<<

This article was downloaded by: [EBSCOHost EJS Content Distribution - Superceded by 916427733] On: 29 June 2010 Access details: Access Details: [subscription number 911724993] Publisher Psychology Press Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37- 41 Mortimer Street, London W1T 3JH, UK

Cognition & Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713682755

The perception and categorisation of emotional stimuli: A review Tobias Broscha; Gilles Pourtoisab; David Sandera a University of Geneva, Geneva, Switzerland b Ghent University, Ghent, Belgium

First published on: 19 June 2009

To cite this Article Brosch, Tobias , Pourtois, Gilles and Sander, David(2010) 'The perception and categorisation of emotional stimuli: A review', & Emotion, 24: 3, 377 — 400, First published on: 19 June 2009 (iFirst) To link to this Article: DOI: 10.1080/02699930902975754 URL: http://dx.doi.org/10.1080/02699930902975754

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. COGNITION AND EMOTION 2010, 24 (3), 377400

INVITED REVIEW

The perception and categorisation of emotional stimuli: A review

Tobias Brosch University of Geneva, Geneva, Switzerland Gilles Pourtois University of Geneva, Geneva, Switzerland, and Ghent University, Ghent, Belgium David Sander University of Geneva, Geneva, Switzerland

When we perceive our environment, we rapidly integrate large amounts of incoming information into categories that help to guide our understanding of the world. Some stimuli are more relevant for our well-being and survival than others, for example stimuli that signal a threat or an opportunity for growth and expansion. In this review we examine the special role of such ‘‘emotional’’ stimuli in perception and categorisation. To this end, we first discuss some fundamental aspects of perception, with an emphasis on the cognitive process of categorisation. We then tackle the questions: ‘‘What is an emotional stimulus?’’ and ‘‘What is an emotion category?’’ Afterwards, we illustrate, with a review of key findings from the empirical literature, (i) how stimuli are categorised as emotional, and (ii) how the perceptual processing of emotional stimuli is prioritised to allow for a rapid preparation of adaptive responses. To conclude, we discuss how research on the perception of emotional stimuli can contribute to current debates in psychology, namely (i) about the role of bottom-up vs. top-down factors in emotional processing and experience, and (ii) about the nature of the relationship between cognition and emotion.

Keywords: Emotion; Perception; Categorisation.

INTRODUCTION to rapidly make sense of our multifarious and fast- changing environment. To do so, we create an In order to successfully move about in the world internal mental representation of the stimuli that and respond to its permanent challenges, we have are immediately present in our surroundings. Any Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010

Correspondence should be addressed to: Tobias Brosch, Swiss Centre for Affective Sciences, University of Geneva, 7, Rue des Battoirs, 1205 Geneva, Switzerland. E-mail: [email protected] The preparation of this article was supported by the National Centre of Competence in Research for Affective Sciences, financed by the Swiss National Science Foundation.

# 2009 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business 377 http://www.psypress.com/cogemotion DOI:10.1080/02699930902975754 BROSCH, POURTOIS, SANDER

given external object in the environment, the may differ in some respects from other categories. distal stimulus (e.g., a stone) is not processed as In line with this, a lot of empirical evidence such, but is represented in the organism as a illustrates how people make use of such special physical pattern on the senses, the categories to guide their perception of the envir- proximal stimulus (e.g., the pattern of light on the onment. For example, it has been shown that retina reflected by the stone). Perception is the people are able to rapidly and accurately classify transformation of the proximal stimulus into a emotional expressions into emotional categories, percept, the accessible, subjective, reportable even when the sender has a cultural background experience that takes the form of an activation different from their own. of a certain category in the mind (e.g., the In this paper, we will examine the special role accessible visual experience of the stone). How of emotional stimuli in perception and categorisa- we perceive our environment is thus profoundly tion. We will first discuss some fundamental shaped by categorisation. When we categorise a aspects of perception in general, with an emphasis stimulus, we group certain objects or concepts as on the central cognitive process of categorisation. equivalent or analogous, thereby reducing the By introducing general principles of perception information complexity of the external world. At and categorisation, we will be able to investigate the same time, a lot of information about the whether similar principles apply to the perception stimulus is inferred due to its association with a and categorisation of emotional and neutral category. The act of categorisation is therefore stimuli, or if different mechanisms may be critical to cognition (see Harnad, 2005) and involved. We will then tackle the question allows us to give meaning to the world. ‘‘What is an emotional stimulus?’’ taking into Sometimes we are confronted with classes of account various definitions from different theories stimuli that have more direct relevance for our of emotion, and review different suggestions of well-being and survival than others. For instance, how the emotional categories we use to classify some stimuli may signal danger or threat, such as and label highly relevant stimuli are determined, predators or enemies, whereas other stimuli signal learned, and eventually used to guide our percep- chances for growing and expansion, such as tion. Afterwards, we will illustrate the preferential potential mates or food sources. Such stimuli perception of emotional stimuli by reviewing require rapid adaptive responses, such as evading some of the key findings from the empirical the threat or approaching the positive stimulus. literature. We will address two main lines of One might expect that, given the high importance research, (i) research focusing on qualitative of such ‘‘emotional’’ stimuli for the organism, effects of emotional stimuli, i.e., research addres- the perceptual processing of these stimuli should sing the question of how people are able to be prioritised to allow for a rapid appraisal of the categorise different stimuli into emotional cate- situation and consequently the rapid preparation gories, and (ii) research focusing on quantitative of an appropriate behavioural response. In line effects of emotional stimuli, i.e., research addres- with this, many everyday examples suggest that sing the question of how the emotionality of a the perception of emotional stimuli is somewhat stimulus can modulate and sometimes even trans- special, or heightened, relative to non-emotional form perception, independent of whether people Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 stimuli. Smiling people, cute babies, erotic scenes, are asked to (consciously) categorise them. We but also poisonous snakes or scenes of war and will conclude with some reflections on how mutilations seem to catch one’s eye more easily research on the perception of emotional stimuli than emotionally ‘‘neutral’’ stimuli. Moreover, as can contribute to some current debates in psy- the process of categorisation is crucial for the chology, namely (i) about the role of bottom-up organisation of perception, one may furthermore vs. top-down factors in emotional processing and expect that extremely relevant stimuli are cate- experience, and (ii) about the nature of the gorised into special emotion categories, which relationship between cognition and emotion.

378 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

A review on a topic as large as ‘‘perception and language, or consciousness. Many scholars have categorisation of emotional stimuli’’ necessarily stressed that perception profoundly depends on has to be selective. For example, we will restrict the process of top-down categorisation (Barrett, ourselves to research using relatively simple 2006b; Davidoff, 2001; Palmeri & Gauthier, stimuli, such as emotional words, pictures of 2004; Rosch, 1975; Rosch, Mervis, Gray, John- emotional scenes, or emotional expressions con- son, & Boyes-Braem, 1976). How we perceive our veyed by face, body, or voice. We will not survey environment is thus shaped by categorisation more complex emotional behaviours, behaviour processes that guide and constrain the organisa- descriptions or emotional events and their effect tion of incoming stimulus information and thus on people’s inferences of , traits, compe- make a conscious representation and identifica- tences, or status and power characteristics. There tion of this information possible. This principle is is a large literature in social and clinical psychol- supposed to hold for all kinds of categories, no ogy on these topics (see, e.g., Augoustinos, matter whether the perceived content is colour, Walker, & Donaghue, 2006, for a review). certain objects, faces, facial expressions of emo- Furthermore, perception does not always involve tion, emotional or any other attribute a conscious subjective experience and we do not (Barrett, 2006b; Cohen & Lefebvre, 2005; David- intend to reduce categorisation to conscious off, 2001). experience. Research on ‘‘unconscious perception’’ During categorisation, a continuously chan- shows that under certain conditions such as ging stimulus is identified against discrete and degraded stimulus input or lapse of or pre-existing categories or conceptual boundaries. awareness, stimuli can nonetheless be categorised This can be experienced, for example, when we to some extent, be partly processed and eventually are watching a rainbow. Even though a rainbow is have an impact on behaviour without being composed of a continuous range of varying consciously experienced (Merikle & Daneman, wavelengths, we perceive chunks of colours rather 1998; Winkielman, Berridge, & Wilbarger, 2005; than a gradual continuum of changing colours. Zajonc, 1980). Here, however, we focus on the Due to the influence of top-down information perception of emotional stimuli, when the cate- about colour categories, the linear physical gorised percept of these stimuli presumably enters changes of the distal and proximal stimuli have awareness. non-linear effects on the percept. Categorisation allows us to structure stimuli by grouping or classifying them according to certain PERCEPTION AS A FUNDAMENTAL principles, such as perceptual similarities (Rosch, CATEGORISATION PROCESS 1978), semantic rules or theories (Murphy & Medin, 1985), implications for goal states What is categorisation? (Barsalou, 1983) or evoked emotional responses How do we perceive and categorise objects? (Niedenthal, Halberstadt, & Innes-Ker, 1999). By Which fundamental psychological mechanisms categorising a stimulus we give meaning to it, as underlie this ability? These questions have been categorisation allows us to make inferences, central to psychology for many decades (see analogies, and predictions about a stimulus and Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 Cohen & Lefebvre, 2005; Palmeri & Gauthier, to communicate about the stimulus with people 2004, for reviews). In fact, perception and object who share our concepts (Niedenthal et al., 1999). recognition sometimes are considered the crucial Object categorisation occurs very rapidly issues that research on human cognition has to (Thorpe, Fize, & Marlot, 1996). An efficient explain (Kourtzi & DiCarlo, 2006), as a deeper categorisation process allows us to constrain, insight into these processes will also substantially guide and summarise the processing of stimuli further the understanding of downstream higher- encountered in the environment with minimal order cognitive processes such as memory, cognitive effort. When a stimulus is categorised, a

