist Vol 13( March) 109113. Held, Richard and Allan Hein (1963), Movement produced Stimulation in the Development of Visually Guided Behavior, Journal of Comparative and Physiological Psychology 56 (October), 87276. Hering, Ewald (1964), Outlines of a Theory of the Light Sense Harvard University Press, Cambridge, MA. Hoebel, Bartley, Nicole M. Avena, and Pedro Rada (2008), An Accumbens DopamineAcetylcholine System for Approach and Avoidance, in Handbook of Approach and Avoidance Motivation, ed. Andrew J. Elliot, New York, NY: Psychology Press, 89107. Hofer, Myron (1987), Early Social Relationships: A Psychobiologists View, Child Development 58 (June), 63347. Hubel, David H., and Torsten N. Wiesel (1963), Receptive Fields of Cells in Striate Cortex of Very Young, Visually Inexperienced Kittens, Journal of Neurophysiology, 26 (November), 9941002. Hurvich Leo M. and Dorothea Jameson (1957), An O pponent process T heory of C olor V ision, Psychological Review 64(November), 384404. Ikemoto, Satoshi and Jaak Panksepp (1999), The Role of Nucleus Accumbens Dopamine in Motivated Behavior : A Unifying Interpretation with Special Reference to Rewardseeking, Brain Research Reviews 31 (December), 641. Imanaka, Akihiro, Shigeru Morinobu, Shigeru Toki, and Shigeto Yamawaki (2008), Neonatal Tactile Stimulation Reverses The Effect of Neonata l Isolation On OpenField and Anxiety like Behavior, and Pain Sensitivity in Male and Female Adult SpragueDawley Rats, Behavioural Brain Research 186 (January), 9197. Isen, Alice, Kimberly Daubman, and Gary Nowicki (1987), Positive Affect Facilitates Creative Problem Solving, Journal of Personality and Social Psychology 52 (June), 112231.
149 Isen, Alice, Thomas Shalker, Margaret Clark, and Lynn Karp (1978), Affect, Accessibility of Material in Memory, and Behavior: A Cognitive Loop? Journal of Personality and Social Psychology 36 (January), 1 12. Johnson, Allison R., and David W. Stewart (2004) A ReAppraisal of the Role of Emotion in Consumer Behavior: Traditional and Contemporary Approaches, in N. Malhotra (Ed.), Review of Marketing Research Volume 1 (p. 333), Armonk NJ: M. E. Sharpe, Inc. Kadohisa, Mikiko Edmund T. Rolls, and Justus V. Verhagen (2005), Neuronal Representations of Stimuli in the Mouth: The Primate Insular Taste Cortex, Orbitofrontal Cortex and Amygdala, Chemical Senses 30 (June), 40119. Kafka, Martin P. (2001), The Paraphiliarelated Disorders: A Proposal for a Unified Classification of Nonparaphilic Hypersexuality Disorders, Sexual Addiction & Compulsivity 8 (July), 22739. Kahn, Barbara, and Alice Isen (1993), The Influence of Positive Affect on Variety Seeking Among Safe, Enjoyable Products, The Journal of Consumer Research 20 (September), 25770. Kalin, Ned H., Steven E. Shelton, and Charles M. Barksdale (1989), Behavioral and Physiologic Effects of CRF Administered to Infant Primates Undergoing Maternal Separation, Neuropsychopharmacology 2 (June), 97104. Kandel, Eric, JH Schwartz, and TM Jessell (2000), Principles of Neural Science McG raw Hill, New York. Kang, Min Jeong, Ming Hsu Ian Krajbich George L oewenstein, Samuel McClure Joseph Wang, Joseph Taoyi and Colin Camerer (2009), The W ick in the C andle of L earning: Epistemic C uriosity A ctivates R eward C ircuitry and E nhances M emory, Psychological Science Vol 20(8), Aug 2009, 963973. Keller, Anand, Punam, Isaac M. Lipkus, and Barbara K. Rimer (2003), Affect, Framing, and Persuasion, Journal of Marketing Research 40 (February), 5464. ___________ (2002), Depressive Realism and Health Risk Accuracy: The Negative Consequences of Positive Mood, Journal of Consumer Research 29 (June), 5769. Keller, Kevin (2003), Brand Synthesis: The Multidimensionality of Brand Knowledge, Journal of Consumer Research 29 (Ma rch), 595600. Keppel, Geoffrey and Thomas D. Wickens (2004), Design and Analysis: A Researchers Handbook Upper Saddle River, NJ: Pearson/Prentice Hall. Keverne, Eric (2005), Odor Here, Odor There: Chemosensati on and Reproductive Function, Nature Neuroscience, 8 (December), 163738.
