Models are fundamental in any scientific field. They summarize and simplify phenomena that are otherwise extremely difficult to grasp, as is the case with the role of emotions – particularly negative emotions – in art appreciation (Silvia Reference Silvia2009; Reference Silvia, Shimamura and Palmer2012). In this sense, Menninghaus et al.'s Distancing-Embracing model (DEM) is a welcome addition to the literature. But scientific models need to satisfy a set of criteria that establish their validity and ensure their potential to advance scientific understanding of the modeled phenomenon – in this case, the enjoyment of negative emotions in art reception. The validity and relevance of a model depend on the degree to which:
-
1. It summarizes systematically collected experimental or observational data: Models should be grounded on evidence.
-
2. It describes and explains the evidence: Models should highlight the causal mechanisms that bring about the evidence.
-
3. It formulates empirically testable hypotheses: Models should predict future evidence.
-
4. It brings theoretical clarity by integrating and relating diverse concepts and observations.
-
5. It is compatible with models developed at other explanatory levels.
-
6. It clarifies its relation to other existing models.
Menninghaus et al.'s DEM does not fare well when measured against these criteria. First and foremost, the model is not motivated by a comprehensive body of empirical evidence. For instance, Menninghaus et al. produce only indirect and inconsistent evidence supporting the existence and engagement of the art, representation, and fiction schemata they suggest underlie the Distancing factor (sect. 3). The DEM predicts that these schemata, activated by art framing, should influence the experience of negative emotions. However, Wagner et al. (Reference Wagner, Menninghaus, Hanich and Jacobsen2014) and Gerger et al. (Reference Gerger, Leder and Kremer2014) – cited as supporting studies – found no such influence. The DEM is grounded not on empirical data but on assumptions about art and emotion derived from philosophical and poetic theorizing. One such assumption, presented in Section 2, is that negative emotions are general resources predestined for the arts' purposes, implying that the experience of negative emotions is pervasive in art reception. This assumption, however, is clearly untenable, because not all interactions with art involve negative emotions (Martindale Reference Martindale2001; Smith Reference Smith2014), and all forms of art, including prose and poetry, abound with works intended to evoke positive emotions (subgenres of comedy, erotica, lullabies, etc.). Moreover, no evidence is marshalled to bolster the paper's main claim, namely, that art's function is to transform negative emotions into pleasurable responses, nor to refute alternative possibilities.
Second, the DEM makes no reference, whether in cognitive or in neural terms, to the fundamental explanatory mechanisms underlying the distancing and embracing processes. Impressionistic descriptions are used to illustrate how, for instance, conceiving an art object as fictive can provide “an awareness that no real person (or animal) has been physically harmed” (sect. 3.3, para. 1). The notions invoked by such descriptions are conceptually obscure.
Moreover, the model's central concepts and processes, such as “experiential spaces” and to “keep felt negative emotions at some psychological distance” (sect. 3, para 1), are inadequately specified, bordering on folk-psychological. No attempt is made to relate them to psychological or neurobiological processes commonly studied by cognitive neuroscience. Does regarding something as fictive elicit top-down regulative mechanisms? What processes would such putative regulative mechanisms modulate? The perceptual processing of the percept being computed? The affective responses associated with these attenuated percepts? The DEM cannot answer such questions. And this is a crucial limitation, because in addition to describing a phenomenon, a scientific model should be able to explain how it works.
Furthermore, because the DEM does not postulate adequately specified explanatory mechanisms, the role of these mechanisms in the enjoyment of negative emotions cannot be tested empirically (sect. 4). For example, much is made about the co-occurrence of positive and negative emotions during the embracing phase. Testing this would require measuring the occurrence of negative and positive emotions separately and then registering their co-occurrence. Evidence from social neuroscience, however, suggests that rather than being dissociable, emotions are rooted in core affect – an organism's level of pleasant or unpleasant arousal (Wilson-Mendenhall et al. Reference Wilson-Mendenhall, Barrett and Barsalou2013) – and mapped onto a common reference space (Barrett & Wager Reference Barrett and Wager2006; Barrett et al. Reference Barrett, Mesquita, Ochsner and Gross2007). Because the DEM does not lend itself to deriving testable hypotheses nor to falsification, its potential to motivate research in empirical studies of the arts is uncertain.