COGNITION AND EMOTION, 2010, 24 (3) 379 BROSCH, POURTOIS, SANDER

large amount of relevant information related to two basic terms for the whole range of visible the category is activated and made available, colours (Rosch Heider, 1972). This was inter- whereas irrelevant distinctions within categories preted as showing that a certain categorical are omitted for the sake of cognitive economy organisation of colours may be found universally (McClelland & Rumelhart, 1985; Rosch, 1978; in humans and may be predetermined genetically. Rosch et al., 1976). Ultimately, categorisation On the other hand, results showing that the operates as a strong filter, which drastically possession of certain linguistic colour terms reduces the information content available in the influences the organisation of categories suggested external world by grouping certain objects as that it is mostly language and semantic concepts equivalent. Categorisation furthermore facilitates that shape the organisation of incoming stimulus rapid object discrimination. It is easier to dis- information (Roberson, Davies, & Davidoff, criminate two colours of different shades when 2000). In this view, the placement of the bound- they cross colour category boundaries than aries between categories is not considered to be when they are within the same category, even based on pre-existing universals, but rather on though the differences in wavelength are identical conventions within a cultural group. for the two pairs (Bornstein & Korda, 1984). As will be outlined in the next section, a Similar results have been obtained, e.g., for the similar debate exists for emotional categories (see differentiation of speech phonemes (Liberman, also Boster, 2005). Theoretical suggestions on Harris, Hoffman, & Griffith, 1957) and the how emotional categories are defined go from the discrimination of familiar faces (Beale & Keil, universalistic perspective that there are biologi- 1995). This effect, basically enhancing perceived cally based universal emotion categories, the 1 between-category differences and reducing per- ‘‘basic emotions’’ (Ekman, 1992; Izard, 2007; ceived within-category differences, was coined the Panksepp, 1998), to the notion that multicompo- categorical perception effect (Harnad, 1987). nent patterns of emotional responses that occur with a relatively high frequency are categorised into ‘‘modal emotions’’ (Scherer, 1994b), and How are categories determined? finally to the constructivist perspective that emo- How exactly categories and their boundaries are tional experience is based on the categorisation of determined is a matter of debate. Even for a rather a raw affective quality into emotional categories, simple domain such as colour categories, the which are considered to be man-made concepts theoretical positions that have been advanced (Barrett, 2006a). span all the space between universalistic (cate- gories are determined by perceptual factors based on the properties of colour-coding neurons; Berlin EMOTIONAL CATEGORIES & Kay, 1969) and relativistic views (categories are arbitrarily set based on language and cultural What is an emotional stimulus? conventions about concepts of colour; Whorf, When comparing the role of emotional and non- 1956). Strong support for the universalistic view emotional stimuli in perception, an important came from evidence showing that the Dani, a Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 issue that needs to be tackled is the definition of remote branch of a hunter-gatherer tribe, showed ‘‘emotional stimulus’’: Why is a stimulus perceived the same cognitive organisation of colour as or categorised as ‘‘emotional’’ at all? One way to speakers of English, even though they only used address this question is to find a definition based

1 In this context, one finds frequent analogies between perceptual and emotional categories. For example, according to Izard (2007), ‘‘it is possible to argue by analogy that the capacity to discriminate among basic-emotion feeling states, like discriminating among basic tastes, is innate and invariant across the lifespan’’ and ‘‘the data relating to the underlying neural and behavioral processes suggest that the emergence of discriminable basic emotion is analogous to that for basic tastes’’ (see Sander, 2008).

380 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

on functional considerations. Following this line, prepare the organism for action, while allowing one can begin by asking why ‘‘emotional for some flexibility in terms of the response, as an stimuli’’*which constitute a group of rather emotional stimulus is not associated reflex-like heterogeneous stimuli*should be categorised with a specific response, but an emotion primes an together and treated in a preferential manner by arsenal of potential adaptive responses (Frijda, the organism. To this end, we will briefly review 2007; Scherer, 1994a). In the context of adaptive how different influential psychological theories of responding, it appears furthermore functional to emotion provide important clues to what actually assume that the perceptual processing of emo- renders a stimulus emotional. One should keep in tional stimuli is prioritised to allow for a more mind that, generally, theories of emotion are rapid computation and situation analysis. concerned with the elicitation of emotional Whereas most theories of emotion agree that responses (see also Moors, 2009), but less with emotions serve to organise adaptive responses to the definition of ‘‘emotional stimulus’’ or the stimuli that are important for the survival and processes involved in the perception of such well-being of the organism, different theories stimuli. However, these issues are highly inter- disagree with regards to the mechanisms under- twined, as, very often, the perception of a stimulus lying this adaptive function. as ‘‘emotional’’ will subsequently elicit an emo- Basic emotion theories assume a number of tional response. Thus, it should be possible to distinct basic emotions, including, e.g., , draw some conclusions about the conditions , , , or . In under which a stimulus is perceived as ‘‘emo- this tradition, the term ‘‘basic’’ is used to express tional’’ (and another one is not) based on the three postulates (Ekman, 1992): First, it conveys different theoretical approaches. In addition to the notion that ‘‘there are a number of separate the distinction between emotional and non-emo- emotions which differ from one another in tional stimuli, different theoretical views further- important ways’’, second, it indicates that ‘‘evolu- more allow us to gain insight on the actual tion played an important role in shaping both the cognitive processes involved in the perception unique and the common features which these and categorisation of emotional stimuli. In re- emotions display as well as their current function’’, viewing the different theoretical approaches, we will first outline the main points of agreement and finally, the term refers to the notion that the disagreement between the theories. Afterwards, existence of non-basic emotions can be explained we will discuss some of the issues about which the by combinations of the basic emotions. Basic different approaches disagree in the light of the emotions are defined as programmes that empirical evidence on the perception and categor- are triggered by appropriate eliciting events to isation of emotional stimuli. produce emotion-specific response patterns such Most emotion theories agree that emotional as prototypical facial expressions and physiological stimuli represent a special type of stimulus as they reactions (Ekman, 1992), driven by specific neural possess high relevance for the survival and well- response systems (Panksepp, 1998). According to being of the observer. For instance, some stimuli basic emotion theories, perceptual processing of may signal threats, such as predators or enemies, emotional stimuli is assumed to be essentially Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 whereas other stimuli signal chances for growing organised in a categorical manner, with innate and expansion, such as potential mates or food categories being universally found in humans. sources. Such stimuli require rapid responses, like Some theorists suggest a special role for the basic evading the threat or approaching the positive emotion of fear (O¨ hman & Mineka, 2001). Due stimulus. Emotional responses are adaptive re- to evolutionary reasons, fear/threat-related stimuli sponses to an eliciting stimulus, including action such as angry facial expressions, snakes or spiders tendencies, bodily responses, behavioural re- are thought to be attended to, perceived, recog- sponses and a change in subjective feeling. They nised, remembered and associated with adaptive

COGNITION AND EMOTION, 2010, 24 (3) 381 BROSCH, POURTOIS, SANDER

behavioural output faster or more readily than any Constructivist theories of emotion emphasise the other emotional stimulus. role of culture, language, and high-level cognition Appraisal theories of emotion suggest that emo- in the emergence of emotional experience. Barrett tional processes are elicited as the individual (2006a,b) recently adapted Schachter’s theory of continuously appraises objects, behaviours, events emotion (see also Moors, 2009) to propose a and situations with respect to their relevance for conceptual act approach based on the core affect his/her needs, goals, values, and general well- notion put forward by Russell (2003). According being (Ellsworth & Scherer, 2003). Emotions are to this view, emotion categories are not natural elicited and differentiated on the basis of the entities, but man-made concepts. The experience subjective evaluation of a stimulus or event on a of emotions is based on a process that categorises set of standard criteria or objectives such as a readout of core affect into language-based novelty, intrinsic pleasantness, goal conducive- emotion categories. According to this relativistic ness, and normative significance, as well as the view, language provides conceptual categories, coping potential of the organism. Appraisal is which in turn constrain the process that attributes subjective and thus a function of the individual meaning to stimuli, including emotional ones, and and the specific situation/context, therefore al- thus biases perception. Thus, the perception of a lowing for differences between, e.g., species, age stimulus as emotional should depend heavily on groups, personal dispositions, and cultural con- the language context of the observer (Barrett, texts. The outcome of the appraisals of these Lindquist, & Gendron, 2007b). different criteria is predicted to directly drive To summarise, different theories of emotion response patterning of physiological reactions, differ in the way they conceptualise how stimuli motor expression, and action preparation are categorised as emotional. Basic emotion (Scherer, 2001). With regards to emotion cate- theories claim that certain classes of stimuli gories, it has been suggested that the subjective trigger predefined affect programmes, which experience of the response patterning (‘‘qualia’’)is then elicit specific response patterns. In contrast categorised into ‘‘modal emotions’’ (Scherer, to this rather inflexible mechanism, which is 1994b), reflecting the relatively frequent occur- mainly based on a schema evaluation or a rence of some patterns of responses that are pattern-matching process between a stimulus associated with core concerns or core relational and a template, appraisal theories emphasise the themes (Smith & Kirby, in press; Smith & importance of the subjective evaluation of the Lazarus, 1990). stimulus according to its importance for the Dimensional theories of emotion emphasise the individual. This allows for a greater amount of role of a few key dimensions, usually valence and flexibility and individual adjustment of person, , in the organisation and categorisation of situation and what is perceived as emotional. emotional stimuli. The dimensional approach Dimensional theories propose a very general, allows us to distinguish between negative and economical mechanism linking stimulus proces- positive emotions of different intensities, which sing to an evaluation that basically distinguishes reflects two basic motivational systems, the appe- between positive and negative stimuli and be- titive and the aversive systems that underlie tween stimuli leading to different degrees of Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 approach and withdrawal behaviour, respectively activation. Finally, constructivist theories empha- (Davidson & Irwin, 1999; Lang, 1995; Schneirla, sise the constraining role of language context on 1959). According to Russell (2003), core affect, the the mapping of these dimensions into emotion primary, consciously accessible internal emotional categories. Thus, different theories of emotion state, consists exclusively of an integral blend of differ with regards to how much emphasis they valence and arousal. The affective quality of a put on bottom-up mechanisms and top-down stimulus is the capacity of this stimulus to change mechanisms determining what makes a stimulus core affect. emotional, how it is categorised and how it is

382 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

perceived, with basic emotion theories arguing matter of degree rather than all-or-nothing that it is mainly (but not always) stimulus driven (internal structure), and no sharp boundaries bottom-up processes, appraisal theories suggest- separate category members from non-members ing a more flexible and dynamic mechanism (fuzzy boundaries). Based on early theoretical work taking into account the interaction of stimulus by Rosch and colleagues (1976), emotion cate- and the needs and goals of the observer, and gories (and the category ‘‘emotion’’) have been constructivist theories mainly focusing on the suggested to be hierarchically organised, with constraining top-down effects of mental repre- positive and negative emotions as superordinate sentations and language knowledge. categories, categories such as anger, fear, or happiness at the basic level, and subordinate categories such as wrath, , fury, How are emotion categories determined? (Russell & Fehr, 1987; Shaver, Schwartz, Kirson, Whereas the previous section mainly focused on &O’Connor, 1987). theoretical issues regarding the distinction be- Many prototype categories we use in daily life tween emotional and non-emotional stimuli, now are created based on the correlational structure of we will address the related question of how our properties that observers perceive in the world. internal emotional categories, i.e., the categories Properties of objects do not occur randomly, but we use to classify and label emotions and some combinations tend to co-occur frequently emotionally relevant stimuli, are determined, and (such as ‘‘feathers’’ and ‘‘wings’’), whereas some to which extent these categories may be similar to other combinations rarely occur together (such as or different from non-emotional categories. ‘‘fur’’ and ‘‘wings’’). Although atypical cases do Across many languages we find a large variety exist, in general different stimuli sharing some of categories that we use to describe our emotional common properties can be put into discrete experiences (e.g., ‘‘fear’’) or the elicitors of such categories (e.g., ‘‘birds’’) to simplify the organisa- experiences (e.g., ‘‘threats’’). According to basic tion of the environment (Rosch, 1978; Rosch et emotion theories, the basic emotions constitute al., 1976). Similar processes might be involved in innate categories, which are shaped by - the development of emotion categories, concepts, ary pressures. However, as will be outlined in schemas, or scripts. An event that interrupts goal more detail in a later section of this paper, there attainment might frequently be paired with a are several lines of argumentation that go against subjective experience of and arousal, the notion of universal or innate emotional behavioural attempts at overcoming the blockage, categories. and typical facial and vocal expression patterns If emotion categories like ‘‘fear’’ or ‘‘anger’’ are (Hess, Philippot, & Blairy, 1998), all of which not innate, they may be learned, just like other might then be integrated into a semantic network categories such as ‘‘birds’’ or ‘‘furniture’’. In this representing the anger concept, which guides the context, it has been argued that the boundaries of categorisation of emotional expressions (Russell, emotion categories are not well-defined,2 but that 1991). membership in an emotion category is based on In addition to categories based on the correla- the extent of resemblance to a prototype (proto- tional structure of the environment, other cate- Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 type theory of emotion; Russell & Fehr, 1987). gorical grouping mechanisms have been proposed. States elicited by a certain event or stimulus are For example, it has been suggested that facial perceived to be instances of fear, anger, or expressions of emotion are examples of goal- happiness to the extent that they resemble certain derived categories, optimised to reach the goal of ideal cases. Category membership, thus, is a emotion communication, rather than taxonomic

2 In well-defined categories, category membership can be defined by one or more individually necessary and jointly sufficient features, as is the case for example for ‘‘square’’ or ‘‘grandfather’’.

