The Distancing-Embracing model of the enjoyment of negative emotions in art reception

Abstract

Why are negative emotions so central in art reception far beyond tragedy? Revisiting classical aesthetics in the light of recent psychological research, we present a novel model to explain this much discussed (apparent) paradox. We argue that negative emotions are an important resource for the arts in general, rather than a special license for exceptional art forms only. The underlying rationale is that negative emotions have been shown to be particularly powerful in securing attention, intense emotional involvement, and high memorability, and hence is precisely what artworks strive for. Two groups of processing mechanisms are identified that conjointly adopt the particular powers of negative emotions for art's purposes. The first group consists of psychological distancing mechanisms that are activated along with the cognitive schemata of art, representation, and fiction. These schemata imply personal safety and control over continuing or discontinuing exposure to artworks, thereby preventing negative emotions from becoming outright incompatible with expectations of enjoyment. This distancing sets the stage for a second group of processing components that allow art recipients to positively embrace the experiencing of negative emotions, thereby rendering art reception more intense, more interesting, more emotionally moving, more profound, and occasionally even more beautiful. These components include compositional interplays of positive and negative emotions, the effects of aesthetic virtues of using the media of (re)presentation (musical sound, words/language, color, shapes) on emotion perception, and meaning-making efforts. Moreover, our Distancing-Embracing model proposes that concomitant mixed emotions often help integrate negative emotions into altogether pleasurable trajectories.

1. Introduction

Enjoyment associated with negative emotions in art reception has been a central issue in poetics and aesthetics ever since Aristotle's theory of tragedy (1961). Many plays, operas, poems, and films elicit feelings of sadness and melancholy, and horror movies are a very popular genre. Even the emotion of disgust plays an important role in contemporary art and entertainment, from photography and installation art to teenage comedies. As these examples indicate, our model of art reception is not limited to the highbrow arts, but applies to a broad range of media products, as well. Throughout this article, we use the term art in this broader sense.

Exposure to artworks is widely believed to be driven by hedonic expectations and actual hedonic reward (Arnold 1960; Berenbaum 2002; Dubé & Le Bel 2003; Knobloch-Westerwick & Keplinger 2006; Zillmann 1988). Moreover, only positive emotions or, put more generally, only positive affect (for the distinction between emotion and affect, see Russell [2003] and Scherer [2005]) is known to support hedonic, or friendly, approach behavior, whereas negative affect primes avoidance, defensive action responses, or hostile approach behavior (Norris et al. 2010). Self-sought hedonic exposure to negative emotions in art reception has therefore come to be called a “paradox” (Hume 1757). This introduction outlines the model we have developed to solve this paradox (see Fig. 1). Previously proposed solutions are reviewed in detail throughout the pertinent sections of this article as we spell out the individual components of our model. In a nutshell, what distinguishes the model presented here from earlier efforts can be highlighted through three points:

1. A. This is not a processing model for rare, exceptional cases in which enjoyment is associated with negative emotions. Rather, we suggest that negative emotions are quite generally a resource that is predestined for the arts' purposes. Section 2 derives this basic assumption from juxtaposing key findings of recent psychological research on negative emotions and key tenets of classical poetics. Psychological research suggests that negative emotions have a distinct potential for high intensity of subjective feeling, a powerful grip on attentional resources, and privileged storage in memory. Poetics suggests that these powers are precisely what the arts strive for. Hence, it appears that negative emotions and art reception may be a perfect match. Guided by this assumption, our model addresses the following questions: How can the arts adopt the particular powers of negative emotions to secure attention, intense emotional involvement, and high memorability without recipients experiencing the nonhedonic consequences of negative affect? Moreover, does negative affect also contribute to enjoyment because it is negative affect, and not just as a promoter of high emotional intensity and memorability? If so, which processing components and/or mechanisms support such hedonic benefits of experiencing negative affect in art contexts?

2. B. Responding to these questions, our model proposes two complementary processing factors, each of which includes several components. The first, already fairly well researched group of components (the “Distancing” factor) consists of the cognitive schemata of art, representation, and fiction. Situational activation of these schemata precedes the online processing and is maintained throughout it. (Reflecting this assumption, the color assigned to the Distancing factor in Figure 1 is also a background color for the Embracing factor and for Enjoyment.) The art, representation, and fiction schemata modify several important appraisal dimensions of negative emotions. As a result, they keep negative emotions at some psychological distance, thereby safeguarding the hedonic expectations of art reception against being inevitably compromised by the experience of negative emotions.

Figure 1. The Distancing-Embracing model. Solid rectangles indicate the processing components that are always involved in making negative emotions enjoyable, whereas the dashed rectangles indicate components that are likely to be only occasional contributors to such enjoyment (for details, see sects. 3 and 4 of this article).

On this basis, the second group of processing components (the “Embracing” factor) positively integrates, assimilates, or adopts the powers of negative emotions in the service of making art reception more emotional, more intense, more interesting, and, in the end, more rewarding. This factor, which is the prime focus of the present article, consists of five components. Whereas the activation of particular acquired genre scripts mostly precedes the online processing and is analogous to the top-down cognitive activation of the situation schemata of the Distancing factor, the other four components are operative during online processing, with the component of meaning construction being the only component that can well be operative past the end of the actual exposure. Here is a brief description: (1) compositional interplays of positive and negative feelings are hypothesized to render art processing richer in emotional variation and less prone to induce boredom than types of pleasure that involve exclusively positive feelings; (2) concomitant mixed emotions are hypothesized to serve as bipolar mediators for incorporating negative emotions into positive enjoyment; (3) aesthetic virtues of the artistic representation itself (for instance, the beauty of the wording, musical sound, or painterly execution in terms of color, shape, and abstract patterns) promote dimensions of liking and enjoyment that are based primarily on low-level perceptual processing, thereby creating a (more) positive environment for the processing of concomitant negative emotions; (4) processes of (symbolic) meaning construction can redeem negative emotions on the level of higher cognitive processes; and (5) the emotion-regulatory implications of particular acquired genre scripts, such as the power of the normative happy end of (prototypical) fairy tales, allow readers/listeners to go through the preceding dire situations of need and conflict in a less desolate way than could be expected in the absence of an established mental model of a fairy tale.

Thus, the model includes eight components that recruit a variety of cognitive and perceptual processes at different time points before, during, and after the exposure; some of these components are likely to be operative simultaneously during online processing. All of these components are hypothesized to exert specific emotion-regulatory/transformative effects on the processing of negative emotions. Notably, regardless of the assignment of the eight components to different time frames, our multicomponent model is not a component-process model. A detailed process model could not possibly cover music, literature, and the visual arts at the same time. After all, existing process models are invariably and for good reasons confined to particular art domains (Brattico et al. 2013; Jacobs 2015; Juslin 2013; Leder et al. 2004; Pelowski et al. 2016). The present article aims, however, precisely at identifying processing components that are hypothetically relevant for the hedonic processing of negative emotions across the art domains of music, literature, and – with a limitation mentioned at the very end of this Introduction –the visual arts.

Our Distancing-Embracing model shares with existing comprehensive models of art reception the general assumption that all aesthetic appreciation involves interactions of person variables; situational, cultural, and historical context; as well as stimulus characteristics (Bullot & Reber 2013b; Chatterjee 2013; Jacobsen 2006; Leder et al. 2004).

1. C. In the existing literature, the eight components of our two-factor model have been discussed largely as alternatives (to the extent that the individual components were actually discussed at all). In contrast, our model stipulates that at least one component of the Distancing factor and one component of the Embracing factor need to be combined to satisfactorily explain the pleasure in art reception that is associated with negative emotions. Indeed, in most cases, two, three, or even all five components of the Embracing factor are likely to play a role.

All eight components of our model identify features that distinguish the experience of negative emotions in art reception from their analogues in real-life contexts. Based on some of these components – most notably, the cognitive art and fiction framing – philosophers have suggested conceiving of art-elicited emotions as “make believe,” “as if,” or “quasi” emotions (Gaut 2003; Levinson 1997a; Mulligan 2009; Skulsky 1980; Solomon 2003). Despite the partial overlap with our argument, we refrain from using any of these terms for two reasons. First, by their very linguistic nature – and hence irrespective of how they are defined by individual authors – these terms evoke the notion (which we consider misleading) that art-elicited emotions may be somehow a species of inauthentic emotions. Second, the nuances of meaning distinctive of the above-quoted terms are highly debated even within the philosophy of emotions and barely known beyond this field. Therefore, projecting the eight processing components along which we distinguish art-elicited and ordinary emotions onto the above-quoted terms did not seem a promising effort for the purposes of the present article.

The present article focuses entirely on the core components of our model. Other important variables that are represented by the left-hand side of Figure 1, specifically, individual, historical, and cultural differences that are likely to influence the enjoyment of negative emotions in art contexts, lie beyond the scope of this article. Moreover, because our model exclusively accounts for the immediate experiential correlates of exposure to artworks that elicit negative emotions, we likewise do not discuss short-, medium-, or long-term functional benefits for psychological well-being that might be served by experiences of this type, such as catharsis (Aristotle 1961), exercising and developing one's cognitive and affective capabilities (Bloom 2011, p. 173; Kidd & Castano 2013; Kumschick et al. 2014; Nussbaum 2008), vicarious acting, coping with fears and terror management (Goldenberg et al. 1999), and homeostatic regulation (Sachs et al. 2015).