Finally, the DEM neglects existing models and explanations of the role of negative emotions in art reception (e.g., Sachs et al. Reference Sachs, Damasio and Habibi2015). For example, our understanding of visual art perception has been advanced recently by incorporating principles that account for the interplay between prediction error and reward (Kesner Reference Kesner2014; Van de Cruys & Wagemans Reference Van de Cruys and Wagemans2011). The basic idea behind predictive coding is that the brain actively anticipates incoming sensory input. When predictions are accurate, efficient processing of the input occurs. Conversely, when there is a difference between prediction and the actual state of affairs, a prediction error ensues. Prediction errors are therefore typically emotional and negative in valence. Van de Cruys and Wagemans (Reference Van de Cruys and Wagemans2011) proposed that artists habitually manipulate conditions that initially increase viewers' prediction error, which is subsequently resolved as the stimulus becomes predictable. This transition from unpredictability to predictability is experienced as rewarding. This account is consistent with evidence that visual perception has an affective component (Barrett & Bar Reference Barrett and Bar2009) and that attaining conditions that favor processing fluency is pleasurable (Reber et al. Reference Reber, Schwarz and Winkielman2004; Zajonc Reference Zajonc1968).
In sum, because the model presented in the target article fails to meet basic validity and relevance criteria, it has limited potential to advance scientific understanding of the role of negative emotions in art reception. Without strong grounding evidence, clear explanatory mechanisms, and direct testability, the DEM is more a description of the phenomenology of art reception than a causal model that explains how negative emotions can lead to pleasurable engagements with art. What empirical aesthetics and neuroaesthetics require at present are models that make sense of the available mountain of empirical evidence and that make testable claims about neural mechanisms that can advance our understanding of how the human brain, as a matter of empirical fact, produces art experiences.
Models are fundamental in any scientific field. They summarize and simplify phenomena that are otherwise extremely difficult to grasp, as is the case with the role of emotions – particularly negative emotions – in art appreciation (Silvia Reference Silvia2009; Reference Silvia, Shimamura and Palmer2012). In this sense, Menninghaus et al.'s Distancing-Embracing model (DEM) is a welcome addition to the literature. But scientific models need to satisfy a set of criteria that establish their validity and ensure their potential to advance scientific understanding of the modeled phenomenon – in this case, the enjoyment of negative emotions in art reception. The validity and relevance of a model depend on the degree to which:
1. It summarizes systematically collected experimental or observational data: Models should be grounded on evidence.
2. It describes and explains the evidence: Models should highlight the causal mechanisms that bring about the evidence.
3. It formulates empirically testable hypotheses: Models should predict future evidence.
4. It brings theoretical clarity by integrating and relating diverse concepts and observations.
5. It is compatible with models developed at other explanatory levels.
6. It clarifies its relation to other existing models.
Menninghaus et al.'s DEM does not fare well when measured against these criteria. First and foremost, the model is not motivated by a comprehensive body of empirical evidence. For instance, Menninghaus et al. produce only indirect and inconsistent evidence supporting the existence and engagement of the art, representation, and fiction schemata they suggest underlie the Distancing factor (sect. 3). The DEM predicts that these schemata, activated by art framing, should influence the experience of negative emotions. However, Wagner et al. (Reference Wagner, Menninghaus, Hanich and Jacobsen2014) and Gerger et al. (Reference Gerger, Leder and Kremer2014) – cited as supporting studies – found no such influence. The DEM is grounded not on empirical data but on assumptions about art and emotion derived from philosophical and poetic theorizing. One such assumption, presented in Section 2, is that negative emotions are general resources predestined for the arts' purposes, implying that the experience of negative emotions is pervasive in art reception. This assumption, however, is clearly untenable, because not all interactions with art involve negative emotions (Martindale Reference Martindale2001; Smith Reference Smith2014), and all forms of art, including prose and poetry, abound with works intended to evoke positive emotions (subgenres of comedy, erotica, lullabies, etc.). Moreover, no evidence is marshalled to bolster the paper's main claim, namely, that art's function is to transform negative emotions into pleasurable responses, nor to refute alternative possibilities.