150 Keverne, Eric, Nicholas Martensz, and Bernadette Tuite (1989), Betaendorphin Concentrations in Cerebrospinal Fluid of Monkeys Are Influenced By Grooming Relationships, Psychoneuroendocrinology 14 (January), 15561. Kim Un kyung, Eric Jorgenson, Hilary Coon Mark Leppert Neil Risch, and Dennis Drayna (2003), Positional cloning of the human quantitative trait locus underlying taste sensitivity to phenylthiocarbamide, Science 299 ( February) 12211225. Kringelbach, Morten L. (2009), The Pleasure Center: Trust Your Animal Instincts, New York, NY, US: Oxford University Press. Kringelbach, Morten L. and Kent Berridge (2010) Pleasures of the Brain, New York, NY, US: Oxford University Press. Krishna, Aradhna and Maureen Morrin (2008), Does Touch Affect Taste? The Perceptual Transfer of Product Container Haptic Cues, Journal of Consumer Research 34 (April), 80718. Kuhn, Cynthia M., John Pauk, and Saul Schanberg (1990), Endocrine Responses to Mother Infant Separation in Developing Rats, De velopmental Psychobiology 23 (May), 395410. LaBar, Kevin S. and Joseph E. LeDoux (2006), Fear and Anxiety Pathways, in Understanding Autism: From Basic Neuroscience to Treatment ed. Steven O. Moldin and John L. R. Rubenstein, Boca Raton, FL, US: CRC Press, 13354. Labroo, Aparna A. and Anirban Mukhopadhyay (2009), Lay Theories of Emotion Transience and the Search for Happiness: A Fresh Persp ective on Affect Regulation, Journal of Consumer Research 36 (August), forthcoming. Labroo, Aparna A. and Vanessa M. Patrick (2009), Providing a Moment of Respite: Why a positive mood helps seeing the big picture, Journal of Consumer Research 35(5), 8 009. Labroo Aparna A. and Suresh Ramanathan (2007), The Influence of Experience and Sequence of Emotions on Ad Attitudes, Journal of Consumer Research 33 (4), 5238. LeDoux, Joseph (2002), Synaptic Self: How Our Brains Become Who We Are, Viking Pengui n, NY. Lee, Angela and Brian Sternthal (1999), Effects of Positive Mood on Memory, Journal of Consumer Research 26 (2), 115127. Lee, Leonard, On Amir, and Dan Ariely (forthcoming). In Search of Homo Economicus: Cognitive Noise and the Role of Emotion in Preference Consistency, Journal of Consumer Research.
151 Lewis, Marc (2000), Emotional Self organization at Three Time Scales, in Emotion, Development, and Self Organization: A Dynamic Systems Approach to Emotional Development ed. Marc Lewis and Isabela Granic, New York, NY: Cambridge, 3769. Lim, Juyun and Barry Green (2007), The Psy chophysical Relationship between Bitter Taste and Burning Sensation: Evidence of Qualitative Similarity, Chemical Senses 32 (January), 3139. Li ndemann, Bernd (2001), Receptors and Transduction in Taste, Nature, 413 (September), 21925. Loewenstein, George (2006), The Pleasures and Pains of Information, Scien ce, 312 (5774), 70406. Losada, Marcial (1999), The Complex Dynamics of High Performance Teams Mathematical and Computer Modelling, 30 (November) 17992. Losada, Marcial and Emily Heaphy (2004), The Role of Positivity and Connectivity in the Performance of Business Teams: A Nonlinear Dynamics Model, American Behavioral Scientist 47 (6), 74065. Luce, Mary Frances, Payne, John W., and Bettman, Jim R. (1999), "Emotional Trade off Difficulty and Choi ce," Journal of Marketing Research, 36, 143 159. MacDonald, Geoff and Rachel Kingsbury (2006), Does Physical Pain Augment Anxious Attachment? Journal of Social and Personal Relationships 23 (April), 291304. MacKenzie, Scott B., Richard J. Lutz, and George E. Belch (1986), The Role of Attitude toward the Ad as a Mediator of Advertising Effectiveness: A Test of Competing Explanations, Journal of Marketing Research, 23 (May), 13043. Mansbach, Robert S. Mark A. Geyer, and David L. Braff (1988), Dopaminergic Stimulation Disrupts Sensorimotor Gating in the Rat, Psychopharmacology, 94 (April), 50714. Martel, Frances M., Claire M. Nevison, David F. Rayment, Michael J.A. Simpson, and Eric B. Keverne (1993), Opioid Receptor Blockade Reduces Maternal Affect and Social Grooming in Rhesus Monkeys Psychoneuroendocrinology 18 (March) 307 21. MacInnis, Deborah J., and Gustavo DeMello (2005), The Concept of Hope and its Relevance to Product Evaluation and Choice, Journal of Marketing, 69 (1), 114. MacInnis, Deborah J., Vanessa Patrick, and C. Whan Park (2006), Looking Through the Cryst al Ball: Affective Forecasting and Misforecasting in Consumer Behavior, Review of Marketing Research 2, 4380.