Notably, treatises on literature were the first to stipulate that their compositional trajectories should go through a broad variety of stages and highly varying emotional implications (cf. Menninghaus 2003). In line with this tradition, recent empirical research into the enjoyment of negative emotions in art reception also focuses preeminently on the arts that unfold in time, specifically, film, music, and literature. Temporal artworks require far longer exposure times and attention spans than purely spatial objects, such as painting and sculpture; accordingly, composition-guided emotional variation over longer periods tends to be greater in these temporal arts. Given that the present article not only proposes novel designs for empirically testing our hypotheses, but even more strongly also relies on reviewing available empirical evidence, we cannot but share this primary focus on the temporal arts. Still, we also include the relatively few studies on negative emotion processing in response to visual artworks in our treatment of the topic. Clearly, to the extent that photographs, paintings, and statues tell a story (cf. Picasso's Guernica) or represent objects or scenes of negative emotional valence, they also allow for investigating aesthetic trajectories that involve negative emotions (cf. Gerger et al. 2014). At the same time, we do by no means rule out that visual representations of beautiful humans, animals, landscapes, stills, and so forth can well be enjoyed as beautiful without necessarily co-evoking any negative emotional associations. In any event, systematic comparisons across all art domains, as well as across different types of presentation (live vs. media-based), are clearly called for to determine the extent to which our model fits all individual art forms equally well.

2. The goals of art, characteristics of negative emotions, and their strong role in artworks

Ever since Greek and Latin antiquity, treatises on rhetoric and poetics widely assume that artworks compete for attention, intense emotional involvement, and high memorability (Lausberg 1998; Quintilian 1920). Recent psychological research has provided evidence that many negative emotions have a particularly powerful grip on attentional resources, are intensely felt, are less prone to fade or return to baseline level than positive emotions, and often have access to privileged storage in memory (Cacioppo & Gardner 1999; Frijda 1988; Larsen & Prizmic 2008; Musch & Klauer 2003; Rozin & Royzman 2001; Vaish et al. 2008). The formula “bad is stronger than good” (Baumeister et al. 2001) provocatively summarizes this line of research. Putting together these strands of rhetoric/poetics and psychological theory, we conclude that negative emotions may actually not be an exceptional phenomenon licensed only in special art genres, but rather a key resource to be drawn upon by many, if not most, artistic efforts.

Conforming to this reasoning, the representational arts, particularly those based on narrative plot, routinely involve social conflicts and both represent and elicit negative emotions in response to such conflicts (Grodal 2007; Krämer & Witschel 2010; Scalise Sugiyama 2005). For dramas, novels, an epic poems this is true to an intriguing degree: Whereas failing marriages, unhappy love, long separations, adultery, betrayed friendship, and the like are routinely represented in great detail, we know of no novel that extensively depicts 20 years of happy married life, let alone the lifelong happy marriage of an elderly couple. Moreover, because the composition of narratives of all kinds is typically far more condensed than the unfolding of real-life events over a longer trajectory (Mar & Oatley 2008), the underlying conflicts and corresponding negative emotions tend to become more pronounced in artistic representations. Poems, too, thrive at least as often on sad or even desperate feelings of uncertainty or negative certainty regarding the responsiveness of the beloved rather than on happy feelings of having one's love reciprocated. Correspondingly, Batteux noted in his influential treatise The Fine Arts Explained through a Single Principle: “Artists succeed much more easily with unpleasant objects in the arts than with pleasant ones” (1746/2015, p. 48). To be sure, many comedies and popular novels are precisely about happy love relations, yet they, too, typically fade out – just like all fairy tales invariably do – once a state of happiness has been reached, and do not depict the affectively positive states in nearly as much detail as the preceding states of uncertainty, conflict, and suffering.

Some authors have discarded the notion that artworks in fact involve recipients in negative emotions. They stipulated that recipients may only erroneously report experiencing (expectable) negative emotions in response to artworks with negative emotional implications, but actually not feel any relevant level of such emotions (cf. Kivy 1991, Ch. 8; Krämer & Witschel 2010) because exposure to artworks is categorically different from ordinary “real-life” contexts. The majority of the studies on responses to horror films (Andrade & Cohen 2007), sad music (Vuoskoski & Eerola 2012), sadly moving films (Hanich et al. 2014), and disgusting images (Wagner et al. 2014) have relied exclusively on subjective self-reports and, therefore, cannot rule out this possibility. However, several studies on sad music (Lundqvist et al. 2009), affectively negative pictures (Gerger et al. 2014), and anger-inducing performances (Wagner et al. 2016) have also reported objective physiological responses of the autonomic nervous system (ANS) that were in line with the patterns expectable for genuine episodes of the respective negative emotions. All of these studies converge in showing (a) that negative emotions are not just represented in plots and expressed by characters, but are also, at least partly, felt by audiences (cf. also Juslin 2013; Mills 1993), and (b) that such felt negative emotions appear to covary with aesthetic liking and positive enjoyment.

Regarding instrumental music, the historical and theoretical discourse is far less focused on the apparent paradox of negative emotions than it is for literature. Many empirical studies emphasize the role of positive emotions (Blood & Zatorre 2001; Hunter et al. 2008; 2010; Keller & Schubert 2011; Salimpoor et al. 2011; Schellenberg et al. 2008; Witvliet & Vrana 2007; Zentner et al. 2008). Still, negative emotions, especially sadness, have also been reported to be both expressed by purely instrumental music and felt by listeners (Hunter et al. 2010; Schellenberg et al. 2008), and there is substantial evidence that music-elicited feelings of sadness correlate positively with liking the pertinent pieces of music (Garrido & Schubert 2011; Panksepp 1995; Panksepp & Bernatzky 2002; Schellenberg et al. 2008; Schubert 2013; Taruffi & Koelsch 2014).

Research into different types of pleasure has shown that intellectual and emotional pleasures (including those of art reception) routinely “encompass negative emotions like sadness … and positive emotions that entail complex appraisal” (Dubé & Le Bel 2003, p. 291; see also Berenbaum 2002). In contrast, physical pleasures tend to be least affected by concomitant negative emotions (Dubé & Le Bel 2003). This provides another piece of evidence in favor of the assumption that the inclusion of negative emotions should not be considered an exceptional case of art reception, but rather constitutes a general distinctive trait of it.

But do artworks that elicit higher levels of negative emotions indeed support more intense emotional responses and are also more memorable than artworks not involving negative emotions, or involving them only to a lower degree? No such comparisons have been performed to date in any systematic fashion. Still, throughout this article, we report individual pieces of evidence in favor of this assumption. To start with, we refer to a particularly conspicuous example. Not coincidentally, it comes from research into highly intense emotional responses to artworks: those associated with chills and goosebumps (Benedek & Kaernbach 2011; Blood & Zatorre 2001; Goldstein 1980; Grewe et al. 2007; Panksepp 1995; Rickard 2004; Salimpoor et al. 2011; 2009). Activation of chills and goosebumps in listening to music or watching a film is accompanied both by increased electrodermal activity as indicative of emotional arousal of the autonomic nervous system and a strong activation of the primary reward network (Blood & Zatorre 2001; Salimpoor et al. 2011); physiological arousal reaches maximal levels when goosebumps co-occur with emotional tears (Wassiliwizky et al. 2017a). Importantly, recent research has shown that chills and goosebumps in art contexts are highly likely to be not just “peak emotional responses” of an unspecific nature, but physiological correlates of states of being emotionally moved (Benedek & Kaernbach 2011; Wassiliwizky et al. 2015). This emotion state, in turn, has been shown to routinely involve interplays, or blends, of positive and negative emotional ingredients, not just in its sad, but also in its joyful variant (Menninghaus et al. 2015b).

Moreover, in a free-recall task in which participants were asked to remember feelings of being moved in response to real-life and fictional events (the latter being nearly always related to artworks), episodes of being joyfully moved were far more frequently recalled for the real-life memories than episodes of being sadly moved; inversely, memories related to fictional artworks showed a strong bias toward the sad variant of being moved (Menninghaus et al. 2015b; for more details on sadness and being moved, see sect. 4.2.1). Sadly moving films and narratives were also rated as affecting the audience more intensely than joyfully moving ones (Goldenberg et al. 1999; Menninghaus et al. 2015b; Wassiliwizky et al. 2015). Finally, experiencing chills and goosebumps in emotionally moving art context also has strong memory effects: Many recipients not only remember which artworks elicited these responses; they even remember the specific moment in the respective artworks that repeatedly elicited them (Blood & Zatorre 2001; Panksepp 1995; Salimpoor et al. 2011; Sumpf et al. 2015). Hence, these emotional responses can be considered as prime examples of high emotional intensity, high levels of aesthetic enjoyment, and high memorability as associated with artworks, the processing of which includes marked levels of negative affect. (Because the respective studies relied largely on participants' choices of artworks based on their memories of prior exposure, they moreover show that the power of familiarity to predict aesthetic liking [Bornstein 1989; Reber et al. 1998; Zajonc 1968] also extends to cases in which aesthetic liking is specifically associated with concomitant negative feelings.)