COGNITION AND EMOTION, 2010, 24 (3) 383 BROSCH, POURTOIS, SANDER

categories, which help to economically describe the To sum up, empirical evidence and theoretical environment (Barsalou, 1985; Horstmann, 2002). considerations indicate that emotion categories When participants were asked to choose typical are not principally determined universally or examples of facial emotion expressions out of biologically, but are learned and continuously several exemplars with different expression inten- adjusted in a flexible way. Grouping of different sities, they chose the most extreme version stimuli into an emotional category can be based (Horstmann, 2002). Thus, the facial expressions on a number of different principles, reflecting the of emotion that are perceived as ‘‘most typical’’ are correlational structure of the environment, the not the ones that are encountered frequently, but optimisation of communicative goal attainment, the ones that are most suitable to communicate a the combination of a set of appraisal outputs, or certain emotion (see also Smith, Cottrell, Gosse- the subjective emotional feeling elicited by a lin, & Schyns, 2005). This suggests that the basic stimulus. Emotion categories thus can be con- emotion categories may have been created to sidered as adaptive and flexible emotion scripts, maximise communicative goal attainment. integrating aspects of emotion elicitation (apprai- Furthermore, the psychological concept of core sal, core relational themes) and of the emotional relational themes (Smith & Kirby, in press; Smith response toward the emotional stimulus (bodily & Lazarus, 1990) is of particular with responses, motor responses, action tendencies, and respect to its categorical function. Core relational subjective feeling). Some aspects of these may be themes are categorical conceptualisations of emo- relatively hard-wired (e.g., a simple response like a tion-eliciting appraisals. Each core relational startle reflex), others mainly determined by cul- theme is specific to a given emotion and refers ture (e.g., a more complex appraisal of norm to a combination of a set of appraisal outputs (e.g., compatibility). Together, the integrated emotion high importance, high undesirability and other- category then may guide the perception and accountability define the core relational theme of categorisation of emotional stimuli, as will be ‘‘other-blame’’, which may elicit ‘‘anger’’). There- outlined in the remainder of this paper. fore, each core relational theme may be seen as a functional category that has the potential to elicit a specific emotion. EMPIRICAL STUDIES ON THE The subjective emotional feeling elicited by a PERCEPTION AND stimulus may be a further central feature in CATEGORISATION OF determining category membership. Stimuli that EMOTIONAL STIMULI evoke the same emotional response may thus be grouped together and treated as equivalent things, The theoretical considerations reviewed in the even when they are perceptually, functionally and first part of this paper point to important ques- theoretically different. When participants had to tions that can be addressed in empirical research categorise triads of concepts that shared both on how people actually perceive emotional sti- emotional and non-emotional relations (e.g., joke, muli. For example, as discussed above, a crucial speech, and sunbeam), participants for whom prediction of basic emotion theories is the uni- affective information was made salient by experi- versal and presumably innate organisation of fixed Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 encing a positive or negative mood used emotional emotion categories, whereas other theories em- response categories (i.e., grouped together joke phasise the flexibility, malleability, versatility, and and sunbeam), whereas participants in a neutral context-dependency of emotional processing. mood grouped the concepts into non-emotional Thus, in the next section, we will discuss what categories (joke and speech), indicating that the empirical research on the perception of emotional evoked subjective emotional feeling can be a stimuli can contribute to the question of the category-defining property (Niedenthal et al., relative contribution of bottom-up and top-down 1999). factors in emotional processing. Furthermore, we

384 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

will evaluate empirical results showing how meta-analysis showing that facial expressions of perception actually profits from having some emotion are correctly categorised across cultures stimuli tagged as emotional and others not, and with an accuracy of 58% (Elfenbein & Ambady, we will discuss how such studies can help to 2002). Similar results have been reported for the answer the question of what actually defines an categorisation of vocal emotional expressions. emotional stimulus, essentially by taking into Another meta-analysis (Juslin & Laukka, 2003) consideration what kind of emotional stimuli are showed that vocalisations of emotions are cor- prioritised in perception. To shed some light on rectly categorised across cultures with largely these questions, we will discuss primarily two above chance accuracy. Other studies have broad lines of research: (i) research focusing on furthermore demonstrated above-chance categor- qualitative effects of emotional stimuli, addressing isation of bodily expressions of emotion (Atkin- the question of how people categorise different son, Dittrich, Gemmell, & Young, 2004; de stimuli into emotional categories, and (ii) research Gelder, 2006). High intraindividual correlations focusing on quantitative effects of emotional have been observed for correct identification of stimuli, referring to how the emotionality of a different emotional signals in facial, prosodic and stimulus may modulate and transform perception. lexical channels. People who excel at correctly categorising facial expressions of emotion perform Qualitative effects of emotion on perception: similarly highly using other channels, leading to The categorisation of stimuli as emotional the suggestion that an amodal system might be involved in perceptual identification of various Is there universality in emotional categorisation? emotional expressions in different communication A central tenet of basic emotion theories is the channels (Borod et al., 2000). assumption that emotional stimuli, especially People are not only able to label emotional facial emotional expressions, are universally per- expressions with categories, but the actual percept ceived in a categorical manner. In two very of seems to be influenced by influential lines of research, Ekman (1972, 1992, category boundaries. Just as top-down category Ekman & Friesen, 1975) and Izard (1971) have information transforms the gradually changing investigated the universality of the recognition of continuum of wavelengths in a rainbow into the emotional facial expressions. In their studies, they perception of chunks of different colours (catego- asked participants to categorise facial displays of rical perception effect; Harnad, 1987), instances emotion into several basic emotion categories. For of emotional facial expressions that are morphed example, participants were shown a photograph of into each other along a continuum between two a person expressing prototypical facial configura- emotions (e.g., from happiness to fear), are tion of fear, and then were given a number of perceived as belonging to discrete categories response alternatives such as ‘‘fear’’, ‘‘happiness’’, (either happiness or fear). Moreover, pairs of ‘‘anger’’ or ‘‘disgust’’ to choose from. Alternatively, emotional faces that differ from each other by a they were asked to freely describe the emotion given physical amount on such a continuum can they recognised in the picture without being given be discriminated more accurately when the pairs labels. Ekman and Izard both found that their belong to two different emotion categories than Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 participants were able to correctly categorise the when they belong to the same category (Calder, facial expressions into a number of basic emotions. Young, Perrett, Etcoff, & Rowland, 1996; Etcoff This was the case even when the receiver (the & Magee, 1992; Young et al., 1997). Similar participant asked to categorise the expression) was evidence has been provided for a categorical from a different culture than the sender (the perception of vocal emotion expression (Laukka, person posing for the photograph) and was in fact 2005). These effects illustrate that, even though never exposed to the sender’s culture. These the participants’ task is not to assign the faces or results have been confirmed in a more recent voices to emotional categories, the incoming

COGNITION AND EMOTION, 2010, 24 (3) 385 BROSCH, POURTOIS, SANDER

information on facial and vocal emotion expres- emotional categories are no innate universals, but sion seems to be automatically transformed into (at least to some extent) shaped by top-down categories. Furthermore, these results allow some cultural factors. conclusions about the actual percept of the per- ceived stimuli, namely that there is a qualitative The role of context and top-down effects in emotional difference in how similar expressions actually categorisation. Alternative approaches to facial appear to a perceiver depending on whether or expression and its recognition have been devel- not they belong to the same emotional category: oped based on dimensional theories of emotion, two facial expressions that differ by an exact predicting that facial expressions are not cate- physical amount on a continuum between two gorised directly into specific basic categories, but expressions appear more distinct from each other convey values of valence and arousal, which are when they cross a category boundary, but more subsequently used to attribute an emotion to the similar when they do not cross such a boundary. face (Russell, 1997), and on appraisal theories of Results on the categorisation of emotional emotion, emphasising the link between appraisal expression have provoked a great amount of outcomes and facial expression patterns (e.g., debate (see, e.g., Ekman, 1994; Izard, 1994; Scherer, 1992). According to the latter view, the Russell, 1994, 1995). Main criticisms stemmed facial expression of a given emotion expresses a from some methodological aspects of the con- differential sequential and cumulative response ducted research. For example, the caricatural pattern based on a series of appraisal outcomes. nature of facial expression stimuli used in the Decoders should thus be able to recognise a facial research was questioned, which mainly showed expression of emotion from the outcomes of the extreme versions of facial expressions, which are pattern of cognitive appraisals that have produced rarely observed in daily life (Carroll & Russell, the emotion. Sander, Grandjean, Kaiser, Wehrle, 1997). Furthermore, people make more errors and Scherer (2007) tested the hypothesis that when they are not given forced-choice response operations involved in orienting the focus of alternatives, but have to respond freely. This attention (e.g., gaze direction) and operations suggests that available language-based emotion concerned with evaluation of events would inter- categories drive the answer in a top-down man- act in the decoding of facial emotions. They found ner. A third point of criticism concerns the fact that the perceived specificity and intensity of fear that, even though there is still above-chance and anger depended on gaze direction (direct gaze accuracy when sender and receiver come from for anger and averted gaze for fear; see also Adams different cultures, accuracy is reliably higher when & Kleck, 2003; Sander, Grafman, & Zalla, 2003). both come from the same culture (Elfenbein & Using a judgement paradigm, Scherer and Grand- Ambady, 2002; Juslin & Laukka, 2003), suggest- jean (2008) had people assign pictures of facial ing that emotional categories are to some extent expressions of emotions to underlying patterns of shaped by cultural factors and language-based appraisal (‘‘something unexpected has happened’’, representations. ‘‘I am in a dangerous situation and I don’t know Thus, consistent with the predictions of basic how to get out of it’’) and to basic emotion labels emotion theories, it has been shown that humans (‘‘surprise’’, ‘‘fear’’) and demonstrated similarly Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 are able to categorise prototypical facial, vocal and high success rates both for the appraisal criteria bodily expressions of emotion into discrete emo- categories and the basic emotion categories. tion categories with above-chance accuracy. How- Constructivist theories of emotion, appraisal ever, consistent with theoretical approaches theories and, to a lesser extent, also dimensional emphasising the role of culture- and language- theories underline the importance of context in based top-down factors, cultural knowledge has determining why and how a stimulus is perceived been shown to further improve performance on as emotional. Confirming the important role of categorisation tasks, leading to the conclusion that context, it has been shown that the same facial