152 McCabe, Ciara and Edmund T. Rolls (2007), Umami: A Delicious Flavor formed by Convergence of Taste and Olfactory Pathways in the Human Brain, European Journal of Neuroscience 25 (March), 185564. McGlone, Francis, Ake Vallbo, Hakan Olausson, Line Loken, and Johan Wessberg (2007), Discriminat ive Touch and Emotional Touch Canadian Journal of Experimental Psychology/Revue Canadienne de Psychologie Exprimentale, 61 (September), 17383. Meyerhof, Wolfgang (2008), Sour Taste Cells Functionally Identified, The Journal of Physiology 586 (June), 281919. Mill, John Stuart (1861), Utilitarianism London, England: Longmans. Miller, George (1956), The Magical Number Seven, Plus or Minus Two: Some Limits On Our Capacit y for Processing Information, Psychological Review, 63 (March), 8197. Mitchell, Andrew (1986), The Effect of Verbal and Visual Components of Advertisements on Brand Attitudes, Journal of Consumer Research, 13 (June), 1224. Mitchell, James E., Melis sa Pederson Mussell, Carol B. Peterson, Scott Crow, Stephen A. Wonderlich, Ross D. Crosby, Traci Davis, and Chris Weller (1999), Hedonics of Binge Eating in Women with Bulimia Nervosa and Binge Eating Disorder, International Journal of Eating Disorders 26 (September), 16570. Mora Ferrer, Carlos and Volker Gangluff (2000), D2Dopamine Receptor Blockade Impairs Motion Detection in Goldfish, Visual Neuroscience 17 (March), 17786. ___________ (2002), D2Dopamine Receptor Blockade Modulates Temporal Res olution in Goldfish, Visual Neuroscience 19 (November), 80715. Morales, Andrea C. and Gavan J. Fitzsimons (2007), Product Contagion: Changing Consumer Evaluations through Physical Contact with 'Disgusting' Products, Journal of Marketing Research 44 ( May), 27283. Mueller Ken L. Mark Hoon, Isolde Erlenbach, Jayaram Chandrashekar Charles Zuker and Nicholas Ryba (2005) The R eceptors and C oding L ogic for B itter T aste, Nature, 434 ( March) 225229. Munro, Donald H. (1967), Empiricism and Ethics, Oxford University Press London, UK N ational A eronautics and S pace A dministration (2007), Electromagnetic Spectrum Properties p ermission granted to copy, distribute and/or modify this document under the terms of the GNU Free Do cumentation License, Version 1.2 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front Cover Texts, and no Back Cover Texts.