In a similar vein, the peak-end rule – according to which the mean of peak and end affect is a better predictor of post hoc liking ratings for an affective episode than the average of all continuous ratings (Do et al. 2008; Geng et al. 2013; Kahneman et al. 1993) – has been successfully applied to art reception in a way that implicitly relies on the very hypothesis we advocate here. The respective studies (Rozin et al. 2004; Schäfer et al. 2014) collected peak and end ratings not for pleasure and/or displeasure, positive and/or negative affect, but for intensity and, hence, for an experiential quality that applies equally to positive and negative affect. This choice of rating item implies the notion that not only positive, but also negative affect can equally contribute to overall liking if, and to the extent that, it supports felt emotional intensity.

Summing up, important characteristics of negative emotions can be theoretically conceived as predestined resources for the artists' efforts to produce artworks that powerfully and pleasurably affect recipients. A survey of the plots of literature, of research on music, and of studies on intellectual pleasures altogether supports the notion that negative emotions are not a liability or an exceptional license, but rather a resource on which these pleasures rely. Moreover, there is evidence that negative emotions are not mere misattributions, but that recipients actually feel them and that this is beneficial rather than detrimental for the intense emotional involvement, the highly rewarding quality, and the high memorability of art reception. At the same time, the emotional powers of negative emotions are by their very definition tied to experiencing negative affect and not just affective intensity. Thus, access to the distinctive art-compatible powers of negative emotions does come at a price, and the question arises how the arts can reap the benefits of negative emotions without compromising their ultimate goal. Our response to this question is the Distancing-Embracing model as developed in Sections 3 and 4.

3. Cognitive mechanisms that keep negative emotions at a psychological distance: The Distancing factor

The present section spells out the first pillar of our model as outlined in Figure 1: the Distancing factor. Specifically, we focus on the effects of the situational activation of three cognitive schemata or frames: those of art, representation, and fiction (for the schemata and cognitive framing theory in general, see Abelson 1981; Brewer & Nakamura 1984). We discuss evidence for the following hypothesis: Activation of these cognitive schemata/frames should keep felt negative emotions at some psychological distance (= primary effect) and thereby open experiential spaces in which negative emotions are not inevitably incompatible with art-specific expectations of hedonic reward (= secondary effect). Following the construal-level theory of psychological distance (for this concept, see Trope & Liberman 2010; Trope et al. 2007), we assume that the baseline, or reference point, for psychological distance is the art recipient's “me” in the “here” and “now” and that the various dimensions of psychological distance discussed below tend to be similar in effect (see also Cupchik 2002; Hanfling 2000).

Because the cognitive schemata of representation and fiction can be activated far beyond the confines of art, we define the three schemata under consideration separately. Notably, the three schemata are not mandatory for all arts: Instrumental music is mostly neither fictional nor representational, and the same holds for other forms of art such as abstract paintings and many experimental films. By contrast, the cognitive schema for reading literature and viewing/hearing films, theatrical plays, and operas typically simultaneously encompasses the schemata of art, representation, and fiction.

3.1. The art schema

Art reception typically implies an ongoing situational awareness that one is reading a book or watching a movie or listening to a piece of music rather than being involved in ordinary action or communication contexts (for applications to art contexts see Brewer & Lichtenstein 1982; Dixon & Bortolussi 2009; Hoeken & Van Vliet 2000; Leder et al. 2004; Magliano et al. 1996; Mandler 1984; Schubert 2016; Visch & Tan 2008; Zwaan 1994). This situation concept includes the notion of personal safety, because nothing represented on stage or in a book can directly harm viewers and readers (but see Hanich 2014). Moreover, readers, listeners, and viewers are in control of the situation in that they typically both seek out and continue or discontinue their exposure to a work of art in a self-motivated way (Andrade & Cohen 2007; Apter 1992; 1993; Bartsch et al. 2008; Bloom 2011, p. 196; Rozin et al. 2013; Tan 2008). This control thesis (cf. Eaton 1982; Morreall 1985; Witasek 1904, pp. 116–17; for a critique see Iseminger 1983; Packer 1989; Smuts 2009a) is in line with experimental findings that tolerance for pain is substantially higher if one has the power to stop a painful treatment at one's will and, thus, is in control of the situation (Litt 1988).

Projecting the features of safety and control onto a cognitive appraisal account of emotions (Scherer 2005), all emotions experienced in an art-framing context should be more “self-intended” than analogous emotions in a different framing, because they are typically intentionally self-sought. Moreover, art-elicited emotions should entail little direct personal goal relevance and goal conduciveness for immediate practical purposes (Hunter & Schellenberg 2010; Scherer 2012) and should not challenge our coping potential (apart from demands of cognitive understanding) or stimulate strong action responses (cf. Meinong 1902/1977; Scherer 2005). All of these cognitively distinctive features of art-elicited negative emotions should be reflected in how these emotions are subjectively experienced (Scherer 2005; Scherer et al. 2006) and should specifically work in favor of reducing their potential adverse effect on hedonic processing. Thus, the psychology of emotions clearly predicts differences in emotional responses dependent on the ontological nature of the emotion elicitors. Philosophical theories have discussed comparable distinctions (Gaut 2003; Levinson 1997b; Mulligan 2009; Skulsky 1980; Walton 1990).

3.2. The cognitive schema of representation: Effects of temporal, spatial, and cultural distancing

Representations typically refer to events or scenarios that are not co-extensive in time and/or space with what they represent, with live media coverage overcoming the distance in time, but not in space. Moreover, representations typically highlight some of the perceptual features of an event at the expense of others, and writing even suppresses all direct visual, acoustic, and olfactory input features of the real or fictional events to which it refers. Once the temporal and/or spatial distance becomes large enough, another factor almost invariably comes into play, namely, the cultural difference of contexts. Thus, representations support only distanced, indirect, and –compared with their real antecedents – incomplete exposure. This should, in principle, work in favor of a psychologically more distanced response, as predicted by the construal-level theory of psychological distancing (cf. Trope & Liberman 2010; Trope et al. 2007). For example, a temporal distance of only a few weeks has been shown to make it far more likely to find humorous aspects in the media coverage of a deadly hurricane (McGraw et al. 2012; 2014).

3.3. Distancing effects as a result of fictional status

A further dimension of distancing is exclusively involved in a subgroup of representations, namely, fictional (vs. factual) representations, including fictional artworks. The ontological shift into the realm of fiction has long been considered to work in favor of greater tolerance for and enhanced enjoyment of artworks that elicit negative emotions (cf. de Fontenelle 1692/1968; Zelle 1987, p. 162). Cognitively framing an act of murder or a scene of violent mutilation as fictional provides an awareness that no real person (or animal) has been physically harmed; this should alter emotional responses in comparison to witnessing comparable real acts. To be sure, the arts clearly thrive on the human propensity not to consistently maintain a clear-cut distinction between imagination, fiction, and belief systems, on the one hand, and reality, on the other (cf. Bloom 2011, pp. 155–76). However, even when we are immersed or absorbed in a story world (Busselle & Bilandzic 2008; Green et al. 2008; 2012; Kuijpers et al. 2014; Zwaan 1999), our mental situation model is likely to retain at least some background awareness of this important ontological distinction (Tan 2008). This may, however, apply to children to a lesser degree (Weisberg et al. 2013).

In most cases of narrative artworks and media products, the fiction framing is likely to be largely co-extensive with the appraisal profile of the art framing sketched above. However, studying nonfictional art, non-art fiction, and fake documentaries (such as Orson Welles's radio production based on H. G. Wells's The War of the Worlds, or The Blair Witch Project) could yield additional interesting perspectives on both the overlaps and the differences between the art-versus-reality framing, on the one hand, and the fiction-versus-reality framing, on the other. There are even clear cases of fascination when viewing live footage of real disastrous events (such as the eruption of volcanos, major tornados, tsunami waves, plane crashes, and other accidents) and gigantic crimes (such as the attack on the Twin Towers in 2001), with no temporal, cultural, or art- or fiction-based distancing mechanisms and real humans involved (cf. Rimé et al. 2005). Documentary tragedies that go beyond mere live coverage of disastrous events have likewise been argued to yield some sort of pleasure, at least if a number of constraints are met (Friend 2007). In any event, even in live footage, where no distancing effects of fiction and temporal distance are involved, some spatial distance – sufficient to support the precondition of the viewer's personal safety – is still required, and the media-transmitted sensory impression (loudness, olfactory sensations, etc.) is also different from experiencing, for example, a real earthquake.

3.4. Empirical evidence

Whereas we theoretically introduced the art, representation, and fiction schemata as distinct components of our model, we review the pertinent empirical evidence conjointly. We do so because the few available studies on the distancing effects of the cognitive art schema mostly draw on representational and typically also fictional arts, thereby ruling out a strict separation of the three cognitive schemata. Some of the assumptions outlined in sections 3.1 through 3 have been tested in empirical studies. Episodes of being sadly moved when responding to fictional events have been shown to score significantly higher for the appraisals “self-sought,” “self-caused,” and “power to modify consequences” and lower for the appraisal “undesired outcome” than episodes of responding to real events (Menninghaus et al. 2015b).

