Many, if not most, poems are “sad” in terms of their emotional content, with their artistic construction (word choice, prosody) also expressing feelings of sadness, loss, and despair. Paul Celan's Death Fugue, Walt Whitman's O Captain! My Captain! and W.H. Auden's Funeral Blues are only three of the myriad examples for this. Importantly, readers do not just cognitively decode the emotional context and decipher the emotional expression of poems, but apparently also genuinely feel the sadness by way of empathy, emotional contagion, identification, or other means of emotional transfer (Lundqvist et al., 2009; Gerger et al., 2014). However, is sadness not an emotion we prefer not to feel? Or do we appreciate sadness in aesthetic contexts, such as reading poetry, as something positive? And do we appreciate happier, more joyous poems less than sad poems, however paradoxical this may seem?

Intuitively, positive aesthetic evaluation and the emotional classification of artworks as joyful or affectively positive seem very closely related. However, movies, music, and poems with sad, i.e., affectively negative, content have repeatedly been reported to be highly appreciated aesthetically. Notably, a rating study of the perception of sad and joyful music excerpts found a significant positive correlation between perceived sadness and perceived beauty (Eerola and Vuoskoski, 2011). Likewise, Oliver and Bartsch (2011, p. 31) suggested that the “experience of appreciation is often thought to be tied more closely with sad than joyful affect.”

Throughout the history of poetics and aesthetics, philosophers and poets have tried to tackle the question of why people enjoy and appreciate feelings of sadness (e.g., Hume, 1757/1793; Schiller, 1792/2006). Hanich et al. (2014) suggested that the overall positive feeling of being moved can be understood as a cause of the pleasure associated with negative emotions expressed in or elicited by sadly moving films (for similar findings, see Wassiliwizky et al., 2015). Other mediator emotions that appear to have the power to integrate feelings of sadness into an overall pleasurable emotional trajectory are feelings of nostalgia (Sedikides et al., 2008), tenderness, peacefulness, and relaxation (Taruffi and Koelsch, 2014).

The enjoyment of negative emotions in art reception has also been shown to be influenced by individual differences regarding tendencies to experience states of absorption and music-elicited empathy (Garrido and Schubert, 2011; Taruffi and Koelsch, 2014). Subjectivist theories understand aesthetic evaluation to be mainly determined by individual differences in prior experiences and personal attitudes (e.g., Dewey, 1934/2005). Since frequency measures of exposure to literature and scales like the Author-Recognition-Test (ART; Stanovich and West, 1989; Aacheson et al., 2008) focus mainly on narratives, they are of little use for assessing exposure to or familiarity with the genre of poetry. We here consider readers' general affinity for poetry (see below) as a trait variable that may influence their appreciation of given poems.

Aesthetic judgments, such as those of liking and beauty are often correlated (Brattico et al., 2011; Lüdtke et al., 2014) and understood to be closely related (cf. Reber et al., 2004). However, this is by no means always the case. For instance, horror films are clearly liked by their customary viewers, but research on horror films has not reported any strong experiences of beauty in this context; rather, liking appears to be driven primarily by high affective arousal, thrills, and suspense (Sparks and Ogles, 1994; Hoffner and Levine, 2005; Andrade and Cohen, 2007; Robinson et al., 2014). Similarly, artworks can be liked for being interesting, shocking, a good satire, or even for being markedly ugly (Schlegel, 1795–1797/1979; Rosenkranz, 1853/2015). In such cases, attributions of beauty are apparently no prerequisite for liking. In fact, the partial separation of perceived aesthetic appeal from beauty is one of the major topics and achievements of later eighteenth century, and specifically of post-Kantian aesthetics (for a programmatic volume of essays on this issue see Jauß, 1991).

However, for all these reasons not to commingle judgments of beauty and aesthetic liking, there is some empirical evidence that suggests a very close association between the two judgments in particular contexts. Sad music is one important example; liking of sad music routinely coincides with perceiving high degrees of beauty (Eerola and Vuoskoski, 2011; Taruffi and Koelsch, 2014). Regarding poetry, beauty has been shown—in pronounced contrast to novels and plays—to be (still) the prime expectation of perceived aesthetic appeal among non-professional contemporary readers, regardless of the key emotional tonality (Knoop et al., 2016). A recent experimental study has shown—with a specific focus on the role of parallelistic diction—that liking judgments and beauty attributions for poetry correlate positively both with each other and with self-reported feelings of joy, sadness and being moved (Menninghaus et al., 2016). In light of these data, we decided to follow previous studies on both music and literature (Brattico et al., 2011; Lüdtke et al., 2014) in measuring perceived overall aesthetic appeal by using the average of beauty and liking ratings as a composite index for aesthetic appreciation.

