Emotions and temperature are closely related through embodied processes, and people seem to associate temperature concepts with emotions

Barbosa Escobar F, Velasco C, Motoki K, Byrne DV, Wang QJ (2021) The temperature of emotions. PLoS ONE 16(6): e0252408, Jun 3 2021. https://doi.org/10.1371/journal.pone.0252408

valance >>> valence

Abstract: Emotions and temperature are closely related through embodied processes, and people seem to associate temperature concepts with emotions. While this relationship is often evidenced by everyday language (e.g., cold and warm feelings), what remains missing to date is a systematic study that holistically analyzes how and why people associate specific temperatures with emotions. The present research aimed to investigate the associations between temperature concepts and emotion adjectives on both explicit and implicit levels. In Experiment 1, we evaluated explicit associations between twelve pairs of emotion adjectives derived from the circumplex model of affect, and five different temperature concepts ranging from 0°C to 40°C, based on responses from 403 native speakers of four different languages (English, Spanish, Japanese, Chinese). The results of Experiment 1 revealed that, across languages, the temperatures were associated with different regions of the circumplex model. The 0°C and 10°C were associated with negative-valanced, low-arousal emotions, while 20°C was associated with positive-valanced, low-to-medium-arousal emotions. Moreover, 30°C was associated with positive-valanced, high-arousal emotions; and 40°C was associated with high-arousal and either positive- or negative-valanced emotions. In Experiment 2 (N = 102), we explored whether these temperature-emotion associations were also present at the implicit level, by conducting Implicit Association Tests (IATs) with temperature words (cold and hot) and opposing pairs of emotional adjectives for each dimension of valence (Unhappy/Dissatisfied vs. Happy/Satisfied) and arousal (Passive/Quiet vs. Active/Alert) on native English speakers. The results of Experiment 2 revealed that participants held implicit associations between the word hot and positive-valanced and high-arousal emotions. Additionally, the word cold was associated with negative-valanced and low-arousal emotions. These findings provide evidence for the existence of temperature-emotion associations at both explicit and implicit levels across languages.

General discussion

In the present study, we aimed to uncover how people associate a range of adjectives spanning the emotional circumplex model with different temperature concepts. To this end, we conducted two experiments. In Experiment 1, we evaluated the explicit associations between twelve different emotion adjectives, varying in valence and arousal, and five different temperature concepts on native speakers of four different languages (English, Spanish, Japanese, and Chinese). In Experiment 2, we evaluated native English speakers in terms of their implicit associations between temperature words (hot and cold) and emotion adjectives at the opposite ends of both the valence (Unhappy/Dissatisfied and Happy/Satisfied) and the arousal (Passive/Quiet and Active/Alert) dimensions.

Altogether, the results provided evidence for the existence of explicit and implicit associations between emotions adjectives and temperature concepts. The results of Experiment 1 showed that, regardless of language, the peak of the association ratings moved counterclockwise in the canonical circumplex model of core affect from the lower left side (third quadrant) to the upper left side (second quadrant) as temperature increased from 0°C to 40°C. The results of the IATs in Experiment 2 revealed that participants had faster response times when the word hot was independently matched with the positive-valence and the high-arousal emotion words, than when these emotion words were matched with the word cold. Furthermore, as evidenced by the magnitude of the D values, the associations in the arousal dimension were stronger than in the valence dimension, potentially due to a more linear relationship between temperature and arousal, compared to valance. Therefore, consistent with Experiment 1, the results of Experiment 2 revealed that participants implicitly associated the word cold with the low arousal emotion and the word hot with the high arousal emotion. While some studies have hinted at the existence of temperature-emotion associations, to the best of our knowledge, this is the first study to uncover associations between them explicitly and implicitly and explore their similarity across languages.

Our results may be interpreted from the theory of grounded cognition [51, 53]. It is possible that the associations uncovered here arise from the multimodal representations encoded in the brain incorporating the temperature (body or ambient) experienced during specific emotional states repeated. For instance, the robust associations of 20°C with positive valence, low arousal emotions can be a product of the physical comfort this temperature generates for most people. It is also possible that the associations between emotions and temperature arose because there is causal relationship between them, whether it is in the case where emotions trigger physiological responses that affect bodily temperature [92], or in the case where ambient temperatures trigger affective states [17].