A study that presented photographs of disgusting matter as either art photography or documentary photographs made for purposes of hygiene instruction found higher levels of positive affect in the art-framing group (Wagner et al. 2014); at the same time, reported feelings of disgust did not differ between the two conditions. Accordingly, a study on perceived sadness and anxiety in a fiction-versus-nonfiction framing did not find any differences for the two conditions (Goldstein 2009). This suggests that the distancing effect of the art framing does not convert, let alone erase, negative emotional responses and need not even reduce the felt intensity of these responses to make them (more) compatible with positive enjoyment. Similarly, a study on moral feelings (McGraw & Warren 2010) reported that a psychologically distancing framing rendered a moral violation (a man having sexual intercourse with a chicken) more benign and even somewhat amusing, even though feelings of disgust remained at the same level as for participants who were not primed in a psychologically distancing manner.

Using both pictures of the International Affective Picture System and genuine artworks and targeting a variety of both positive and negative emotions, a study by Gerger et al. (2014) yielded similar results. Although negative valence ratings did not differ for the two framings, pictures of negative valence were again more aesthetically liked in the art-framing condition. Another study (Wagner et al. 2016) reported an analogous positivity/enjoyability effect of an art versus non-art framing for an elaborate identical anger-inducing treatment of participants. Earlier studies had already shown similar effects of cognitive framings on emotional processing (Lazarus 1964; Lazarus & Alfert 1964; Lazarus et al. 1962; 1965; Legrand & Apter 2004). However, these studies did not specifically investigate the role of an art versus non-art framing, but of other mental framings. Finally, a recent electroencephalography study (Van Dongen et al. 2016) yielded neuroscientific evidence for implicit emotion regulation activated through an art framing.

3.5. Downsides of emotional distancing mechanisms

Classical aesthetics suggests that emotional distancing may be disadvantageous for positive emotional content: “The effects of imitation, so advantageous for unpleasant objects, work completely against pleasant ones for the same reason. The impression made by art is weakened …. Thus, all other things being equal, the heart must be much less satisfied by pleasant objects in the arts than it is by unpleasant ones” (Batteux 1746/2015, p. 48). For example, regarding beautiful humans or landscapes, it may in many cases be more pleasurable to behold the real objects rather than their painterly representations. Being part of a joyful festive event may also produce more intense joy than reading a description or looking at well-taken photographs of the event. Correspondingly, photographs of persons and genuine visual artworks that were both perceived as positive in affective valence received lower ratings for positive valence when presented in an art framing compared with a reality framing (Gerger et al. 2014); at the same time, aesthetic liking ratings were at the same level for both framings. Only the pictures with negative emotional valence profited from the art framing: They received significantly higher liking ratings, even though the negative valence ratings and facial electromyography-based measures of positive and negative affect remained virtually unchanged. Thus, Batteux was right: The art framing, as compared with the non-art framing, yielded asymmetrical effects on pictures of positive and negative valence such that the “impressions made by art” were more positively appreciated in the case of pictures of negative valence.

Future research might investigate whether, and to what extent, different dimensions of distancing – spatial, temporal, historical, symbolic, fiction based – exert different effects on negative emotion processing, the extent to which these effects are cumulative, and whether there are ideal levels of distancing as compared with over- or underdistancing (cf. Bullough 1912; Scheff 1979, Ch. 5).

4. How the arts of selecting, combining, and formally elaborating constituent parts of artworks promote the enjoyment of negative emotions: The Embracing factor

Psychologically distancing effects as a result of the cognitive schemata of art, representation, and fiction do not by themselves explain why the arts might not be better advised to wholly avoid negative emotions and to focus exclusively on beauty and concomitant positive emotions. The activations of these cognitive situation schemata exclusively secure preconditions for the enjoyment in question. Additional psychological mechanisms of a more positive enabling nature are needed that take advantage of this precondition and actually adopt the powers of negative emotions for pleasurable purposes (cf. Andrade & Cohen 2007). The present section introduces and discusses the five processing components that make up this positive Embracing factor of our model.

Projected onto the framework of poetics (Lausberg 1998; Quintilian 1920), the first two components bear on the dimensions of inventing/selecting the constituent parts of an artwork (inventio) and combining them into a well-composed temporal or spatial order (dispositio). The third bears on the material execution of artworks in terms of wording (elocutio), sound patterns, coloring, and so forth. A fourth component (which hermeneutics added to the system of poetics) concerns processes of seeking or constructing meaning (interpretation) in response to artworks. The fifth component is another classical aspect of poetics, namely, the emotion-regulatory power of specific acquired genre scripts, that is, the anticipatory adjustment of our emotional expectations and processing routines depending on whether we read a text known to be a tragedy, a fairy tale, a satire, or some other genre.

4.1. Interplays of positive and negative emotions

In this subsection, we advocate the following hypothesis: Composition-driven trajectories of aesthetic processing (cf. Fitch et al. 2009) that involve negative emotions are conducive to enjoyment because the pleasure taken in the beautiful representation of wholly positive and beautiful objects and narratives tends to be less intense, profound, and self-supportive and more prone to induce boredom than pleasure that includes a dynamic interplay of positive and negative emotional responses. An earlier version of this hypothesis was put forward by Kant: “What makes theater plays (whether tragedies or comedies) so enticing? The fact that certain difficulties emerge in all of them – anxiety and perplexity between hope and joy – so that the interplay of opposite feelings sets the mind of the spectator in motion” (Kant 1798/1996, p. 232, emphasis added). Other treatises of classical philosophical aesthetics have similarly argued that interplays of positive and negative feelings are aesthetically superior to a purely and thoroughly positive affective content and tonality (Mendelssohn 1759/1991; 1761/1997, p. 143; Wezel 1785/1971).

Zillmann's (2006) theory of dramatic plot and, specifically, of the “excitation transfer” from the peak moment of threat and uncertainty to the peak moment of a happy end can be understood as a genre-specific variant of the general hypothesis of a pleasurable interplay of positive and negative emotions. Solomon's opponent-process theory even proposes that not just artworks, but all emotion-eliciting events give rise to both negative and positive (i.e., opposite) affective processes (Solomon 1980; Solomon & Corbit 1974). As a consequence of its generality, this theory is, however, as much designed to account for maladaptive (drug addiction) as for hedonic outcomes of such opponent processes. Still, the affective trajectory of parachute jumping as conceptualized by Solomon (1980) is a good non-art example of a trajectory in which negative affect (temporary anxiety) is more than counterbalanced by a state of relief, accomplishment, and even euphoria, which, in turn, is energized by the preceding components of negative affect. Notably, however, Solomon himself never followed up on remarks that art reception might be a good test case of his domain-independent model, too (Rozin 1999). In fact, he even speculated that “perhaps some aesthetic pleasures have no opponent process” (Solomon & Corbit 1974, p. 142). In any event, because Solomon's theory temporally separates hedonic and aversive episodes as first and “after” processes, it does not entail provisions for genuine co-activations of positive and negative affect (cf. Andrade & Cohen 2007).

The compositional rule of going through emotional antitheses is based on the assumption that the inclusion of “unpleasant feelings” provides aesthetic trajectories with a greater affective amplitude, emotional depth, and rate of dynamic change. A series of purely “pleasant feelings” is held to be aesthetically inferior to this interplay, because it is hypothetically limited in its capacity to support intellectual interest over longer trajectories and consequently more prone to wear out. Prominent terms that designate profoundly negative responses to too positive or even sweet content and to too much unadulterated beauty in art reception are “boredom,” “satiation,” and even “satiation-driven disgust,” as first discussed by Mendelssohn and Kant (cf. Menninghaus 2003). One of the most fundamental principles of aesthetics – the rule of uniformity amidst variety advocated by virtually all eighteenth-century aesthetics, as well as by Fechner (1876) and Berlyne (1971a; 1974) – supports this assumption, as well: If richness in “variety” is a preeminent feature of aesthetically appealing stimuli, then variety should also apply to the emotional aspects and effects at which they are aimed, with “variety” in emotional effect requiring the inclusion of negative emotions. These considerations suggest that negative emotions are conducive to enjoyment not just because their negative affective nature is outbalanced by their effects on the intensity of emotional involvement, but also because this negative affective nature acts as a remedy against aesthetic failures that could result from artistic compositions that rely exclusively on positive affect. In this capacity, the contribution negative emotions make to pleasurable interplays of positive and negative emotions relies primarily on their very negative affective nature.

To test and further refine the hypothesis of this subsection, we propose experimentally investigating trajectories that feature varying proportions of positive and negative emotional ingredients. In all likelihood, increasing the share of negative emotions from near zero to ever higher levels will not monotonically increase aesthetic appreciation; rather, it is likely to show an inverted U-shape (see also Berlyne 1971a; 1974). To test this hypothesis, we suggest developing fictional scenarios that manipulate negative and positive emotion cues in a systematic manner, ranging from solely negative or solely positive emotion cues to different admixtures of these. The scenarios that make use of only one type of emotion cue should receive comparatively low ratings on measures for aesthetic appreciation. Adding positive emotion cues to previously exclusively negative ones and adding negative cues to previously exclusively positive ones should, in both cases, result in higher scores on a variety of measures for aesthetic appreciation and emotional involvement (liking, beauty, interest, suspense, being moved, intensity of involvement, etc.). Studies of this type also have a potential to investigate upper limits for and optimal levels of the inclusion of negative emotions for pleasurable purposes, both in general and with respect to the different domains and genres of art. Speeches consisting wholly of epideictic praise (eulogy) may provide a good test case for a too exclusive focus on positive affect.