386 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

expression can be interpreted as showing different proach, but rather compatible with several of emotional states (e.g., fear or anger, surprise or them. happiness)*and thus classified into different However, what can be concluded from the emotion categories*depending on the situational empirical results reviewed here is that contextual context that has been given to the observer top-down effects are extremely important for the (Carroll & Russell, 1996; Kim, Somerville, John- categorisation of a stimulus as emotional. Con- stone, Alexander, & Whalen, 2003; Kim et al., textual or cultural information strongly influence 2004; Russell & Fehr, 1987; Wallbott & Ricci- the outcome of categorisation. Contextual effects Bitti, 1993). For example, if a participant is shown do not just modulate late, high-level interpreta- a prototypical expression of anger, together with tion processes, but impact at the most basic levels the information that this person has just been in a of visual processing. It is nevertheless possible that frightening situation, the face will be categorised there is some innate or universal core that plays a as fearful. Contextual information influences role in the definition of the extension of already at very early perceptual levels. categories. For example, with regards to facial When subjects judged a facial expression of expressions of emotion there may be aspects that disgust presented in an anger context, not only are conserved genetically due to their high did they drop substantially in their categorisation adaptive functionality. It has been demonstrated accuracy (from 87% to 13% compared to a disgust that when subjects show facial expressions of fear, context), but their early eye movements followed they have a larger visual field, allowing for a more efficient scanning of the environment for threats, the visual scan path usually elicited by facial whereas when they pose expressions of disgust, expressions of anger (Aviezer et al., 2008). nasal volume and air velocity during inspiration Furthermore, perceptual memory encoding has are reduced, lowering the intake of potentially been shown to be influenced by conceptual repulsive substances (Susskind et al., 2008). Such knowledge: When participants viewed ambiguous features may be universal parts of the definition of facial stimuli (morphed faces depicting a blend of the extension of emotion categories, and one can two emotion categories) while category knowl- speculate that they may contribute to the above- edge about one of the emotions was made more chance performance in categorisation tasks that is accessible, participants later remembered the face not due to methodological artefacts and cultural stimuli in line with the conceptual knowledge that facilitation. However, emotional categories and was active during encoding (Halberstadt & Nie- emotional categorisation are also to a large extent denthal, 2001). These results highlight the role of shaped by top-down contextual and cultural contextual information in the perceptual categor- factors determined by language. Emotion cate- isation of emotional stimuli. gorisation serves as a rapid and reliable mechan- To summarise, research has shown that when ism for complexity reduction and response people are asked to classify expressions of emo- preparation, however, it is highly sensitive to tion, they are able to do so with high accuracy. situational and contextual factors. Similarly good performance is observed whether the classification is based on basic emotion

Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 categories (Elfenbein & Ambady, 2002), a dimen- Quantitative effects of emotion on sional system (Bradley & Lang, 1994; Russell & perception: The perception of emotional Fehr, 1987), or appraisal criteria (Scherer & stimuli Grandjean, 2008). Thus, even though results So far we have considered studies where partici- from categorisation or classification tasks are pants are asked to make direct/explicit categorisa- frequently taken as main evidence supporting tions of emotional stimuli, mainly expressions of basic emotion theory, they seem not to be emotion in different modalities. Whereas such unequivocally in favour of any theoretical ap- results may shed light on emotion categories, their

COGNITION AND EMOTION, 2010, 24 (3) 387 BROSCH, POURTOIS, SANDER

boundaries, how they are defined and how people within a search grid containing the target as well as use them when they are asked to, they do not a varying number of distracter stimuli, which may address a crucial question, maybe the most im- or may not share some similarities with the target portant one: What is the advantage of having stimulus. The task is either to indicate whether all special emotion categories, how does perception stimuli belong to the same object category or not profit from having some stimuli tagged as emo- (thus a categorisation task, but not an explicit tional and others not? In summarising some of the categorisation of ‘‘emotional’’ vs. ‘‘neutral’’)orto theoretical and empirical work reviewed above, one search for a predefined target. Typically, faster can conclude that categorisation serves complexity detection times are obtained when the target has reduction, whereas emotions serve the optimisa- some emotional value, such as an angry face among tion of adaptive behaviour towards stimuli that are neutral faces (Hansen & Hansen, 1988; O¨ hman, relevant for the needs, goals and well-being of the Lundqvist, & Esteves, 2001b) or a snake among organism. The role of emotion in perception and flowers (O¨ hman, Flykt, & Esteves, 2001a), indi- categorisation thus should ultimately be related to a cating that the emotional target is either identified perceptual prioritisation of categories of relevant in a preattentive manner or that processing re- stimuli in order to facilitate further processing and sources are very rapidly allocated toward its posi- response preparation. tion. The search advantage for emotional stimuli In the next part of the paper, we thus focus on was originally interpreted as a parallel search ¨ quantitative effects of emotional stimuli in percep- leading to a ‘‘pop out’’ effect (Ohman et al., tion, investigating how the emotionality of a 2001a; Treisman & Gormican, 1988). However, stimulus can modulate and transform perception, there is now increasing consensus that the search even when people are not consciously categorising process for emotional stimuli is essentially serial, the stimulus as emotional. Thus, we will consider but characterised by smaller increases in response studies where participants’ explicit task is not a time when more distractors are added (see, e.g., categorisation task, but, for example, to search for a Horstmann, 2007). Emotional stimuli have been specific picture in a display, or to identify rapidly shown both to speed up the orienting of attention presented words. Nevertheless, the tasks require and to prolong the disengagement of attention some implicit form of emotional categorisation and (Fox, Russo, & Dutton, 2002; Koster, Crombez, contain emotional and neutral stimuli. In these Van Damme, Verschuere, & De Houwer, 2004); paradigms, the emotionality of the stimuli mod- faster detection of an emotional target among ulates the efficiency with which the task is neutral distractors (compared with detection of a performed, suggesting interaction between emo- neutral target among emotional distractors) may tion and perception. By integrating these studies in thus be due to either faster orienting of attention to our review, we will be able to provide a more the target or faster disengagement from the neutral targeted overview of the diversity of subprocesses distracters. Studies using a full factorial design (including, e.g., threatening targets and threaten- and effects involved in perceptual processing of ing distractors) suggest that both attention capture emotional stimuli. Furthermore, studies on the and disengagement prolongation by emotional categorisation of emotion usually employ facial (or stimuli contribute to the search advantage for

Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 less often vocal and bodily) expressions of emotion, emotional stimuli in the visual search task (Flykt, whereas the studies we are going to present now 2005). also use other kinds of stimuli, such as emotional In the attentional blink paradigm (Raymond, words and affective pictures of scenes or objects. Shapiro, & Arnell, 1992), participants are pre- sented with a series of stimuli such as words or How does perception profit from having stimuli pictures at high presentation rates (rapid serial tagged as emotional? In the visual search task visual presentation, RSVP, around 10 stimuli per participants are instructed to search for a target second). Participants then have to identify one or

388 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

more of these targets. Any single target can be emotion theories state that rapid perceptual reported accurately, but reporting a second target processing is specific for threat stimuli that are is considerably impaired when the two targets are evolutionarily prepared (O¨ hman & Mineka, presented within a short interval (200500 ms). 2001), it has repeatedly been shown that onto- Impaired performance is though to reflect capa- genetically acquired threatening stimuli (such as city limitations which restrict access to awareness guns or knives) show similar effects (Blanchette, (Shapiro, Arnell, & Raymond, 1997). It has been 2006; Brosch & Sharma, 2005; Fox, Griggs, & shown that the deficit in performance is greatly Mouchlianitis, 2007). Furthermore, even though attenuated for emotional stimuli, which can be the visual search paradigm is frequently cited to reported with higher accuracy than neutral stimuli support preferential detection of threat-related when appearing as second target (Anderson, stimuli, and some early visual search studies found 2005). Conversely, the deficit in performance faster detection of threatening information when may increase for a neutral target that follows an comparing symbolic happy and angry faces (‘‘smi- emotional stimulus. These results indicate that lies’’ and ‘‘frownies’’; e.g., Eastwood, Smilek, & emotional stimuli are selected preferentially from Merikle, 2001), recently it has been argued that a perceptual temporal stream, thus facilitating the finding of preferential attention capture by processes leading to stimulus awareness. angry compared to happy faces is due to the lower Increased processing of emotional stimuli was relevance of happy faces compared to angry faces, furthermore demonstrated in a perceptual identifi- but that attention capture as such is driven by cation task (Zeelenberg, Wagenmakers, & Rotte- stimulus relevance in general, not by fear-rele- veel, 2006), where emotional words were vance (Brosch, Sander, Pourtois, & Scherer, 2008; presented for around 25 ms and masked immedi- Brosch, Sander, & Scherer, 2007). ately. Afterwards, subjects had to indicate the With regard to the empirical evidence for a word they had seen by choosing between two potential threat specificity in the visual search words. Word recognition was increased for both paradigm, results are quite mixed (Wolfe & positive and for negative emotional words com- Horowitz, 2004). Some studies report a search pared to neutral words. Interestingly, in a similar advantage for angry faces, that is sometimes task, when people were asked to recognise rapidly driven exclusively by the eye region (Fox & presented and masked images of snakes and Damjanovic, 2006), sometimes by the mouth spiders, recognition rates were correlated with region (Horstmann & Bauland, 2006). Other individual disgust sensitivity, whereas individual studies report a search advantage both for angry fear of spider correlated with the tendency to and happy faces (Williams, Moss, Bradshaw, & falsely report having perceived spiders (Wiens, Mattingley, 2005), or for happy compared to Peira, Golkar, & O¨ hman, 2008), showing that angry faces (Juth, Lundqvist, Karlsson, & O¨ hman, individual differences in emotional sensitivity 2005). Furthermore, faster detection of animals is measures can influence perception both objec- not specific to threat-related animals such as tively (improved perception) and subjectively snakes or spiders (O¨ hman et al., 2001a), but has (increased misperceptions). also been observed for cute, positively valenced Taken together, these results show that the animals (Tipples, Young, Quinlan, Broks, & Ellis, Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 emotional quality of a diverse range of stimuli 2002). Generally, the results from visual search (such as words, pictures, or faces) can be extracted studies for emotional stimuli do not indicate a rapidly under suboptimal processing conditions faster detection specific to threat stimuli, but and facilitate the further perceptual processing of rather seem to support the notion of a faster the stimulus (Phelps, 2006; Vuilleumier, 2005). detection of emotional stimuli in general (see Frischen, Eastwood, & Smilek, 2008, for a similar What is the defining ‘‘emotionality’’ criterion for argumentation). Similarly, the attenuation of the perceptual prioritisation? Although some basic attentional blink has been demonstrated both for