To date, no empirical investigation has considered aesthetic evaluation(s) of poetry in light of the respective poems' emotional classification and of readers' affinity for poetry. We set out to do precisely this. We expected higher aesthetic appreciation for sad poems than for joyful ones. Further, we hypothesized readers' self-reported affinity for poetry to be positively related to their aesthetic evaluations.

To test whether participants confirmed our pre-classification of the poems as either joyful or sad, we inspected the mean values of all poems on the item Emotion. The means of the poems that were pre-classified as joyful (M = 2.72, SD = 0.65, min = 1.63, max = 3.69) were all below the midpoint of the scale (4), whereas the means of the poems that were pre-classified as sad (M = 5.93, SD = 0.46, min = 5.25, max = 6.88) were all above the midpoint (for mean-Emotion ratings, see Table 1). To control for possible effects of participants' familiarity with the poems, we excluded two joyful poems that were familiar to more than 10% of the participants from further analyses³. On average, participants indicated an affinity of 5.05 for reading or listening to poetry (SD = 1.58, min = 1, max = 7).

Using the emotional pre-classification of the poems (coded in a binary way: 1 (joyful) vs. −1 (sad)), participants' Affinity and the number of words per poem as independent variables, we applied a linear mixed effects analysis with which we predicted Aesthetic Appreciation (as defined above). We also included intercepts for participants and poems as random effects in this analysis (Baayen et al., 2008).

Results show a significant effect of the emotional classification (t = −2.24; p = 0.03) and a significant effect of participants' Affinity for poetry (t = 3.45; p ≤ 0.001) on Aesthetic Appreciation. The number of words per poem was unrelated to participants' ratings (t = 0.72; p = 0.47; see Table 2 for estimates and standard errors of fixed effects and the intercept).

An inspection of the mean values for Aesthetic Appreciation showed that sad poems were rated higher (M = 4.70, SD = 1.45, n = 24) than joyful poems (M = 4.36 SD = 1.51, n = 22; see Figure 1).

Our analyses show that Affinity for poetry clearly affects ratings of Aesthetic Appreciation. Results also show a significantly higher Aesthetic Appreciation for sad than for joyful poems.

With all due caution, our findings can be interpreted as supporting theories of pleasure in negative affect that suggest a positive relation between sad stimuli and aesthetic appreciation (cf. Taruffi and Koelsch, 2014). Considering the well-known effect of familiarity on aesthetic evaluation (cf. Zajonc, 1968; Calvo-Merino et al., 2008), an explanation for our finding of higher Aesthetic Appreciation for sad poems could be that sad poems simply constitute a greater share of the (Western) tradition of poetry than joyful poems. This higher familiarity with sad poems might be the reason why they are generally more appreciated than joyful poems.

The results presented here are certainly limited by the chosen corpus, as well as the personal and textual variables that were analyzed. Therefore, they could be complemented by follow-up studies that include additional situational factors, incorporate a broader exploration of readers' characteristics, and do not exclusively rely on behavioral data.

Furthermore, due to the theoretical separation between beauty and other forms of appreciation within the realm of the arts, future studies exploring empirically possible differences between judgements of liking and beauty are called for. Future investigations on this topic should consider co-occurrence patterns of different linguistic concepts that might reflect different mental constructs by applying corpus-linguistic or qualitative approaches. In addition, such future studies should also aim at explaining a differentiation between different forms of aesthetic appreciation by integrating psychological models of aesthetic appreciation and experience as well as an explanation of the underlying psychological processes (for a review on current psychological models of art experience for the visual arts, see for example Pelowski et al., 2016; see also Jacobs et al., 2016, this issue). Whether poetry is the appropriate genre for differentiating judgments of beauty and liking is, however, an open research question. Alternatively, other literary genres of fictions or media forms, such as movies that foster an involvement of readers and viewers with the expression of ugliness, disgust and horror might be more prone for a differentiation of different forms of positive evaluations.

Summing up, our study indicates that sad poetry indeed is appreciated more than joyful poetry. Furthermore, the higher our affinity to poetry in general, the higher our positive evaluations tend to be, independent of a poem's emotional content.

All experimental procedures were ethically approved by the Ethics Council of the Max Planck Society, and were undertaken with informed consent of each participant. Human participants read and rated six German poems each. No vulnerable populations were involved.

MK and WM jointly designed the study and interpreted the data. MK compiled the poetic corpus, gathered behavioral data, conducted data analyses and wrote the paper.

Data acquisition for this paper was made possible through the support of the Research Cluster Languages of Emotion (EXC302), which was funded by the German Research Association DFG and hosted by the Freie Universität Berlin. The writing and data analysis was conducted at the Max Planck Institute for Empirical Aesthetic in Frankfurt am Main, Germany.