4.2. Concomitant mixed emotions as mediators of negative emotions’ positive contributions to enjoyment

This subsection adds a third player to the interplay of positive and negative emotions as discussed in the preceding subsection, namely, mixed emotions. Based on a review of research on sad films and sad music, as well as on horror films, we advocate the following hypothesis: Concomitant mixed emotions serve as bipolar mediators for incorporating negative emotions into positive enjoyment. This hypothesis has no direct antecedent in classical poetics and aesthetics. It is based wholly on recent empirical research on the enjoyment of negative emotions (Hanich et al. 2014; Wassiliwizky et al. 2015), while in general informed by research on co-activations of positive and negative affect (e.g., Larsen et al. 2003). Specifically, we discuss evidence for the hypothesis that the principal mediator emotions for sad films and horror films are feelings of being moved and suspense, respectively. We argue that both mediator emotions are of a mixed emotional nature and that this facilitates their role in reconciling negative emotions with the hedonic expectations of art reception (for the concept of mixed emotions, see (Carrera & Oceja 2007; Ersner-Hershfield et al. 2008; Larsen & McGraw 2011; Oceja & Carrera 2009; Rafaeli et al. 2007; Schimmack 2001). A study on benign moral violations (McGraw & Warren 2010) has already shown that mixed emotional responses to these violations can be a way to retain feelings of moral rejection while simultaneously finding the violations amusing and, hence, enjoyable.

Whereas the interplay of positive and negative emotions discussed in the previous subsection relies on a series of affective antitheses (and the interaction between them), the present subsection addresses phenomena that in terms of rhetoric and poetics are metonymical in nature (cf. Jakobson 1973). Metonymies are based on contiguity relations, that is, the meaning of one element of a representation is strongly informed by a neighboring or co-occurrent element. For example, in a novel, the deranged state of a handbag can be described to shed light on the affective state of the person carrying it (who might otherwise successfully conceal his or her state of embarrassment). Regarding emotions, contiguity in affect space underlies typical “family resemblances” between individual emotions. Thus, feelings of being moved and of being touched cluster closely with feelings of sadness and nostalgia; moreover, all four of these emotion states frequently co-occur (Menninghaus et al. 2015b; Sedikides et al. 2008). The principle of metonymical transfer then predicts that the more unambiguously negative emotions (for instance, sadness) that are part of such contiguity-based clusters can profit from adjacent emotions that are more positive in affective nature (for instance, being moved and nostalgia). Similarly, at the opposite end of affect space, horror and suspense are both high-arousal emotions that frequently co-occur in response to specific genres of artworks and media products, yet they differ in affective valence: suspense is of a mixed affective valence (for details, see sect. 4.2.2), whereas horror is typically of an unambiguously negative nature. The principle of metonymical contiguity then predicts that the more positive affective nature of suspense can inform and partly transform co-occurrent or adjacent feelings of horror and, hence, lead to perceiving horror as (more) enjoyable. Thus, in both cases, the arts take advantage of latent affinities between emotions that are contextually activated by means of compositional contiguity.

4.2.1. The case of sadness

For eighteenth-century Scottish moral sense philosophers such as Lord Kames (1751/2005; see also Zelle 1987, p. 176), feeling compassion for the plight of others conforms to the standards of moral virtue and humanity; those who experience these feelings can therefore be pleased with their own emotional responses (cf. Aikin & Aikin 1773; Princess Elisabeth of Bohemia & Descartes 1645/2007, p. 118; for a more recent philosophical version, see Feagin 1983; for a more recent psychological version, see Schaller 1993). In recent media psychology, this focus on compassion, or empathy (for subtle distinctions between these terms, see Klimecki et al. 2012) as self-gratifying prosocial responses to the suffering of others has been reformulated in the “meta-emotions” model (Bartsch 2007; 2008; Bartsch & Viehoff 2003; Bartsch et al. 2008, 2010; Mills 1993; Oliver 1993; Schramm & Wirth 2010). Typologically, the different versions of the hypothesis of enjoyment qua self-gratifying empathy suggest a transformation of sadness into a source of pleasure by means of a mediation, a detour through a concomitant emotion – in this case empathy/compassion – together with a shift in focus from the object of the empathic response to the implication of this response for how the onlooker feels about him- or herself. Empirical evidence that the positive affect found in responses to sad films may actually be causally a result of such meta-emotional (re)appraisals rather than other response dimensions is, however, only tentative. Moreover, David Hume already offered a powerful argument against this hypothesis: If awareness of our own prosocial sympathetic impulses were by itself a sufficient reason for deriving pleasure from sad scenes, it would follow that “a hospital would be a more entertaining place than a ball” (Hume 2004, p. 243).

Another explanation stipulates that sad music, poems, narratives, and films can be enjoyed because, and to the extent that, they are blended with, or integrated into, episodes of being moved (for general definitions of this emotion state, see Kuehnast et al. 2014; Menninghaus et al. 2015b; Tokaji 2003) and/or nostalgia (Juslin 2013; Taruffi & Koelsch 2014), and that this association allows sadness to partake in the mixed, yet predominantly positive affective nature that is characteristic of being moved (Hanich et al. 2014; Wassiliwizky et al. 2015) and/or nostalgia (Routledge et al. 2011; Sedikides et al. 2008; Wildschut et al. 2006; 2010).

Two prototypes have been shown to account for a great portion of episodes of being moved. In the sadly moving prototype, experiences of loss (separation, death) or acts of sacrifice are associated with a positive appreciation of the value and memory of a beloved and/or with feelings of love and/or empathy on the part of bystanders and onlookers (Menninghaus et al. 2015b; Tokaji 2003). Similarly, situations of bidding farewell or separation can be represented –and have been shown to be experienced – as not just sad, but sadly moving (Wassiliwizky et al. 2015), because they co-activate both the pain of temporarily or permanently severing an important social bond and a heightened sense of the value of this bond and also, in part, hopes for restoring it. The second prototype of being moved is of a joyfully moving nature. Typical eliciting events include nostalgic memories of one's childhood or a former romantic relationship, as well as marriages, reunions, and reconciliations (Kuehnast et al. 2014; Menninghaus et al. 2015b; Taruffi & Koelsch 2014). In all of these cases, the predominant positive feelings come with some negative antidotes, such as an awareness that the happy times of childhood are forever gone, that a happy reunion was preceded by a painful period of separation, or that a couple will invariably have to face days and experiences that are less joyful than their wedding day.

Regardless of the inverse proportions of sad and joyful feelings in the two prototypes of being moved, overall affective responses to emotionally moving artworks show a predominance of positive affect (Hanich et al. 2014). This implies that the partially bipolar nature of states of being moved does not turn them into examples of disconcerting ambivalence and, hence, examples of a response pattern that psychological research has closely associated with the very notion of mixed emotions (Larsen et al. 2001; Norris et al. 2010). Thus, concomitant negative memories of a previous separation typically do not turn a reconciliation into an event that is emotionally ambivalent or that calls for a difficult decision between embracing the positive and embracing the negative response dimensions. Moreover, both the sad and the joyful/happy variants of being moved virtually always include more or less oblique references to positive feelings of bonding and attachment and to prosocial norms, values, and self-ideals (Fukui & Toyoshima 2014; Konečni 2005; Konečni et al. 2007; Menninghaus et al. 2015b; Panksepp 1995; Stel et al. 2008; Tokaji 2003).

These findings entail two further insights into the art of selecting and combining negative emotion cues. First, for all of their reliance on sadness, emotionally moving artworks by no means draw on all possible instances of sadness. Rather, only fairly circumscribed varieties of sadness are eligible to be associated with feelings of being moved. For example, forgetting one's favorite jacket in a restaurant and not retrieving it another day can elicit regret, and even some anger, about one's own forgetfulness and may later lead to feelings of a saddening loss; however, this type of sadness is not likely to elicit feelings of being emotionally moved. Briefly put, only sad feelings that have a direct bearing on social bonding and attachment can become ingredients of feelings of being moved (Menninghaus et al. 2015b); sentimental and nostalgic feelings place similar constraints on including ingredients of sadness (Sedikides et al. 2008; Wildschut et al. 2006). Hence, possible incorporation into altogether pleasurable mixed feelings does not extend to all instances of sadness, but is subject to selective constraints. We therefore propose testing with a broader corpus of “sad” artworks the hypothesis that a great share of sad feelings elicited by artworks cluster closely with feelings of being moved, attachment feelings, nostalgia, sentimental feelings, and feelings of quiet (the latter because sadness is low in arousal; cf. Kreibig 2010; Russell 2003).