COGNITION AND EMOTION, 2010, 24 (3) 389 BROSCH, POURTOIS, SANDER

negative and for positive high-arousing stimuli the eyebrows in a threatening facial expression (Anderson, 2005; Anderson & Phelps, 2001; Keil (Aronoff, Barclay, & Stevenson, 1988) or the & Ihssen, 2004; Most, Smith, Cooter, Levy, & basic configuration of the baby schema (Lorenz, Zald, 2007). 1943). Taken together, the empirical evidence indi- cates that increased perception is not restricted to Are there differences between the effects of emotional fear-relevant stimuli, but is observed for stimuli words and emotional pictures? The various studies with both positive and negative valence, consistent reviewed in this paper used emotional and neutral with the assumption of a perceptual prioritisation words as well as images of emotional or neutral of highly relevant information (Brosch et al., scenes or expressions. With regard to the activa- 2008). tion of emotion concepts, it would be interesting to know whether there are fundamental proces- Is increased perception due to emotional effects or basic sing differences between emotional words and stimulus characteristics? Most researchers study- non-verbal displays of emotion. One might ing the preferential perception of emotional assume, for example, that words activate emotion stimuli assume that the prioritisation of the concepts more easily than other emotional stimuli, stimuli is due to the emotional quality of the and thus lead to stronger top-down effects on stimulus as assessed by the individual. Never- processing. On the other hand, one may argue theless, one cannot exclude that the effects are due that pictures represent stronger or ecologically to associated characteristics of stimulus or task more valid stimuli than words, and thus may lead (e.g., spatial frequencies, low level perceptual to stronger concept activation. The studies re- correlates) and not direct effects of the stimulus viewed so far do not give any definitive answers on emotionality (see, e.g., Cave & Batty, 2006). It this question. For example, studies investigating has been shown, for example, that the degree of the attentional blink have found rapid attentional attentional capture by an emotional stimulus does prioritisation of both emotional words (Anderson, not always correspond to the strength of affective 2005) and emotional pictures (Most et al., 2007), evaluations for the same stimulus when measured but no study has compared the perceptual effects by implicit tests such as affective priming (Purkis of the two types of stimuli. Linguistic studies in & Lipp, 2007). However, in other tasks, ratings of general find that pictures allow access to semantic emotional intensity correlate with degree of information more rapidly than words, as the latter response facilitation (Brosch et al., 2007). A role have to go through phonological processing first; of emotional processes is also supported by the only after a word string has been recognised as a findings that attentional biases can be modulated word will its semantic properties be accessed (see by individual state and trait differences related to Glaser, 1992, for a review). Some more direct emotion. For example, attentional bias toward evidence for stronger or more automatic concept threatening information is often enhanced in activation by emotional pictures than words comes people with specific phobia: attention is directed from studies showing that the emotionality of faster to pictures of snakes than spiders in snake a picture interferes with the affective categorisa- phobics, but vice versa in spider phobics (O¨ hman tion of words, whereas the emotionality of Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 et al., 2001a). Such individual differences strongly words does not (or only to a lesser extent) suggest that prioritised attention is determined by interfere with the affective categorisation of an appraisal of the emotional meaning and pictures (Beall & Herbert, 2008; De Houwer & personal relevance of a stimulus, rather than just Hermans, 1994). salient sensory features. Furthermore, it is im- portant to consider that the emotionality of a Which mechanisms underlie the prioritised perception stimulus may actually be conveyed by some very of emotional stimuli? The increased perception of simple perceptual features, such as the v-shape of emotional stimuli, shown across a wide range of

390 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

paradigms and methods, may depend both on DISCUSSION memory-based processes and an online evaluation Why do we need emotional stimulus of the stimulus. Emotional stimuli such as words categories? or objects may have stronger memory representa- tions than neutral ones (LaBar & Cabeza, 2006; The research that we have reviewed here shows Phelps & Sharot, 2008), probably due to the that emotion is a strong incentive for perception higher implicit or explicit goal relevance of such and that emotionally relevant words or images stimuli for the organism (see Levine & Edelstein, may produce both qualitative and quantitative in press). The emotional significance of a stimulus changes in the speed and amount of what is enhances the formation of long-term memory eventually perceived by the individual. People are traces, shown by better memory performance for able to categorise stimuli into emotional cate- emotional than neutral stimuli (Hamann, Ely, gories, be they based on basic emotions, dimen- Grafton, & Kilts, 1999). The stronger memory sional approaches, or appraisal criteria. The category boundaries affect the actual percept of representation may lead to a facilitated activation emotional stimuli, as shown by the categorical of the representations, which accounts for find- perception effect. Furthermore, independent of any ings such as the preferential identification of categorisation task, emotional stimuli are per- masked emotional words compared to neutral ceived preferentially. words (Zeelenberg et al., 2006) or the facilitated A special role for emotional stimuli in percep- selection of emotional words from a temporal tion is obviously useful, as emotional stimuli, i.e., stream of rapidly presented words (Anderson, stimuli that possess high relevance for the survival 2005). Furthermore, memory representations of and well-being of the observer, usually require emotional stimuli, either on the basis of individual rapid behavioural responses, a preferential percep- stimuli or on the basis of emotional stimulus tion being the first step in the co-ordination of categories, may already include emotional infor- such an adaptive response. Emotional categories mation in the sense of evaluation results, i.e., one allow the organism to rapidly organise the might already know from earlier interactions with processing of environmental information based a stimulus that it has high relevance, both via on the relevance of the information with regards explicit knowledge (e.g., when I see the face of a to current needs and goals. Just as one main person that I don’t like because I am aware that he function of categorisation in general is the rapid has been mean to me) and via implicit pathways access to and retrieval of a lot of information (e.g., when I smell a food for which I have a taste about the incoming stimulus, a function of aversion). In this sense the stored information acts emotional categorising seems to be the rapid as an evaluation shortcut, so that no new elaborate access to and retrieval of information that makes evaluation is necessary. Other forms of affective a quick adaptive response possible. Emotional evaluation cannot solely rely on memory pro- categorisation thus can be conceptualised as an cesses. Context-sensitive processes need to take automatic, adaptive ‘‘tag for high priority proces- into account the current situation as well as the sing’’ (see Yantis & Johnson, 1990).

Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 need or goal state of the organism and match it with the properties of the stimulus. This kind of How do we categorise emotional stimuli: On processing needs an online appraisal (see, e.g., the influence of bottom-up and top-down Moors, in press). Both kinds of processes may play mechanisms a role in the evaluation of the affective value of a One central, frequently reoccurring debate in stimulus. research on the processing, perception and

COGNITION AND EMOTION, 2010, 24 (3) 391 BROSCH, POURTOIS, SANDER

categorisation of emotional stimuli is centred opening of the eyes in a fear expression (Susskind around the question of the relative importance of et al., 2008), but furthermore includes culture- universal, biological bottom-up factors, as empha- specific, socially determined aspects. Emotional sised, e.g., by adherents of basic emotion theories, categorisation is furthermore based to a great versus culturally and socially determined top-down extent on available information about the current factors, as emphasised especially by constructivist situational context. theories. This question has recently returned to the Thus, whether and how a stimulus is perceived attention of emotion psychology due to a debate in as emotional is not static, fixed or invariant, but Perspectives in Psychological Sciences (Barrett, 2006a; critically depends on and fluctuates according to Barrett et al., 2007a; Izard, 2007; Panksepp, 2007, the person’s particularities (such as the current 2008). Based mainly on the argument that there is a mood or motivations) and specific context (e.g., lack of human neurophysiological evidence for situation, time, culture). Hence, a given stimulus discrete response patterning, Barrett promotes a can be emotional for one person, while being constructivist approach claiming that emotion perceived as carrying less or even lacking any categories are not natural entities, but man-made emotional meaning by another individual. concepts. The experience of emotions is under- Furthermore, the emotional meaning of a given stood as the categorisation of core affect, an stimulus may fluctuate for the exact same indivi- internal state describable only in terms of valence dual according to the specific context (situation and arousal, into language-based emotion cate- and time) in which this stimulus is encountered. It gories (Barrett, 2006a). In contrast, based on a large is mainly for that reason that it is not easily amount of animal data showing discrete emotional possible to compile a fixed list of ‘‘emotional response systems for a number of fundamental stimuli’’ that elicit an emotional effect in all behaviours (e.g., FEAR, RAGE, PLAY; Pank- people and on all occasions. A given stimulus sepp, 1998), Panksepp rejects the extreme con- becomes emotional for a person due to the structivist position and suggests that human individual interaction of that stimulus with the emotion researchers need to take into account perceiving organism, assessing the individual cross-species data indicating basic emotional sys- emotional relevance of the stimulus for the tems to understand the ‘‘primal sources of human person. To put it simply, a snake by itself is not emotional feelings’’ (Panksepp, 2007, 2008). an emotional stimulus, nor does it guarantee the What can the study of the perception of elicitation of an emotion, but it takes a snake and emotional objects contribute to this hotly debated somebody who is afraid of snakes to have an issue? Some of the principal claims of the debate emotion. There might be some stimuli that elicit can be evaluated under the light of the empirical highly similar emotional responses across all data and conceptual arguments that we discuss persons, for example a strong fear-eliciting sti- here. The data on categorisation tasks leave very mulus such as a painful electric prickling. How- little about the fact that the claim of a ever, even such extreme cases should be strong universality of emotion categories put conceptualised as reflecting an interaction be- forward in support of basic emotion theories is tween the person and the stimulus, as demon- not supported by the empirical evidence. The data strated by interindividual differences in Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 rather indicate a large role of culture and context perception related to factors such as race, sex, in an ongoing and flexible development of the catastrophism or level (Ploghaus et al., categories that we use to carve our environment. 2001; Sheffield, Biles, Orom, Maixner, & Sheps, Emotion categories seem to be learned and 2000). refined over time by integrating emotion-related Thus, categorisation, while acting as an effi- information that frequently occurs together. This cient mechanism for rapid complexity reduction, may nevertheless include some biological bottom- takes into account situational and contextual up aspects, such as adaptive responses like the factors. The outcomes of the rapid categorisation