The results of Experiment 1 showed that there was a positive relationship between temperature concepts and the arousal dimension of the emotions. In line with our hypotheses, the results also revealed an inverted U-shaped relationship between temperature concepts and valence since the extreme temperature concepts, both cold and hot, were associated with negatively valanced emotions, whereas the milder ambient temperature was associated with positive valanced emotions. A possibility is that valence is related to embodied process of comfort as warmer temperatures are comfortable but extreme temperatures at both ends can generate discomfort [47]. These results are consistent with Wilkowski et al. [93] as the authors suggested people from different cultures use metaphorical expressions of hot and negative emotions (e.g., anger). These results also are also in line with Baylis et al. [62], who found that expressions of positive emotions in social media were the highest at 20°C and decreased beyond 30°C, at which point negative emotions also increased. It is worth noting that the present study did not control for whether participants interpreted the temperature concepts presented as coming from the environment or from a specific object, despite the visual representations used. Hence the temperature range considered comfortable might differ. The results of Experiment 2 were partially consistent with studies that have implied that warmer temperatures are positively valanced [17, 59, 60]. It is important to note that only two temperature words were used. Experiment 2 also revealed that the association between temperature and arousal was more robust than that between temperature and valence, potentially because associations with valence at higher temperatures is less clear as hotter temperatures can be evaluated positively or negatively, as Experiment 1 showed.

Furthermore, people may associate high arousal emotions, whether they are positively or negatively valanced, with higher temperatures because body temperature increases when they experience those emotions. Some studies that have shown that the temperature of peripheral body regions decreases during negative-valanced, high-arousal emotional states [31, 67, 94, 95]. Nevertheless, other studies have indicated that body temperature changes which are triggered by emotions generally accompany arousal, but are independent of the valence of the emotions [92, 96], which seems to be consistent with the associations of the higher temperature concepts in Experiment 1 and the smaller difference across dimensions of Experiment 2. Examining the inverse relationship between temperature and emotions, in which certain temperatures trigger specific emotional states, the associations can come from high ambient temperatures or activities that increase body temperature and hence arousal. For example, physical exercise increases body temperature and at the same time may increase excitement and energy levels. Similarly, it is possible that the associations between positive emotions with low levels of arousal and ambient temperatures arise because at this temperature, people are at their homeostatic optimum [1, 12, 30] and therefore feel calm, secure, or happy.

The results of Experiment 1 showed that the associations across the four languages exhibited a high degree of similarity and followed the same overall direction towards the two dimensions of the emotions. These findings are consistent with other studies that have found large similarities in associations between emotions and colors [49, 82, 97, 98] and emotions and brightness [99] across languages. The large similarity in temperature-emotion associations can be the result of highly comparable concepts linked to emotions across languages, which can potentially be captured by broad categories. As Ogarkova [37] suggested, emotional categories in most languages have similar hierarchical structures and the variance of emotion lexicons can be explained by a few relevant dimensions. Another potential explanation of these results is that the subjective experience of emotions did not differ significantly across speakers of the various languages. It is possible that the emotion-temperature associations are fundamentally driven by core affect, which according to the constructionist theory of emotion, is parsed into specific emotion categories [24, 100]. As Sievers et al. [101] suggested, there is a high degree of similarity in how people understand expressions of emotional arousal since they are signaled with a multisensory code based on variations in magnitude. Our findings agree with Jackson et al. [27] in that they seem to reflect the existence of a common semantic framework of emotions across language based on valence and arousal, which are linked to neurophysiological systems that keep homeostasis, although there exists cultural differences.

Despite the high degree of similarity in the emotion-temperature associations across languages, small differences were present. These differences may arise because of linguistic discrepancies and what the various emotions mean in across languages, as well as countries [27]. As Lindquist [36] suggested, languages encode emotions differently, and emotional perception is culturally relative. Additionally, these differences may be caused by environmental factors and the degree of exposure native speakers of a given language that predominantly live-in specific countries have with different temperature ranges. For example, Jonauskaite et al. [49] found that the association between yellow and the concept of joy varied depending on overall exposure to sunshine. Temperature may affect the expression of affect, as well as the subjective experience of similarly intense affective stimuli [12].