Most importantly for the present context, feelings of being moved by sad film clips show a robust positive correlation with appreciating the respective films as artistic achievements and with the level of enjoyment they elicit (Hanich et al. 2014; cf. also Tokaji 2003). In contrast, the positive correlation between felt sadness and aesthetic appreciation/enjoyment disappears when potential mediating effects of feelings of being moved are controlled for. Wassiliwizky et al. (2015) replicated this finding and extended it by showing that no such mediation was found for joyfully moving films; felt joy contributed to overall enjoyment in a direct fashion that was wholly unmediated by the concomitant feelings of being moved. Another study replicated the findings of Hanich and colleagues for sad music (Eerola et al. 2016). On a similar note, a recent study has shown that sad music is hedonically rewarding because it also and even predominantly evokes feelings of nostalgia and tenderness in Western listeners and feelings of peacefulness and tenderness in Eastern (Asian) listeners (Taruffi & Koelsch 2014; see also Huron 2011; Kawakami & Katahira 2015). Thus, it is not sadness qua sadness as a negative emotion that is liked and that contributes to aesthetic appreciation. Rather, what ends up being enjoyed are specific metonymical configurations of specific instances of sadness with a very circumscribed range of other emotional responses. Therefore, treating the topic of “sad” artworks in complete abstraction from the nearer and broader neighborhood of the respective sadness elicitors can easily be misleading.

In episodes of being sadly moved, a pity/compassion/empathy factor (de Vignemont & Singer 2006; Eerola et al. 2016; Keen 2006; Singer & Lamm 2009; Singer et al. 2004) is often part of the emotional mélange (cf. Menninghaus et al. 2015b). Therefore, the enjoyment qua meta-emotional reappraisal of one's own feelings of empathy and the enjoyment qua being moved hypotheses do have some overlap. At the same time, the enjoyment qua being moved hypothesis is by no means limited to self-gratifying (meta)implications of feeling empathy for the plight of others. Sad films or narratives typically do not merely represent sad events; rather, they mostly feature friends, family, bystanders, and other witnesses who observe the plight of the protagonists already in the plot of the artwork itself, and at least some of these observers show prosocial responses of empathy and compassion and, occasionally, deep respect and admiration. Hence, more often than not, readers and viewers of sadly moving artworks can directly observe positive prosocial responses to sad events as depicted or displayed in the sadly moving artworks themselves and need not exclusively construe these in a meta-emotional fashion as a wholly self-reflective appraisal of their own empathic feelings. In other words, the enjoyment qua being moved hypothesis provides a broader cognitive basis for the positive feelings that sad films and poems have been shown to elicit, one that is not confined to a meta-appraisal of one's own feelings. At the same time, this hypothesis does not formally reject the meta-emotion hypothesis, but rather integrates it as one potential dimension within a more comprehensive framework.

Sadly moving real events, such as funerals, can similarly elicit positive feelings of high esteem, gratitude, and respect for the deceased person, along with a sense of social bonding among the survivors and, potentially, also of a self-rewarding beauty of the ceremony. Because being moved is mostly tied to a witness position (Tan 2009) and hence typically does not activate any attempt to change the respective emotion-eliciting event (Menninghaus et al. 2015b), emotional responses to real funerals experienced in person, on the one hand, and in response to art- and media-represented funerals, on the other, may in some cases not differ very much. The same, however, would not apply to experiencing real versus media-elicited horror, because in this case the appraisal of one's own safety being challenged would yield a strong affective difference for the two contextual framings. Hence, the example of real funerals does not challenge the importance of an art framing, but only shows that specific cognitive appraisals have a strong influence on how greatly emotional responses differ dependent on the activation of an art or ordinary reality framing. Presenting a filmic representation of a funeral as a documentary or as part of a fictional movie would enable investigation – however, within a shared representation framing – of the subtle differences in emotional responses that are still likely to derive from the ontological distinction between real represented and fictional represented events.

4.2.2. The case of horror (fear, fearful dread)

Research on horror films has provided evidence that increased levels of negative affect while watching horror films are associated with greater enjoyment (cf. Hoffner & Cantor 1991; Sparks 1991; Zillmann et al. 1986; Zuckerman 1979) and, more specifically, that habitual horror film viewers positively embrace not just emotional antidotes, such as moments of relief and happy endings, but the fearful feelings themselves (Andrade & Cohen 2007).

Given the design of the studies mentioned above, the self-reported positive affect is far from easy to interpret. It may have, at least partly, been a response to the actors, superb special effects, the setting, the editing, the plot construction, the soundtrack, and/or other dimensions of the artistic making of the films. However, nuanced measures of aesthetic appreciation were not made in these studies. Regarding emotional responses, data for positive and negative affect were routinely collected, and additional data for happiness and fear were collected in at least some of the studies (Andrade & Cohen 2007). However, even though an important role of suspense and thrill seeking is widely acknowledged in the literature on horror films (Hoffner & Levine 2005), we know of only one study (Sparks & Ogles 1994) that has collected both suspense and enjoyment ratings in addition to those for fear. Only these three measures together – ideally complemented by measures for arousal, positive, and negative affect – would allow a mediation analysis testing the hypothesis that suspense-driven arousal is an important factor in the co-activation of positive and negative affect and that this factor may be instrumental in making fear/horror enjoyable. Pursuing other research questions, however, the aforementioned study did not perform such a mediation analysis and also not a correlation analysis.

Feelings of suspense have been shown to be pleasurably experienced in response to many plot-based artworks, be these literary narratives, dramas, or films, including horror films (for the latter genre, see Hoffner & Levine 2005; Zillmann 1980; Zillmann & Weaver 1996). Narrative suspense is a state of cognitive uncertainty regarding the outcome of a plot trajectory; it can go through varying degrees of fearful and hopeful anticipations, and it creates a need for resolution, which may or may not turn out to comply with the reader's or onlooker's expectations and wishful desires (cf. Anz 1998; Berlyne 1960; Carroll 1996; Fill 2007; Hanich 2014; Lehne & Koelsch 2015; Löker 1976; Zillmann 1980). Oscillating as they do between fearful and hopeful anticipations, states of suspense can be ranked among the states of mixed affective nature (Madrigal & Bee 2005). (Because the pertinent discussion does not address the issue of negative emotions, we here do not discuss the question of whether or not suspense can be experienced in repeated exposure to the same literary or filmic narratives; see Carroll [1996], Gerrig [1997], Mag Uidhir [2011], Prieto-Pablos [1998] Smuts [2009b], and Yanal [1996].)

Zillmann's theory of suspense (1980) stipulates that it is the happy end that turns a suspenseful trajectory into a self-rewarding experience. Alternative theories, however, allow for positively appreciating the suspenseful trajectory itself. Specifically, the psychological construct of sensation seeking (Zuckerman 1979) suggests that the cognitive uncertainty and affective ambiguity between fear and hope can be experienced as inherently self-rewarding, if and to the extent that these negative cognitive and affective aspects of suspense fulfill needs for affective and physiological arousal that are different from a need for a happy resolution. Thus, fearful dread may feed and maximize the suspenseful emotional arousal that horror film viewers might ultimately seek and experience as self-rewarding, irrespective of what the end is like; after all, most typical recent media products of this genre lack a happy ending. Confirming this assumption, frequent viewers of horror films report that “the jolt of horror is exhilarating” and leaves them “feeling invigorated” (Tamborini & Stiff 1987, p. 425); they expressly view horror films “as a way to get an adrenaline high or to feel pumped up and alive” (Robinson et al. 2014, p. 46). Andrade and Cohen's interpretation of their data (2007) is likewise compatible with the hypothesis that the enjoyment of fear/horror may be mediated through experiencing suspense.

If this hypothesis were to withstand further testing, it would follow that the arousing nature of horror films has two sources, fear/dread/horror and suspense-driven arousal, with suspense being of a mixed affective nature and thereby potentially facilitating – in conjunction with the art framing – the positive reevaluation of the negatively arousing nature of fear (horror). (On a qualifying note, this hypothesis may not apply to horror film viewers who appear to directly draw positive enjoyment from endorsing, if not identifying with, the aggressive and often sadistic behavior of the perpetrators [cf. Oliver & Sanders 2004; Shaw 2001].)

Mixed emotional states of suspense are likely to contribute to enjoyment associated with negative emotions way beyond the particular genre of horror films. As already pointed out in Section 2, narratives and dramas, including comedies, routinely involve social conflicts, obstacles, suspenseful states between fear and hope, and the negative feelings associated with such conflicting predicaments (Grodal 2007; Krämer & Witschel 2010; Scalise Sugiyama 2005). Moreover, if one considers musical tension – which has been shown to often involve interplays of positive and negative emotional cues on the levels of tempo, harmony, and other dimensions (Hunter et al. 2008) – as an analogue to narrative suspense, then the composition of temporal works of art, including literature, films, music, and dance, is routinely experienced as a temporal trajectory of suspense/tension and resolution/release (Huron 2006; Meyer 1961; Trehub 2000). Both the “sweet anticipation” of release (Huron 2006) and its actual experience in listening to music depend on a prior buildup of antagonistic tension-release patterns and on postponements of resolution rather than on an ongoing conformity to processing ease or pleasantness at no cost. Dissonances that increase partially unpleasant tension may well support stronger feelings of resolution and relaxation both during the online resolution of tension and afterward (Koelsch 2014).