392 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

mechanisms should not be mistaken for a modular into an emotion category that can be used to mechanism for emotional processing, which is structure the environment, guide perception, and restricted to a few basic categories. Whereas the give rise to a subjective feeling. Like constructivist perception of emotional stimuli can be easily approaches, can thus account for described in terms of basic categories or dimen- the richness and flexibility of the extension of the sions, the actual process that renders a stimulus emotional categories, as appraisal is not hard emotional must be conceived of as more complex, wired, but takes into account individual particula- highly flexible and context dependent. To explain rities and specific contexts (e.g., Frijda, 2007). the perception and categorisation of emotional Unlike constructivist approaches, however, it stimuli, it is thus not sufficient to rely on a few postulates more specific mechanisms that give basic, inflexible, hardwired categories. Emotional rise to the emotional quality that is categorised categorisation is modulated by our language (core affect in the case of Barrett, 2006, versus a capacities and available labels, and emotional response pattern of appraisal results, physiological categories seem to be similar to non-emotional reactions, motor expression, and action prepara- categories in that they reflect correlational struc- tion, in the case of Scherer, in press). From this tures that we experience in our environment. They perspective, effects like the rapid prioritisation of may be different from non-emotional categories, emotional stimuli by the perceptual system thus however, in that they integrate information about can be understood as embedded in the patterning the different components of emotions, such as of appraisal processes, action preparation and appraisal, action tendencies, bodily responses, physiological orienting responses, and may serve behavioural responses and changes in subjective to optimise perception even before a conscious feeling, and in that emotional categorisation categorisation has occurred. functions as a tag for high priority processing in the service of adaptive response preparation The relation of perception/cognition and toward relevant stimuli. emotion It is furthermore doubtful that a purely con- structivist position can account for results ob- Another important psychological debate is tained in research on the perception of emotional centred on the question of whether separate stimuli. Especially in tasks where multiple stimuli mechanisms exist for a dedicated processing of are presented or stimuli are presented in a rapid emotional stimuli or whether emotional and non- visual stream with one stimulus every 100 ms or emotional stimuli are processed by the same less, it is not clear how a preferential processing of cognitive mechanisms (Duncan & Barrett, 2007; emotional stimuli can be accounted for when the Eder, Hommel, & De Houwer, 2007; Lazarus, only emotional quality that is available for guiding 1984; Leventhal & Scherer, 1987; Pessoa, 2008; perception is an unspecified internal core affect Storbeck, Robinson, & McCourt, 2006; Zajonc, reflecting changes in a ‘‘neurophysiological bar- 1980). If emotion and cognition are treated as ometer’’ (Barrett, 2006a). As constructivist the- separate or dissociable psychological processes, ories mainly focus on the mechanisms underlying researchers would gain little insight on emotional the subjective experience of emotions, they do not perception by studying cognitive mechanisms of Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 formulate relevant predictions or explanations perception and categorisation. If, however, emo- about the topic of automatic perceptual prioritisa- tional stimuli are a special class of stimuli, which tion of emotional stimuli. are processed by the same cognitive mechanisms Componential appraisal theories of emotion as ‘‘neutral’’ stimuli, one can investigate cognitive (e.g., Scherer, in press) focus on the effects of mechanisms to elucidate a special role of emo- stimulus appraisal on the response patterning of tional stimuli in perception in the sense of physiological reactions, motor expression, and preferential treatment within cognitive processing action preparation, which then may be integrated (Moors, 2007).

COGNITION AND EMOTION, 2010, 24 (3) 393 BROSCH, POURTOIS, SANDER

The so-called ‘‘trilogy of the mind’’ (Hilgard, that later would assign them to the favoured 1980), separating cognition, emotion and motiva- outcome (Balcetis & Dunning, 2006). Perception tion into distinct entities, still seems to be very can be conceived as an interaction of bottom-up influential in the current literature. However, in sensory signals that are processed and integrated most theories, the question of whether affective by the sensory pathways, and top-down knowl- processes are dissociated from cognitive processes edge systems already present in the observer, (such as Zajonc’s, 1980, strong claim that ‘‘pre- which are used to structure and understand the ferences need no inferences’’) can be reduced to new incoming information (Bar, 2004, 2007; the question of whether sensory processes are Yantis, 1992). The emotional meaning of the considered as cognitive in nature or not (see also stimulus emerges from interactions between the Parrott & Schulkin, 1993). If one defines with stimulus content and the actual state of the Neisser (1967) cognitive processes as those pro- individual, during a proactive process. This con- cesses ‘‘by which the sensory input is transformed, clusion is consistent with a recent analysis of reduced, elaborated, stored, recovered, and used’’, neuroimaging studies suggesting that a segrega- it would still remain to be shown that sensory tion of the brain into ‘‘emotional’’ and ‘‘cognitive’’ input alone can indeed elicit emotion without any areas is not supported by the empirical evidence kind of transformation. (Pessoa, 2008). The evidence reviewed in this article shows The current review thus suggests that the that perceptual and emotional processing are perception of emotional stimuli does not seem highly intertwined. In contrast to the view that to require special dedicated kinds of ‘‘emotional’’ perception is an encapsulated process that is not processing mechanisms. The emotional quality of influenced by top-down influences such as ex- a stimulus rather seems to trigger a high-priority pectations or prior knowledge (Pylyshyn, 1999), processing mode inside an integrated cognitive the evidence suggests that perception is a highly affective system. dynamic, proactive process, which influences and is reciprocally influenced by other processes, including emotional processes, through dynamic CONCLUSIONS interactions. The perception of an emotional stimulus is both stimulus driven and concept In the present paper, we have shed some light on driven, i.e., the result is shaped by sensory the perception and categorisation of emotional information as well as by memory-based concep- stimuli, by integrating theoretical perspectives on tual information and online evaluation capacities. what makes a stimulus emotional and on how In extreme cases, emotional top-down concepts emotional categories are formed, as well as may even bias the perception of non-emotional empirical data illustrating how stimuli are cate- stimuli. For example, in one study participants gorised as emotional as well as how the perception were shown ambiguous figures that could be of emotional stimuli3 is prioritised. interpreted as a ‘‘B’’ or as a ‘‘13’’ and were told Emotional categorisation is a very important that if they saw a ‘‘B’’ they would be assigned to a mechanism by which we structure our environ- condition where they would taste orange juice, ment. Emotion is a strong incentive for percep- Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 whereas if they saw a ‘‘13’’ they would taste a tion, and emotional stimuli may produce both green, foul-smelling vegetable smoothie. Partici- qualitative and quantitative changes in the speed pants tended to report having seen the version and amount of what is eventually perceived by the

3 One should be aware that ‘‘emotional stimuli’’ as used in the studies presented here only very rarely elicit a full-blown emotion with an intense subjective feeling component. Reading the word ‘‘snake’’ will probably not very often be linked with experiencing strong feelings of fear. However, as shown by the evidence reviewed in this article, the perceptual processing of the word may still be increased compared to emotionally neutral words.

394 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

individual. People classify facial, vocal and bodily Aronoff, J., Barclay, A. M., & Stevenson, L. A. (1988). expressions of emotion with high accuracy, allow- The recognition of threatening facial stimuli. Journal ing them to rapidly assess the emotional state of of Personality and Social Psychology, 54, 647655. interaction partners. This classification can occur Atkinson, A. P., Dittrich, W. H., Gemmell, A. J., & according to ‘‘basic’’ emotion categories, dimen- Young, A. W. (2004). Emotion perception from dynamic and static body expressions in point-light sions such as valence or arousal, and appraisal and full-light displays. Perception, 33, 717746. criteria such as relevance or coping potential. Augoustinos, M., Walker, I., & Donaghue, N. (2006). Furthermore, emotional stimuli in general are Social cognition: An integrated introduction. London: prioritised in perception, are detected more Sage. rapidly and gain access to conscious awareness Aviezer, H., Hassin, R. R., Ryan, J., Grady, C., more easily than non-emotional stimuli. Susskind, J., Anderson, A., et al. (2008). Angry, Emotion categories are not determined uni- disgusted, or afraid? Studies on the malleability of versally or biologically, but are flexible and emotion perception. Psychological Science, 19, 724 continuously adjusted. Like other categories, 732. they reflect correlational structures experienced Balcetis, E., & Dunning, D. (2006). See what you want in the environment. However, they are special in to see: Motivational influences on visual perception. that they integrate different aspects of the emo- Journal of Personality and Social Psychology, 91, 612 625. tional response toward a stimulus (such as apprai- Bar, M. (2004). Visual objects in context. Nature sal components, core relational themes, action Reviews Neuroscience, 5, 617629. tendencies, bodily responses, behavioural re- Bar, M. (2007). The proactive brain: Using analogies sponses, subjective feeling). Furthermore, contex- and associations to generate predictions. Trends in tual top-down information is extremely important Cognitive Sciences, 11, 280289. for the categorisation of a stimulus as emotional. Barrett, L. F. (2006a). Are emotions natural kinds? This flexibility in categorisation helps to rapidly Perspectives on Psychological Science, 1,2858. and economically perceive the environment by Barrett, L. F. (2006b). Solving the emotion paradox: focusing on relevant information, but nevertheless Categorization and the experience of emotion. allows for adjusting the definition of what is Personality and Social Psychology Review, 10,2046. relevant at a given moment, congruent with the Barrett, L. F., Lindquist, K. A., Bliss-Moreau, E., Duncan, S., Gendron, M., Mize, J., et al. (2007a). view of emotion as a highly flexible interface Of mice and men: Natural kinds of emotions in the between stimulus input and adaptive response. mammalian brain? A response to Panksepp and Izard. Perspectives on Psychological Science, 2, 297 Manuscript received 24 March 2008 311. Revised manuscript received 10 October 2008 Barrett, L. F., Lindquist, K. A., & Gendron, M. Manuscript accepted 9 April 2009 (2007b). Language as context for the perception of First published online 19 June 2009 emotion. Trends in Cognitive Sciences, 11, 327332. Barsalou, L. W. (1985). Ideals, central tendency, and frequency of instantiation as determinants of graded REFERENCES structure in categories. Journal of : Learning, Memory, and Cognition, 11,

Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 Adams, R. B., Jr., & Kleck, R. E. (2003). Perceived 629654. gaze direction and the processing of facial displays of Barsalou, L. W. (1983). Ad hoc categories. Memory & emotion. Psychological Science, 14, 644647. Cognition, 11, 211227. Anderson, A. K. (2005). Affective influences on the Beale, J. M., & Keil, F. C. (1995). Categorical effects in attentional dynamics supporting awareness. Journal the perception of faces. Cognition, 57, 217239. of Experimental Psychology: General, 134, 258281. Beall, P. M., & Herbert, A. M. (2008). The face wins: Anderson, A. K., & Phelps, E. A. (2001). Lesions of Stronger automatic processing of affect in facial the human amygdala impair enhanced perception of expressions than words in a modified Stroop task. emotionally salient events. Nature, 411, 305309. Cognition and Emotion, 22, 16131642.