153 Nelson, Eric and Jaak Panksepp (1998), Brain Substrates of Infant Mother Attachment: Contributions of Opioids, Oxytocin, and Norepinephrine, Neuroscience & Biobehavioral Reviews 22 (May), 43752. Nettle, Daniel (2006), The Evolution of Personality Variation in Humans and Other Animals, American Psychologist 61, (September), 62231. ____________ (forthcoming), An Evolutionary Model of Low Mood States, Journal of Theoretical Biology Nisbett, Richard and Timothy Wilson (1977), Telling More Than We Can Know: Ver bal Reports about Mental Processes, Psychological Review, 84 (May), 23159. Oliver, Richard L. (1993), Cognitive, Affective, and Attribute Bases of the Satisfaction Response, Journal of Consumer Research 20 (December), 41830. Panksepp, Jaak, Eric Nels on, and Marni Bekkedal (1997), Brain Systems for the Mediation of Separation Distress and Social Reward: Evolutionary Antecedents and Neuropeptide Intermediaries, in The Integrative Neurobiology of Affiliation, ed. C. Sue Carter, Izja I. Lederhendler, and Brian Kirkpatrick, Cambridge, MA: MIT Press, 22144. Panksepp, Jaak (1998), Affective Neuroscience: The Foundations of Human and Animal Emotions, New York, NY: Oxford University Press. ___________ (2003), Feeling the Pain of Social Loss, Science, 302 ( October), 23739. Panksepp, Jaak and Joseph Moskal (2008), Dopamine and SEEKING: Subcortical Reward Systems and Appetitive Urges, in Handbook of Approach and Avoidance Motivation, ed. Andrew J. Elliot, New York, NY: Psychology Press, 6787. Panksepp, J aak, Najma Najam, and Flores Soares (1979), Morphine Reduces Social Cohesion in Rats, Pharmacology, Biochemistry, and Behavior, 11 (August), 13134. Panksepp, Jaak, Eric Nelson, and Marni Bekkedal (1997), Brain Systems for the Mediation of Separation Di stress and Social Reward: Evolutionary Antecedents and Neuropeptide Intermediaries, in The Integrative Neurobiology of Affiliation, ed. C. Sue Carter, Izja I. Lederhendler, and Brian Kirkpatrick, Cambridge, MA: MIT Press, 22144. Panksepp, Jaak, Jay Bean, Paul Bishop, Tom Vilberg, and Tony Sahley (1980), Opioid Blockade and Social Comfort in Chicks, Pharmacology Biochemistry & Behavior 13 (November), 67383.
154 Park, C. Whan, Deborah MacInnis, and Joseph Priester (2006). Beyond Attitudes: Attachment and Consumer Behavior. Seoul National Journal 12(2), 3 36. Partala, Timo and Veikko Surakka (2003), Pupil Size Variation as an Indication of Affective Processing, International Journal of HumanComputer Studies, 59 (July), 18598. Patrick, Vanessa, Debbie MacInnis, C. Whan Park (2007), Not as Happy as I Thought I'd be? The Role of Affective Misforecasting in Consumer Satisfaction, Journal of Consumer Research 33(March), 479490. Pecina, Susana, Barbara Cagniard, Kent C. Berridge, Wayne Aldridge, and Xiaoxi Zhuang (2003), Hyperdopaminergic Mutant Mice Have Higher Wanting but not Liking for Sweet Rewards, Neuroscience 23 (October), 9395402. Pecina, Susana, Kyle S. Smith, and Kent C. Berridge (2006), Hedonic Hot Spots in the Brain, The Neuroscientist 12 (December), 50011. Peter, Paul and Jerry Olson (2001), Consumer Behavior Irwin, Chicago, IL. Pham, Michel Tuan (1998), Representativeness, Relevance, and the Use of Feelings in Decision Making, Journal of Consumer Research 25 (September), 14459. Pham, Michel Tuan and Tamar Avnet (2009), Contingent Reliance on the Affect Heuristic as a Function of Regulatory Focus ," Organizational Behavior and Human Decision Processes 208 (2), 267 78. Pham, Michel Tuan, Joel B. Cohen, John W. Pracejus, and G. David Hughes (2001), Affect Monitoring and the Primacy of Feelings in Judgment Journal of Consumer Research 28 (September), 16788. Posnock, Ross (1991), The Trial of Curiosity, Oxford University Press. Pritchard, Thomas C., David A. Macaluso, and Paul J. Eslinger (1999), Taste Perception in Patients with Insular Cortex Lesions, Behavio ral Neuroscience 113(August), 66371. Raghunathan, Rajagopal, Michel T. Pham, and Kim P. Corfman (2006), Informational Properties of Anxiety and Sadn ess, and Displaced Coping, Journal of Consumer Research 32, 596602. Ramanathan, Suresh, and Geeta Menon (2006), Time Varying Effects of Chronic Hedonic Goals on Impulsive Behavior, Journal of Marketing Research 43 (November), 628641. Ratner, Rebecca K., & Herbst, Kenneth C. (2005). When Good Decisions Have Bad Outcomes: The Impact of Affect on Switching Behavior, Organizational Behavior and Human Decision Processes 96 (1) 23 37.