Finally, to the extent that sad narratives and films also involve feelings of suspense, they are likely to recruit two mixed emotional states – being moved and suspense – for integrating negative feelings into overall pleasurable trajectories (for an example, see Hanich & Menninghaus 2017). Thus, the mediator emotions we have discussed in separate subsections and with reference to two polar genres are by no means clear-cut alternatives, but can well be found in responses to the very same artworks and media products.

4.2.3. The case of disgust

Empirical evidence regarding the adoption of disgust for the pleasurable purposes of the arts is far less available than evidence regarding the enjoyment of sad films, poems, music, and horror films. Accordingly, this subsection is far shorter and of a more theoretical nature than the preceding two. Elaborate reflections by Nietzsche, Freud, Bataille, Sartre, Kristeva, and others (for a detailed account of this tradition see Menninghaus 2003, Ch. 5–9) converge in emphasizing that disgust has a potential to involve us in hidden, if not repressed, dimensions of profound pleasurability, or even jouissance. Freud's insistence that very young children like to play with their feces and even consider them as valuable gifts and that many feelings of disgust may be conceived of as repressed pleasure (cf. Menninghaus 2003, Ch. 6) is only the most prominent among the multiple, mostly philosophical theories that stipulate a partly positive reevaluation of the emotion of disgust specifically in art contexts. An empirical study focusing on the humorous and amusing implications of feelings of disgust in art contexts has provided nuanced evidence for a mixed affective nature of disgust in these contexts (Hemenover & Schimmack 2007). Another recent study (Rozin et al. 2013) has likewise reported evidence for a reverse evaluation of potential elicitors and feelings of disgust. Thus, feelings of disgust in art reception may not only support hedonic processing by virtue of their sheer arousal value; the arts may also bring out (latent) pleasure dimensions (Korsmeyer 2011) that are typically not included in psychological accounts of disgust (Rozin & Fallon 1987; Rozin & Haidt 2013; Rozin et al. 2008; Tybur et al. 2013). Again, by no means do all otherwise disgusting feelings allow for such a positive reevaluation in art contexts; for example, to date, the genuine stench of corpses has never been incorporated into “disgust art,” not even in Damien Hirst's provocative displays of decaying animal matter. Clearly, the arts implement a selective regime of compatibility and noncompatibility with enjoyment in the case of disgust, as well.

4.2.4. Summary

Summing up, this subsection argues – and partly provides empirical evidence – for the hypothesis that the arts tend to draw on negative emotions in such a fashion that their elicitation ends up fueling and energizing neighboring or concomitant feelings of a mixed affective nature and that this metonymical contiguity with mixed emotions plays a mediating role for negative emotions' contributions to overall enjoyment. Supporting these assumptions from other vantage points, recent research in developmental psychology has shown that individuals are more motivated to explore, seek, and maintain negative affect if it is accompanied by positive affect and is hence part of a context that has a mixed affective nature (Riediger et al. 2009). Findings that complex mixed emotions play a stronger role in art reception than pure and simple negative emotions (cf. Krämer & Witschel 2010; Oliver et al. 2009; Wirth et al. 2006) point in the same direction.

The hypothesis discussed throughout this subsection goes beyond the more abstract hypothesis of emotional antithesis discussed in the preceding subsection in that it adds an intermediate – mixed emotions – to the compositional interplay of positive and negative emotions. It attributes a crucial mediator role to this additional player and highlights the importance of particular contiguity relations among the interacting emotions in affect space, thereby substantially limiting the particular ranges, or instances, of individual negative emotions that are eligible for being adopted for pleasurable purposes. Because interplays of positive, negative, and mixed emotion are even more emotionally varied than those of positive and negative emotions only, they are likely to support the benefits of the latter – more emotional variety and dynamic changes – at least to an equal degree. In fact, one might well find upon closer inspection that complex emotional states of a mixed affective nature always play a role in the integration of the powers of negative emotions into altogether pleasurable trajectories. In that case, the compositional interplays discussed in sections 4.1 and 4.2 would not be categorically different anymore.

It is highly likely that other emotional states that are either routinely or frequently of a mixed emotional nature serve a role similar to the one we have diagnosed for being moved and suspense. The validity of this assumption could be tested on awe (Keltner & Haidt 2003; Silvia et al. 2015) and feelings of the sublime (Eskine et al. 2012; Gordon et al. 2017), nostalgia (Wildschut et al. 2006), surprise (Noordewier & Breugelmans 2013; Silvia 2009), and even special variants of confusion (Silvia 2009; 2010). If such tests yield positive results, one could even stipulate in a generalizing fashion that potential enjoyment of a particular negative emotion in art reception is dependent on the availability of a closely related mixed emotion that both is contiguous in affect space and can readily co-occur with the respective negative emotion. Speculative as though this consideration is, it emphasizes the importance of defining, as we did in this subsection, additional constraints that need to be met for a particular negative emotion to support intensely felt and memorable pleasurable experiences in art reception. For the time being, the range of negative emotions that meet the constraints defined above appears to be fairly limited.

4.3. Aesthetic virtues of the artistic uses of the media of representation

Following the system of poetics, the next step after treating the selection/invention and the combination of the major constituent parts of artworks is to consider the ways in which artworks use their specific media of representation for emotion-regulatory purposes. Regarding this dimension that is found in all artworks, we propose the following hypothesis: Aesthetically appealing uses of the media of representation (such as sound/music, words/language, color/shape) have the power to make the processing of negative emotional content or associations more enjoyable while not reducing, let alone erasing, negative emotional responses.

Importantly, all (re)presentational media not only represent something, but also are something on their own. They have their own materiality and specific powers to signify, represent, allude to, or evoke some meaning and/or emotional response. The arts of writing, music, painting, and so forth are typically believed to make a special, more elaborate, and even partially alienating use of language, musical structures/musical performance, or color and shape compared with ordinary or more quotidian forms of language, singing, and painting (Dissanayake 2000). In fact, the artistic use of the very media of representation constitutes the material and phenomenal “reality” of artworks; it takes up a great portion of the artists' efforts, provides a reason for admiring their particular skills and performances (Newman & Bloom 2012), and is likely to make a substantial contribution to aesthetic liking. Pleasure taken in the very art of representation is (primarily) not due to the processing of the object of representation, but rather to the aesthetic properties (Walton 1970) or virtues inherent to the representation itself (coloring, execution, poetic style, etc.; cf. Tinio & Leder 2009; Tinio et al. 2011). Notably, all aesthetically appealing features of the wording used to represent, and reflect upon, a uxoricide in tragedy – Aristotle explicitly stressed the “sweetness” of diction, meter, and melody in the language and the sung portions of tragedy (1961, paras. 1449b and 1450b) – are by definition missing in the real event. As a result, the basis for an affective appraisal is likely to be different for a real event and for its artistic representation (cf. Friend 2007), because in the latter, the event represented interacts with the very means of artistically representing it.

Research on negative emotions in art reception has widely disregarded the aesthetic appeal of the purely formal quality of an artwork or media product. Studies on preferences for horror films are particularly weak in this regard: No other genre has prompted so many studies on the enjoyment associated with negative emotions, yet apparently none of these studies has experimentally modified the patterns of how these films employ lighting, coloring, camera perspective, techniques of cutting, special effects highlighting the monster's dreadful appearance, soundtrack, and so forth (for similar résumés, see the meta-analyses by Hoffner & Levine 2005; Tamborini et al. 1990). Moreover, empirical evidence regarding the effects of special uses of the media of representation on the enjoyment of negative emotion is likewise scarce beyond the genre of horror films.

Positive statistical correlations between aesthetic liking, the intensity of being affected, and felt negative emotions have been reported in several studies on sad films (Bartsch & Viehoff 2003; Bartsch et al. 2010; Hanich et al. 2014; Oliver 1993; Oliver & Bartsch 2010; Wassiliwizky et al. 2015) and soundtracks of sad films (Eerola & Vuoskoski 2011). Experimentally modifying 20 sadly and 20 joyfully moving poems that differ substantially in time of origin, a recent study (Menninghaus et al. 2017) investigated the effects that features of poetic parallelism as defined by Roman Jakobson (1960) exert on emotional response dimensions (being moved, sadness, joy), on unipolar ratings of overall positive and negative affect, on several measures of aesthetic appreciation, and on the general impression of the intensity of being affected by the poems. The presence versus absence of the target patterns of poetic diction increased feelings of sadness, being moved, intensity, and positive affect (in the case of the sadly moving poems) and of joy, being moved, intensity, and positive affect (in the case of the joyfully moving poems). Importantly, ratings for sadness correlated positively with ratings for beauty, aesthetic liking, melodiousness, being moved, intensity, and positive affect. Hence, the artistic treatment of language was shown to enhance the positive aesthetic appreciation of poems of a primarily sad content, even without reducing, let alone converting, ratings for felt sadness and overall negative affect. This parallel increase in perceived feelings of sadness and being moved, perceived aesthetic virtues, and perceived overall intensity strongly supports the model presented here. Because parallelistic diction in general enhances ease of prosodic processing (Obermeier et al. 2016) – albeit often at the expense of making semantic processing more demanding (Menninghaus et al. 2015a) – these findings also speak to an important contribution of processing fluency (Alter & Oppenheimer 2008; Oppenheimer 2008; Reber 2016; Reber et al. 2004) to the enjoyment of negative emotions.