COGNITION AND EMOTION, 2010, 24 (3) 395 BROSCH, POURTOIS, SANDER

Berlin, B., & Kay, P. (1969). Basic color terms: Their boundary between preattentive and attentive vision. universality and evolution. Berkeley, CA: University Visual Cognition, 14, 629646. of California Press. Cohen, H., & Lefebvre, C. (Eds.). (2005). Handbook of Blanchette, I. (2006). Snakes, spiders, guns, and categorization in cognitive science. Amsterdam: Else- syringes: How specific are evolutionary constraints vier. on the detection of threatening stimuli? Quarterly Davidoff, J. (2001). Language and perceptual categor- Journal of Experimental Psychology, 59, 14841504. isation. Trends in Cognitive Sciences, 5, 382387. Bornstein, M. H., & Korda, N. O. (1984). Discrimina- Davidson, R. J., & Irwin, W. (1999). The functional tion and matching within and between hues mea- neuroanatomy of emotion and affective style. Trends sured by reaction times: Some implications for in Cognitive Science, 3,1121. categorical perception and levels of information de Gelder, B. (2006). Towards the neurobiology of processing. Psychological Research, 46, 207222. emotional body language. Nature Reviews Neu- Borod, J. C., Pick, L. H., Hall, S., Sliwinski, M., roscience, 7, 242249. Madigan, N., Obler, L. K., et al. (2000). Relation- De Houwer, J., & Hermans, D. (1994). Differences in ships among facial, prosodic, and lexical channels of the affective processing of words and pictures. emotional perceptual processing. Cognition and Cognition and Emotion, 8,120. Emotion, 14, 193211. Duncan, S., & Barrett, L. F. (2007). Affect is a form of Boster, J. (2005). Emotion categories across languages. cognition: A neurobiological analysis. Cognition & In H. Cohen & C. Lefebvre (Eds.), Handbook of Emotion, 21, 11841211. categorization in cognitive science. Amsterdam: Else- Eastwood, J. D., Smilek, D., & Merikle, P. M. (2001). vier. Differential attentional guidance by unattended Bradley, M. M., & Lang, P. J. (1994). Measuring faces expressing positive and negative emotion. emotion: The self-assessment manikin and the Perception & Psychophysics, 63, 10041013. semantic differential. Journal of Behavioural Therapy Eder, A. B., Hommel, B., & De Houwer, J. (2007). and Experimental Psychiatry, 25,4959. How distinctive is affective processing? On the Brosch, T., Sander, D., Pourtois, G., & Scherer, K. R. implications of using cognitive paradigms to study (2008). Beyond fear: Rapid spatial orienting towards affect and emotion. Cognition & Emotion, 21, 1137 positive emotional stimuli. Psychological Science, 19, 1154. 362370. Ekman, P. (1972). Universals and cultural differences in Brosch, T., Sander, D., & Scherer, K. R. (2007). That baby caught my eye. Attention capture by infant facial expressions of emotion. In J. K. Cole (Ed.), faces. Emotion, 7, 685689. Nebraska symposium on motivation (pp. 207 283). Brosch, T., & Sharma, D. (2005). The role of fear- Lincoln, NE: University of Nebraska Press. relevant stimuli in visual search: A comparison of Ekman, P. (1992). An argument for basic emotions. phylogenetic and ontogenetic stimuli. Emotion, 5, Cognition and Emotion, 6, 169200. 360364. Ekman, P. (1994). Strong evidence for universals in Calder, A. J., Young, A. W., Perrett, D. I., Etcoff, N. facial expressions: A reply to Russell’s mistaken L., & Rowland, D. (1996). Categorical perception critique. Psychological Bulletin, 115, 268287. of morphed facial expressions. Visual Cognition, 3, Ekman, P., & Friesen, W. V. (1975). Unmasking the 81118. face. Englewood Cliffs, NJ: Prentice-Hall. Carroll, J. M., & Russell, J. A. (1996). Do facial Elfenbein, H. A., & Ambady, N. (2002). On the expressions signal specific emotions? Judging emo- universality and cultural specificity of emotion

Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 tion from the face in context. Journal of Personality recognition: A meta-analysis. Psychological Bulletin, and Social Psychology, 70, 205218. 128, 203235. Carroll, J. M., & Russell, J. A. (1997). Facial expres- Ellsworth, P. C., & Scherer, K. R. (2003). Appraisal sions in Hollywood’s portrayal of emotion. Journal of processes in emotion. In R. J. Davidson, H. H. Personality and Social Psychology, 72, 164176. Goldsmith, & K. R. Scherer (Eds.), Handbook of Cave, K. R., & Batty, M. J. (2006). From searching for affective sciences (pp. 572595). Oxford, UK: Oxford features to searching for threat: Drawing the University Press.

396 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

Etcoff, N. L., & Magee, J. J. (1992). Categorical Horstmann, G. (2002). Facial expressions of emotion: perception of facial expressions. Cognition, 44, Does the prototype represent central tendency, 227240. frequency of instantiation, or an ideal? Emotion, 2, Flykt, A. (2005). Visual search with biological threat 297305. stimuli: Accuracy, reaction times, and heart rate Horstmann, G. (2007). Preattentive face processing: changes. Emotion, 5, 349353. What do visual search experiments with schematic Fox, E., & Damjanovic, L. (2006). The eyes are faces tell us? Visual Cognition, 15, 799833. sufficient to produce a threat superiority effect. Horstmann, G., & Bauland, A. (2006). Search asym- Emotion, 6, 534539. metries with real faces: Testing the anger-superiority Fox, E., Griggs, L., & Mouchlianitis, E. (2007). The effect. Emotion, 6, 193207. detection of fear-relevant stimuli: Are guns noticed Izard, C. E. (1971). The face of emotion. New York: as quickly as snakes? Emotion, 7, 691696. Appleton-Century-Crofts. Fox, E., Russo, R., & Dutton, K. (2002). Attentional Izard, C. E. (1994). Innate and universal facial expres- bias for threat: Evidence for delayed disengagement sions: Evidence from developmental and cross- from emotional faces. Cognition and Emotion, 16, cultural research. Psychological Bulletin, 115, 288 355379. 299. Frijda, N. H. (2007). The laws of emotion. Mahwah, NJ: Izard, C. E. (2007). Basic emotions, natural kinds, Lawrence Erlbaum Associates, Inc. emotion schemas, and a new paradigm. Perspectives Frischen, A., Eastwood, J. D., & Smilek, D. (2008). on Psychological Science, 2, 260280. Visual search for faces with emotional expressions. Juslin, P. N., & Laukka, P. (2003). Communication of Psychological Bulletin, 134, 662676. emotions in vocal expression and music perfor- Glaser, W. R. (1992). Picture naming. Cognition, 42, mance: different channels, same code? Psychological 61105. Bulletin, 129, 770814. Halberstadt, J. B., & Niedenthal, P. M. (2001). Effects Juth, P., Lundqvist, D., Karlsson, A., & .O¨ hman, A. of emotion concepts on perceptual memory for (2005). Looking for foes and friends: Perceptual and emotional expressions. Journal of Personality and emotional factors when finding a face in the crowd. Social Psychology, 81, 587598. Emotion, 5, 379395. Hamann, S. B., Ely, T. D., Grafton, S. T., & Kilts, C. Keil, A., & Ihssen, N. (2004). Identification facilitation D. (1999). Amygdala activity related to enhanced for emotionally arousing verbs during the attentional memory for pleasant and aversive stimuli. Nature blink. Emotion, 4,2335. Neuroscience, 2, 289293. Kim, H., Somerville, L. H., Johnstone, T., Alexander, Hansen, C. H., & Hansen, R. D. (1988). Finding the A. L., & Whalen, P. J. (2003). Inverse amygdala and face in the crowd: An anger superiority effect. medial prefrontal cortex responses to surprised faces. Journal of Personality and Social Psychology, 54, Neuroreport, 14, 23172322. 917924. Kim, H., Somerville, L. H., Johnstone, T., Polis, S., Harnad, S. (1987). Categorical perception: The ground- Alexander, A. L., Shin, L. M., et al. (2004). work of cognition. Cambridge, UK: Cambridge Contextual modulation of amygdala responsivity to University Press. surprised faces. Journal of Cognitive Neuroscience, 16, Harnad, S. (2005). To cognize is to categorize: 17301745. Cognition is categorization. In H. Cohen & C. Koster, E., Crombez, G., Van Damme, S., Verschuere, Lefebvre (Eds.), Handbook of categorization in B., & De Houwer, J. (2004). Does imminent threat

Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 cognitive science. Amsterdam: Elsevier. capture and hold attention? Emotion, 4, 312317. Hess, U., Philippot, P., & Blairy, S. (1998). Facial Kourtzi, Z., & DiCarlo, J. J. (2006). Learning and reactions to emotional facial expressions: Affect or neural plasticity in visual object recognition. Current cognition? Cognition & Emotion, 12, 509531. Opinion in Neurobiology, 16, 152158. Hilgard, E. R. (1980). The trilogy of mind: Cognition, LaBar, K. S., & Cabeza, R. (2006). Cognitive neu- , and conation. Journal of the History of roscience of emotional memory. Nature Reviews Behavioral Sciences, 16, 107117. Neuroscience, 7,5464.

COGNITION AND EMOTION, 2010, 24 (3) 397 BROSCH, POURTOIS, SANDER

Lang, P. (1995). The emotion probe: Studies of Journal of Experimental Psychology: General, 130, motivation and attention. American Psychologist, 50, 466478. 372385. .O¨ hman, A., Lundqvist, D., & Esteves, F. (2001b). The Laukka, P. (2005). Categorical perception of vocal face in the crowd revisited: A threat advantage with emotion expressions. Emotion, 5, 277295. schematic stimuli. Journal of Personality and Social Lazarus, R. S. (1984). On the primacy of cognition. Psychology, 80, 381396. American Psychologist, 39, 124129. .O¨ hman, A., & Mineka, S. (2001). , phobias, and Leventhal, H., & Scherer, K. R. (1987). The relation- preparedness: Toward an evolved module of fear and ship of emotion to cognition: A functional approach fear learning. Psychological Review, 108, 483522. to a semantic controversy. Cognition and Emotion, 1, Palmeri, T. J., & Gauthier, I. (2004). Visual object 328. understanding. Nature Reviews Neuroscience, 5, 291 Levine, L. J., & Edelstein, R. S. (in press). Emotion 303. and memory narrowing: A review and goal relevance Panksepp, J. (1998). : The founda- approach. Cognition and Emotion. tions of human and animal emotions. New York: Liberman, A. M., Harris, K. S., Hoffman, H. S., & Oxford University Press. Griffith, B. C. (1957). The discrimination of speech Panksepp, J. (2007). Neurologizing the psychology of sounds within and across phoneme boundaries. affects: How appraisal-based constructivism and Journal of Experimental Psychology, 54, 358368. basic emotion theory can coexist. Perspectives on Lorenz, K. (1943). Die angeborenen Formen mo¨glicher Psychological Science, 2, 281295. Erfahrung [The innate forms of potential experi- Panksepp, J. (2008). Cognitive conceptualism: Where ence]. Zeitschrift fu¨r Tierpsychologie, 5, 233519. have all the affects gone? Additional corrections for McClelland, J. L., & Rumelhart, D. E. (1985). Barrett et al. (2007). Perspectives on Psychological Distributed memory and the representation of Science, 3, 305308. general and specific information. Journal of Experi- Parrott, W. G., & Schulkin, J. (1993). What sort of mental Psychology: General, 114, 159188. system could an affective system be? A reply to Merikle, P. M., & Daneman, M. (1998). Psychological LeDoux. Cognition and Emotion, 7,6569. investigations of unconscious perception. Journal of Pessoa, L. (2008). On the relationship between emo- Consciousness Studies, 5,518. tion and cognition. Nature Reviews Neuroscience, 9, Moors, A. (2007). Can cognitive methods be used to 148158. study the unique aspect of emotion: An appraisal Phelps, E. A. (2006). Emotion and cognition: Insights theorist’s answer. Cognition and Emotion, 21, 1238 from studies of the human amygdala. Annual Review 1269. of Psychology, 57,2753. Moors, A. (2009). A review of theories concerned with Phelps, E. A., & Sharot, T. (2008). How (and why) emotion causation. Cognition and Emotion, 23, 625 emotion enhances the subjective sense of recollec- 662. tion. Current Directions in Psychological Science, 17, Moors, A. (in press). Automatic constructive appraisal 147152. as a candidate cause of emotion. Emotion Review. Ploghaus, A., Narain, C., Beckmann, C. F., Clare, S., Most, S. B., Smith, S. D., Cooter, A. B., Levy, B. N., Bantick, S., Wise, R., et al. (2001). Exacerbation of & Zald, D. H. (2007). The naked truth: Positive, pain by anxiety is associated with activity in a arousing distractors impair rapid target perception. hippocampal network. Journal of Neuroscience, 21, Cognition and Emotion, 21, 964981. 98969903. Murphy, G. L., & Medin, D. L. (1985). The role of Purkis, H. M., & Lipp, O. V. (2007). Automatic

Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 theories in conceptual coherence. Psychological Re- attention does not equal automatic fear: Preferential view, 92, 289316. attention without implicit valence. Emotion, 7, 314 Neisser, U. (1967). Cognitive psychology. New York: 323. Appleton-Century-Crofts. Pylyshyn, Z. (1999). Is vision continuous with cogni- Niedenthal, P. M., Halberstadt, J. B., & Innes-Ker, A. tion? The case for cognitive impenetrability of visual H. (1999). Emotional response categorization. Psy- perception. Behavioral and Brain Sciences, 22, 341 chological Review, 106, 337361. 423. .O¨ hman, A., Flykt, A., & Esteves, F. (2001a). Emotion Raymond, J. E., Shapiro, K. L., & Arnell, K. M. drives attention: Detecting the snake in the grass. (1992). Temporary suppression of visual processing

398 COGNITION AND EMOTION, 2010, 24 (3) PERCEPTION AND CATEGORISATION OF EMOTIONAL STIMULI

in an RSVP task: An attentional blink? Journal of Evidence for appraisal theories of emotion. European Experimental Psychology: Human Perception and Per- Journal of Cognitive Psychology, 19, 470480. formance, 18, 849860. Scherer, K. R. (1992). What does facial expression Roberson, D., Davies, I., & Davidoff, J. (2000). Color express? In K. Strongman (Ed.), International categories are not universal: Replications and new review of studies on emotion (Vol. 2, pp. 139165). evidence from a stone-age culture. Journal of Experi- Chichester, UK: Wiley. mental Psychology: General, 129, 369398. Scherer, K. R. (1994a). Emotion serves to decouple Rosch, E. (1975). Cognitive representations of seman- stimulus and response. In P. Ekman & R. J. tic categories. Journal of Experimental Psychology: Davidson (Eds.), The nature of emotion: Fundamental General, 104, 192233. questions (pp. 127130). New York: Oxford Rosch, E. (1978). Principles of categorization. In E. University Press. Rosch & B. B. Lloyd (Eds.), Cognition and Scherer, K. R. (1994b). Toward a concept of ‘‘modal categorization (pp. 2748). Hillsdale, NJ: Lawrence emotions’’. In P. Ekman & R. J. Davidson (Eds.), Erlbaum Associates, Inc. The nature of emotion: Fundamental questions (pp. Rosch, E., Mervis, C., Gray, W. D., Johnson, D. M., & 2531). New York: Oxford University Press. Boyes-Braem, P. (1976). Basic objects in natural Scherer, K. R. (2001). Appraisal considered as a process categories. Cognitive Psychology, 8, 382439. of multilevel sequential checking. In K. R. Scherer, Rosch Heider, E. (1972). Universals in color naming A. Schorr, & T. Johnstone (Eds.), Appraisal processes and memory. Journal of Experimental Psychology, 93, in emotion: Theory, methods, research (pp. 92 120). 1020. New York: Oxford University Press. Russell, J. A. (1991). Culture and the categorization of Scherer, K. R. (in press). The dynamic architecture of emotion: Evidence for the component process emotions. Psychological Bulletin, 110, 426450. model. . Russell, J. A. (1994). Is there universal recognition of Cognition and Emotion Scherer, K. R., & Grandjean, D. (2008). Facial emotion from facial expressions? A review of the expressions allow inference of both emotions and cross-cultural studies. Psychological Bulletin, 115, their components. Cognition and Emotion, 22, 789 102141. 801. Russell, J. A. (1995). Facial expressions of emotion: Schneirla, T. (1959). An evolutionary and develop- What lies beyond minimal universality? Psychological mental theory of biphasic processes underlying Bulletin, 118, 379391. approach and withdrawal. In Nebraska Symposium Russell, J. A. (1997). Reading emotions from and into on Motivation (pp. 142). Lincoln: Nebraska Press. faces: Resurrecting a dimensional contextual per- Shapiro, K. L., Arnell, K. M., & Raymond, J. E. spective. In J. A. Russell & J. M. Fernandez-Dols (1997). The attentional blink. Trends in Cognitive (Eds.), The psychology of facial expressions (pp. 295 Sciences, 1, 291296. 320). New York: Cambridge University Press. Shaver, E., Schwartz, J., Kirson, D., & O’Connor, C. Russell, J. A. (2003). Core affect and the psychological (1987). Emotion knowledge: Further exploration of construction of emotion. Psychological Review, 110, a prototype approach. Journal of Personality and 145172. Social Psychology, 52, 10611086. Russell, J. A., & Fehr, B. (1987). Relativity in the Sheffield, D., Biles, P. L., Orom, H., Maixner, W., & perception of emotion in facial expressions. Journal Sheps, D. S. (2000). Race and sex differences in of Experimental Psychology: General, 116, 223237. cutaneous pain perception. Psychosomatic Medicine, Sander, D. (2008). Basic tastes and basic emotions: 62, 517523. Basic problems, and perspectives for a nonbasic Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010 Smith, M. L., Cottrell, G. W., Gosselin, F., & Schyns, solution [commentary]. Behavioral and Brain P. G. (2005). Transmitting and decoding facial Sciences, 31, 88. expressions. Psychological Science, 16, 184189. Sander, D., Grafman, J., & Zalla, T. (2003). The Smith, C. A., & Kirby, L. D. (in press). Core relational human amygdala: An evolved system for relevance themes. In D. Sander & K. R. Scherer (Eds.),, The detection. Reviews in the Neurosciences, 14, 303316. Oxford companion to emotion and the affective sciences. Sander, D., Grandjean, D., Kaiser, S., Wehrle, T., & Oxford, UK: Oxford University Press. Scherer, K. R. (2007). Interaction effects of per- Smith, C. A., & Lazarus, R. S. (1990). Emotion and ceived gaze direction and dynamic facial expression: adaptation. In L. A. Pervin (Ed.), Handbook of

COGNITION AND EMOTION, 2010, 24 (3) 399 BROSCH, POURTOIS, SANDER

personality: Theory and research (pp. 609637). New Williams, M. A., Moss, S. A., Bradshaw, J. L., & York: Guilford Press. Mattingley, J. B. (2005). Look at me, I’m smiling: Storbeck, J., Robinson, M. D., & McCourt, M. E. Visual search for threatening and nonthreatening (2006). Semantic processing precedes affect retrie- facial expressions. Visual Cognition, 12,2950. val: The neurological case for cognitive primacy in Winkielman, P., Berridge, K. C., & Wilbarger, J. L. visual processing. Review of General Psychology, 10, (2005). Unconscious affective reactions to masked 4155. happy versus angry faces influence consumption Susskind, J. M., Lee, D. H., Cusi, A., Feiman, R., behavior and judgments of value. Personality and Grabski, W., & Anderson, A. K. (2008). Expressing Social Psychology Bulletin, 31, 121135. fear enhances sensory acquisition. Nature Wolfe, J. M., & Horowitz, T. S. (2004). What Neuroscience, 11, 843850. attributes guide the deployment of visual attention Thorpe, S., Fize, D., & Marlot, C. (1996). Speed of and how do they do it? Nature Reviews Neuroscience, processing in the human visual system. Nature, 381, 5, 495501. 520522. Yantis, S. (1992). Multielement visual tracking: Atten- Tipples, J., Young, A. W., Quinlan, P., Broks, P., & tion and perceptual organization. Cognitive Ellis, A. W. (2002). Searching for threat. Quarterly Psychology, 24, 295340. Journal of Experimental Psychology, 55A, 10071026. Yantis, S., & Johnson, D. N. (1990). Mechanisms of Treisman, A., & Gormican, S. (1988). Feature analysis in early vision: Evidence from search asymmetries. attentional priority. Journal of Experimental Psychol- Psychological Review, 95,1548. ogy: Human Perception and Performance, 16, 812 Vuilleumier, P. (2005). How brains beware: Neural 825. mechanisms of emotional attention. Trends in Young, A. W., Rowland, D., Calder, A. J., Etcoff, N. Cognitive Sciences, 9, 585594. L., Seth, A., & Perrett, D. I. (1997). Facial Wallbott, H. G., & Ricci-Bitti, P. (1993). Decoders’ expression megamix: Tests of dimensional and processing of emotional facial expression: A top- category accounts of . Cognition, down or bottom-up mechanism? European Journal of 63, 271313. Social Psychology, 23, 427443. Zajonc, R. B. (1980). Feeling and thinking: Preferences Whorf, B. L. (1956). The relation of habitual thought need no inferences. American Psychologist, 35, 151 and behavior to language. In J. B. Carroll (Ed.), 175. Language, thought and reality: Essays by B. L. Whorf. Zeelenberg, R., Wagenmakers, E. J., & Rotteveel, M. Cambridge, MA: MIT Press. (2006). The impact of emotion on perception: Bias Wiens, S., Peira, N., Golkar, A., & .O¨ hman, A. (2008). or enhanced processing? Psychological Science, 17, Recognizing masked threat: Fear betrays, but dis- 287291. gust you can . Emotion, 8, 810819. Downloaded By: [EBSCOHost EJS Content Distribution - Superceded by 916427733] At: 09:26 29 June 2010

400 COGNITION AND EMOTION, 2010, 24 (3)