155 Rolls, Edmund (2004), The Functions of the Orbitofrontal Cortex, Brain and Cognition, 55 (June), 1129. ___________ (2005), Emotions Explained, Oxford, UK: Oxford University Press. ___________ (2008), Functions of the Orbitofrontal and Pregenual Cingulate Cortex in Taste, Olfaction, Appetite, and Emotion, Acta Physiologica Hungarica 95 (June), 13164. Rolls, Edmund and Fabian Grabenhorst (2008), The Orbitofrontal Cortex and Beyond: From Affect to Decision making, Progress in Neurobiology, 86 (September), 21644. Rolls, Edmund T., Zenon J. Sienkiewicz and Simon Yaxley (1989), Hunger Modulates the Responses to Taste Stimuli of Single Neurons in the Caudolateral Orbitofrontal Cortex of the Macaque Monkey, European Journal of Neuroscience 1 (January), 53 60. Rowe, Gillian Jacob B. Hirsh and Adam K. Anderson (2007), Positive Affect Increases the Breadth of Attentional Selection, Proceedings of the National Academy of Sciences of the United States of America 104 (January), 38388. Sacks, Oliver, Marjorie Kohl, Charles Messeloff, and W alter Schwartz (1972), Effects of Levodopa in Parkinsonian Patients with Dementia, Neurology 22 (May) 51619. Salamone, John, Michael Cousins, and Sherri Bucher (1994), Anhedonia or Anergia? Effects of Haloperidol and Nucleus Accumbens Dopamine Deplet ion on Instrumental Response Selection in a Tmaze Cost/Benefit Procedure, Behavioural Brain Research 65 (December), 22129. Schiller, Daniela Ifat Levy, Yael Niv, Joseph E. LeDou x and Elizabeth A. Phelps (2008), From Fear to Safety and Back: Reversal of Fear in the Human Brain, Journal of Neuroscience 28 (Novem ber), 1151725. Schultz, Wolfram (2006), Behavioral Theories and the Neurophysiology of Reward, Annual Review of Psychology 57 (January), 87115. Schwarz, Norbert and Gerald L. Clore (1983), Mood, Misattribution, and Judgments of Wellbeing: Informativ e and Directive Functions of Affective States, Journal of Personality & Social Psychology 45 (September), 51323. ___________ (2007), Feelings and Phenomenal Experiences, in Social Psychology: Handbook of Basic Principles ed. E. Tory Higgins and Arie Kruglanski, New York, NY: Guilford, 385407. Shakko, S. (2008), Zeno of Citium reprinted with permission under the Creative Commons Attribution Share Alike 3.0 license.
156 Shiv, Baba (2007), Emotions, Decisions and the Brain, Journal of Consumer Psychology Vol. 17 (3), 174178. Shiv Baba and Alexander Fedorikhin (1999), Heart and Mind in Conflict: The Interplay of Affect and Cognition in Consumer Decision Making, Journal of Consumer Research, 26 (December), 27892. Shiv, Baba, Alexander Fedorikhin, and Stephen M. Nowlis (2005), Interplay of the Heart and Mind in Decision Making, in Inside Consumption: Frontiers of Research on Consumer Motives, Goals, and Desire, ed. Ratti Ratneshwar and David Mick. Skov Martin (2010), The P leasure of A rt. Pleasures of the B rain, Kringel bach, Morten L. (Ed); Berridge, Kent C. (Ed) Pleasures of the brain (pp. 270283). New York, NY, US: Oxford University Press. Small, Deborah, Lerner, J.S., and Fischhoff, Baruch (2006), Emotion priming and spontaneous attributions for terrorism: America ns reactions in a national field experiment, Political Psychology, 27 (2), 289298. Smith, Kyle S. and Kent C. Berridge (2007), Opioid Limbic Circuit for Reward: Interaction between Hedonic Hotspots of Nucleus Accumbens and Ventral Pallidum, The Journal of Neuroscience, 27 (February), 15941605. Sorensen, L. B.; P. Moller, A. Flint, M. Martens, A. Raben (2003), Effect of Sensory Perception of Foods on Appetite and Food Intake: A Review of Studies on Humans, International Journal of Obesity 27 (Octobe r), 115266. Spanagel, Rainer, Albert Herz, and Toni Shippenberg (1990), The Effects of Opioid Peptides on Dopamine Release in the Nucleus Accumbens: An In Vivo Microdialysis Study, Journal of Neurochemistry 55 (October), 173440. Thompson, R.F., and W. A. Spencer (1966), Habituation: A Model Phenomenon for the Study of Neuronal Substrates of Behavior, Psychological Review, 73, 1643. Tindell, Amy, Kent Berridge, Jun Zhang, Susana Pecina, and Wayne Aldridge (2005), Ventral Pallidal Neurons Code Incent ive Motivation: Amplification by Mesolimbic Sensitization and Amphetamine, European Journal of Neuroscience, 22 (November), 261734. Titchener, Edward (1896), An Outline of Psychology London, England: The Macmillan Co. Tomiski, T. (2006), Aristippus of Cyrene, This image is in the public domain because its copyright has expired.