Second, the DEM makes no reference, whether in cognitive or in neural terms, to the fundamental explanatory mechanisms underlying the distancing and embracing processes. Impressionistic descriptions are used to illustrate how, for instance, conceiving an art object as fictive can provide “an awareness that no real person (or animal) has been physically harmed” (sect. 3.3, para. 1). The notions invoked by such descriptions are conceptually obscure.
Moreover, the model's central concepts and processes, such as “experiential spaces” and to “keep felt negative emotions at some psychological distance” (sect. 3, para 1), are inadequately specified, bordering on folk-psychological. No attempt is made to relate them to psychological or neurobiological processes commonly studied by cognitive neuroscience. Does regarding something as fictive elicit top-down regulative mechanisms? What processes would such putative regulative mechanisms modulate? The perceptual processing of the percept being computed? The affective responses associated with these attenuated percepts? The DEM cannot answer such questions. And this is a crucial limitation, because in addition to describing a phenomenon, a scientific model should be able to explain how it works.
Furthermore, because the DEM does not postulate adequately specified explanatory mechanisms, the role of these mechanisms in the enjoyment of negative emotions cannot be tested empirically (sect. 4). For example, much is made about the co-occurrence of positive and negative emotions during the embracing phase. Testing this would require measuring the occurrence of negative and positive emotions separately and then registering their co-occurrence. Evidence from social neuroscience, however, suggests that rather than being dissociable, emotions are rooted in core affect – an organism's level of pleasant or unpleasant arousal (Wilson-Mendenhall et al. Reference Wilson-Mendenhall, Barrett and Barsalou2013) – and mapped onto a common reference space (Barrett & Wager Reference Barrett and Wager2006; Barrett et al. Reference Barrett, Mesquita, Ochsner and Gross2007). Because the DEM does not lend itself to deriving testable hypotheses nor to falsification, its potential to motivate research in empirical studies of the arts is uncertain.
Finally, the DEM neglects existing models and explanations of the role of negative emotions in art reception (e.g., Sachs et al. Reference Sachs, Damasio and Habibi2015). For example, our understanding of visual art perception has been advanced recently by incorporating principles that account for the interplay between prediction error and reward (Kesner Reference Kesner2014; Van de Cruys & Wagemans Reference Van de Cruys and Wagemans2011). The basic idea behind predictive coding is that the brain actively anticipates incoming sensory input. When predictions are accurate, efficient processing of the input occurs. Conversely, when there is a difference between prediction and the actual state of affairs, a prediction error ensues. Prediction errors are therefore typically emotional and negative in valence. Van de Cruys and Wagemans (Reference Van de Cruys and Wagemans2011) proposed that artists habitually manipulate conditions that initially increase viewers' prediction error, which is subsequently resolved as the stimulus becomes predictable. This transition from unpredictability to predictability is experienced as rewarding. This account is consistent with evidence that visual perception has an affective component (Barrett & Bar Reference Barrett and Bar2009) and that attaining conditions that favor processing fluency is pleasurable (Reber et al. Reference Reber, Schwarz and Winkielman2004; Zajonc Reference Zajonc1968).
In sum, because the model presented in the target article fails to meet basic validity and relevance criteria, it has limited potential to advance scientific understanding of the role of negative emotions in art reception. Without strong grounding evidence, clear explanatory mechanisms, and direct testability, the DEM is more a description of the phenomenology of art reception than a causal model that explains how negative emotions can lead to pleasurable engagements with art. What empirical aesthetics and neuroaesthetics require at present are models that make sense of the available mountain of empirical evidence and that make testable claims about neural mechanisms that can advance our understanding of how the human brain, as a matter of empirical fact, produces art experiences.