Analogous effects are likely to be achieved through the special artistic uses of the respective representational media in other art forms, as well. Supporting this assumption, a large share of music that is perceived as “sad” is also perceived as outstandingly beautiful; moreover, self-reports regarding responses to sad music suggest that individuals find it rewarding to imagine that, on occasion, they would have the same expressive power and potency as the sad musical pieces they like (Taruffi & Koelsch 2014). This imaginative transfer, too, appears to be strongly dependent on the perceived artistic achievement of the pieces of music in question and, hence, clearly different from a mere effect of psychological distancing.

4.4. Redeeming negative emotions through (symbolic) meaning making

The construction of (symbolic) meaning is yet another level of art processing for which negative emotions are not just stumbling blocks, but also positive contributors. Here we propose the hypothesis: Interpretive efforts toward meaning making contribute to (re)appraising negative emotional content and concomitant feelings in a (more) positive and enjoyable light. The interpretation of ancient Greek tragedies is a classical case. For example, in line with the understanding of tragedy promoted by Hegel (1970, p. 547) and other philosophers, Bullough (1912, p. 104) suggested that “real tragedy … truly appreciated, is not sad … it is an homage to the great and exceptional in man.”

The search for and discovery of some sort of meaning is a frequently used cognitive strategy for either retroactively or simultaneously (re)appraising negative events in a more favorable light (Giuliani & Gross 2009; Gross 1998; Gross & Thompson 2006; Larsen & Prizmic 2008; Ochsner & Gross 2005; Oliver et al. 2009; Oliver & Woolley 2010). In recent media psychology, reevaluating negative affect from the perspective of a higher-order meaning(fulness) specifically plays a large role in studies on the enjoyment of sad films (Bartsch 2007; 2008; Bartsch & Viehoff 2003; Bartsch et al. 2010; Oliver 1993; Oliver & Bartsch 2010; Oliver & Woolley 2010). Meaningfulness is what Fechner (1876, pp. 238–40) called a “reconciliatory moment.” The heroic death for a good cause is a conspicuous – and often highly ideological – model for such a moment, one that is not even limited to the precondition of an art framing. However, in art contexts, the art framing is always additionally in place, and this should yield differences on some appraisal-driven dimensions of the emotional responses (cf. Menninghaus et al. 2015b).

Thus, contrary to Zillmann's hypothesis (1971; 1991; de Wied et al. 1994), plots that draw heavily on negative emotions by no means need to have happy endings to allow for markedly positive affective (re)appraisals. A prototypical happy ending, although clearly important for reevaluating negative affect in many cases, is only one of many ways to integrate negative emotions into an overall pleasurable trajectory by mixing them with positive ones. Notably, sensation seekers are far less likely to search for a higher-order level of symbolic meaning in horror films, and the same holds for consumers of disgust comedies. We therefore suggest that the meaning-construction route to accommodating negative emotions may only be an option for select cases.

4.5. Genre scripts as emotion-regulation scripts

Acquired genre schemata, or genre scripts, are different from the highly abstract schemata of art, representation, and fiction in that they entail fairly detailed anticipations as to which specific emotions/emotional tonalities recipients are likely to be going through and to what extent emotionally negative content is likely to elicit negative feelings in recipients. In other words, genre concepts entail dimensions of affective forecasting (Wilson & Gilbert 2003; 2005) and hence serve to pre-adjust expected emotional responses (cf. Menninghaus 1999). Here we argue for the following hypothesis: Activation of particular genre scripts can contribute to (re)appraising negative emotions in a positive and enjoyable light.

In his essay The Uncanny, Freud (1955) noted that a narrative content feature that evokes strong feelings of the uncanny in a fantastic horror narrative is likely to elicit no such feelings when encountered in a fairy tale. In a similar vein, exposure to disgusting matters is likely to be perceived differently in a disgust comedy than in a tragedy (cf. the case of Philoctetes as discussed by Lessing 1766/1984). Anticipating recent experimental evidence from film studies (Visch & Tan 2007; Visch & Tan 2008; Visch et al. 2010), Freud contended that an artful writer commands high skills for manipulating the reader's mind frame by means of subtle stylistic cues rather than explicit instructions of a propositional kind (Freud 1955).

To date, genre attributions primed by subtle stylistic cues have not been empirically tested for effects on negative emotion processing. All existing studies have used explicit propositional framing instructions, and these framings have been used exclusively to contrast fictional literary texts with nonfictional news texts (Altmann et al. 2014; Zwaan 1994) rather than different literary genres with specific affective profiles. Freud's remark clearly calls for experimental testing by means of inserting text passages of identical wording into different genre contexts that prime different meaning attributions.

4.6. Summary

This section spells out the five hypotheses that underlie the second pillar of our model (the Embracing factor; see Fig. 1). The majority of these hypotheses have previously not been part of the psychological theorizing about pleasure associated with negative emotions, at least not in a more elaborate form. Notably, our model does not rely on compensatory mechanisms (cf. Carroll 1990b), if compensation means that negative emotions first have a wholly negative effect on the enjoyment of art which is then healed by positive antidotes. After all, the top-down activation of art, representation, and fiction framings preemptively alters important appraisal dimensions of the negative emotions, and genre scripts have similar a priori effects on affective processing. Moreover, from the very beginning, the experiencing of negative emotions during art reception is inextricably linked to the aesthetically rewarding virtues of the artistic representation and to the interplay with positive and mixed emotions as described in this section. Therefore, we conceptualize our model not as a compensation model, but as a two-factor transformation model comprising the a priori Distancing of negative emotions (factor 1) and several mechanisms of Embracing the distanced negative emotions for pleasurable purposes (factor 2). Our model is also not a model of conversion, if conversion means a full-blown transformation of negative into positive affect. After all, we consider it necessary that negative emotions are actually experienced as such, at least within the constraints of the Distancing factor.

5. Limitations and additional future directions

As emphasized in the Introduction, our eight-component model is not a component-process model in any narrower meaning of this term. It hypothesizes that the identified processing components are relevant for negative emotion processing across art domains, yet leaves it to subsequent studies to test how readily these eight components can be integrated into process models of the individual arts.

Some components of the Embracing factor – most notably, the components “compositional interplays of positive and negative emotions,” “aesthetic rewards of the very form of representation,” and “emotion-regulatory implications of particular genre scripts” – have a substantial tradition in treatises on poetics and aesthetics. This raises the question whether artists employed the respective means of representation in a theory-guided fashion or based on intuitive knowledge only. Letters and other testimonies of artists could be scrutinized for evidence of conscious, theory-based anticipations regarding the emotion-regulatory effects of the respective strategies of representation.

Potential additional explanatory mechanisms likewise need to be considered. This applies specifically to the hypothesis that the co-occurrence of enjoyment and negative emotions when viewing horror films and similar media products might be explained as a benign variety of genuine masochism, that is, of the physical and psychological pain that masochists embrace as (sexually) pleasurable (Bloom 2011, pp. 51–52, 194–97). Rozin et al. (2013) have surveyed a broad variety of activities that may be accounted for with the help of this theoretical explanation; art-specific processing mechanisms were not considered in this context. Future studies will therefore need to investigate whether the benign-masochism hypothesis can indeed explain in a very parsimonious fashion all effects that we here ascribe to several art-specific processing components.

For comparative reasons, it would also be interesting – even though very difficult for both reasons of ethics and study design – to investigate non-art instances that eighteenth-century treatises on aesthetics routinely discussed in the context of the topic: the notorious attraction of gladiator's fights and public executions and also of apocalyptic visions and catastrophes.

Finally, the powers of negative emotions to secure attention, intense involvement, and privileged access to memory are likewise recruited – albeit not within the activation of an art framing – by news reports, political speeches and propaganda, and commercial ads. Considering that classical rhetoric and poetics essentially used the very same framework for analyzing poetic, political, and other pragmatic speech, our model may serve as a basis for comparing the role of negative emotions across these different domains.

6. Conclusion

Revisiting a long tradition of rhetoric, poetics, and aesthetics in the light of recent empirical and theoretical work in both the psychology of emotions and aesthetics, we propose a novel integrative account of the aesthetic pleasure/enjoyment associated with negative emotions. Our Distancing-Embracing model (Fig. 1) features two groups of processing components. The first keep negative emotions at a cognitive appraisal–driven distance, thereby preventing them from being outright incompatible with the hedonic expectations of art reception. This sets the stage, or clears the ground, for the second group of components. The latter even positively adopt, or embrace, the particular powers of negative emotions in the service of intensifying overall enjoyment and rendering the trajectory of art reception more varied, interesting, and profound in its affective nature and less prone to induce boredom. In the light of this model, negative emotions are not a special, let alone paradoxical, license for particular art forms only. Rather, their powers are an important, valuable resource for the arts and art reception in general, and our model spells out the mechanisms that allow the hedonic bias of art reception to thrive precisely on the negativity bias (Cacioppo et al. 1999; Ito et al. 1998; Rozin & Royzman 2001) of our emotional system.