157 Trope, Yaacov, and Eva M. Pomerantz (1998), Resolving Conflicts Among Self evaluative Motives: Positive Experiences as a Resource for Overcoming Defensiveness, Motivation and Emotion, 22 (1), 5372. Uvns Moberg, Kerstin (1997), Physiological and Endocrine Effects of Social Contact, Annals of the New York Academy of Sciences, 807 (January), 14663. ___________ (1998), Oxytocin May Mediate the Benefits of Positive Social Interaction and Emotions, Psychoneuroendocrinology 23 (November), 81935. Vallbo, Ake, Hakan Olaus son, and Johan Wessberg (1999), Unmyelinated Afferents Constitute a Second System Coding Tactile Stimuli of the Human Hairy Skin, Journal of Neurophysiology 81 (June), 275363. Velten, Emmett (1968), A Laboratory Task for the Induction of Negative Mood States, Behavioral Research and Therapy, 6 (November), 47382. Vohs, Kathleen D., Roy F. Baumeister, and George Loewenstein (2007), Do Emotions Help or Hurt Decision Making? A Hedgefoxian Perspective New York: NY, Russell Sage. Wadlinger, Heather and Derek Isaacowitz (2006), Positive Mood Broadens Visual Attention to Positive Stimuli, Motivation and Emotion, 30 (March), 89101. Wang, Jing, Nathan Novemsky, Ravi Dhar and Roy Baumeister (2009), "Tradeoffs and Depletion in Choice," Journal of Marketing Research forthcoming issue. Watson, David Lee A. Clark, and Auke Tellegen (1988), Development and validation of brief measures of positive and negative affect: The PANAS scales, Journal of Personality and Social Psychology 54 (June), 106370. Wierenga, Christina E.; Perlstein, William M.; Benjamin, Michelle; Leonard, Christiana M.; Gonzalez Rothi, Leslie; Conway, Tim; Cato, M. Allison; Gopinath, Kaundinya; Briggs, Richard; Crosson, Bruce (2009), Neural substrates of object identification: Functional magneti c resonance imaging evidence that category and visual attribute contribute to semantic knowledge. Journal of the International Neuropsychological Society 15 (March), 169181. Wiesel, Torsten N. and David H. Hubel (1963a), Effects of Visual Deprivation on Morphology and Physiology of Cells in the Cat's Lateral Geniculate Body, Journal of Neurophysiology, 26 (November), 97893. ___________ (1963b), SingleCell Responses in Striate Cortex of Kittens Deprived of Vision in One Eye, Journal of Neurophysi ology, 26 (November), 100317. Zillmann, Dolf (1988), Mood Management through Communication Choices, American Behavioral Scientist 31 (January), 32741.
158 Zubieta, JonKar, Terence A. Ketter, Joshua A. Bueller, Yanjun Xu, Michael R. Kilbourn, Elizabeth A. Young, and Robert A. Koeppe (2003), Regulation of Human Affective Responses by Anterior Cingulate and Limbic MuOpioid Neurotransmission, Archives of General Psychiatry, 60 (November), 114553. Zubieta, JonKar, Yolanda Smith, Joshua Bueller, Yanjun Xu, Michael Kilbourn, Douglas Jewett, Charles Meyer, Robert Koeppe, and Christian Stohler (2001), Regional Opioid Receptor Regulation of Sensory and Affective Dimensions of Pain, Science, 293 (July), 31115.
159 BIOGRAPHICAL SKETCH Dennis Danny O. Uy is a Ph.D. in m arketing from the University of Florida, Warrington College of Business. Prior to matriculating at University of Florida, he received his MBA from the University of Chicago. He is the son of Peter S. Uy and Christine O. Uy (deceased).