Regarding the lower temperature associations for high-arousal emotions in Chinese-speaking participants, it is possible that these differences are the result of a restrained view of the experience and reporting of intense emotions [102]. Intriguingly, there was a slightly higher correlation in the associations between Chinese- and Spanish-speaking participants compared to that between Chinese- and Japanese-speaking participants, as based on geographical and linguistic distance, the latter should be higher [103]. It is possible that this was caused by a greater international cultural exposure from both language groups. However, further research is needed to strip out the effect behind these differences.

Limitations and future directions

One of the main limitations of the present work relates to the set of emotions used. While we focused on emotions that derived from the valence and arousal dimensions, the pool of emotions that can be studied is virtually endless, and other emotions that could have associations with temperature were not included. For instance, romantic or sensual emotions were not analyzed. Future studies may focus on associations with a much more precise set of emotions that have greater relevance for specific fields or applications. That said, the emotion adjectives [70] have been validated across cultures in 23 consumer studies (each with 104–270 participants) involving New Zealand and Chinese consumers. The adjectives were also validated with different types of stimuli (i.e., text, images, aromas, and taste). The emotion circumplex covers a wide range of relevant emotions while remaining parsimonious and is applicable to extensive classes of stimuli. Another aspect to consider when applying these temperature-emotion associations in real world scenarios, is that both temperature and emotions can be product- or context-specific. For instance, while companies may want to generate associations between refreshing beverages and positive emotions, using warm temperatures associations would not be ideal.

Another limitation comes from the method in which the temperatures were presented (visual representations in Experiment 1 and temperature words in Experiment 2). Since no actual temperatures were used, it is not possible to rule out potential semantic effects. People could have also interpreted emotion or temperature words differently, thus introducing some variability. In Experiment 1, people from different countries may not be equally used to certain temperatures. in Experiment 2, people could have had considered diverse temperature ranges for the words hot and cold. Nevertheless, the results provide considerable confidence since the experiment captured relative differences given its within-subjects design. The five temperatures and their visual representations (along with the specific values in°C and°F) in Experiment 1 were chosen as way to cover a broad range of the ambient temperature spectrum, reduce potential language biases, and increase familiarity with temperature measurements. However, it is not certain that participants thought about ambient temperature with these representations. Future studies could expand the range of temperatures and represent them in different ways so that the meaning of temperature is less ambiguous. Moreover, exploring potential differences in the associations between emotions and environmental and object-based temperatures could generate interesting insights. For instance, similar versions of IATs could be designed using pictures of objects or scenes evoking different temperatures combined with facial expressions, such as emojis. Another limitation, especially in the IATs, comes from the possibility that, when evaluating the emotion-temperature associations explicitly or pressing a key in the IAT, participants may not have read the entirety of the pairs of emotion words but instead relied only on the first word. That being said, the use of these emotion adjectives has been extensively validated in multiple studies [70–72].

In recent years, the interest in crossmodal correspondences has seen a rapid growth from academics and practitioners. Research on these correspondences has found a myriad of associations between different modalities (see [104]), and temperature-based correspondences has recently regained the interest of researchers relates to temperature [105, 106]. Spence [107] has recently reviewed the literature on temperature-related crossmodal correspondences. The present study provides valuable insights to advance the study of crossmodal correspondences since the explicit and implicit associations found here may help deepen the understanding of temperature-based crossmodal correspondences mediated by emotions and the role language might play in them. More specifically, these results can guide future studies on the mechanisms behind temperature-based crossmodal correspondences.

To conclude, our findings provide evidence of the existence of consistent associations between emotions and temperature concepts at the explicit level across languages. The findings also provide evidence that some explicit associations also translate to the implicit level. The present study also adds to the literature on emotions and their associations with abstract concepts, and to research on the bidirectionally causal embodied processes between emotions and temperature. Furthermore, the present article contributes to the discussion of how conceptual metaphors can help people understand abstract concepts by interpreting them in terms of concrete experiences, and how using these metaphors can change both how people view the world and their subsequent behavior.