Snakes Represent Emotionally Salient Stimuli That May Evoke Both Fear and Disgust

Rádlová, S.; Janovcová, M.; Sedláčková, K.; Polák, J.; Nácar, D.; Peléšková, Š.; Frynta, D.; Landová, E.

doi:10.3389/fpsyg.2019.01085

ORIGINAL RESEARCH article

Front. Psychol., 09 May 2019

Sec. Evolutionary Psychology

Volume 10 - 2019 | https://doi.org/10.3389/fpsyg.2019.01085

Snakes Represent Emotionally Salient Stimuli That May Evoke Both Fear and Disgust

$\r\nS. Rdlov&#x;$ S. Rádlová^1†

M. Janovcová^1,2†

K. Sedláčková^1,2

J. Polák^1,3

D. Nácar¹

Š. Peléšková¹

D. Frynta^1,2*

E. Landová^1,2

¹National Institute of Mental Health, Klecany, Czechia
²Department of Zoology, Faculty of Science, Charles University, Prague, Czechia
³Department of Psychology, Faculty of Arts, Charles University, Prague, Czechia

Humans perceive snakes as threatening stimuli, resulting in fast emotional and behavioral responses. However, snake species differ in their true level of danger and are highly variable in appearance despite the uniform legless form. Different snakes may evoke fear or disgust in humans, or even both emotions simultaneously. We designed three-step-selection experiments to identify prototypical snake species evoking exclusively fear or disgust. First, two independent groups of respondents evaluated 45 images covering most of the natural variability of snakes and rated responses to either perceived fear (n = 175) or disgust (n = 167). Snakes rated as the most fear-evoking were from the family Viperidae (Crotalinae, Viperinae, and Azemiopinae), while the ones rated as the most disgusting were from the group of blind snakes called Typhlopoidea (Xenotyphlopinae, Typhlopinae, and Anomalepidinae). We then identified the specific traits contributing to the perception of fear (large body size, expressive scales with contrasting patterns, and bright coloration) and disgust (thin body, smooth texture, small eyes, and dull coloration). Second, to create stimuli evoking a discrete emotional response, we developed a picture set consisting of 40 snakes with exclusively fear-eliciting and 40 snakes with disgust-eliciting features. Another set of respondents (n = 172) sorted the set, once according to perceived fear and the second time according to perceived disgust. The results showed that the fear-evoking and disgust-evoking snakes fit mainly into their respective groups. Third, we randomly selected 20 species (10 fear-evoking and 10 disgust-evoking) out of the previous set and had them professionally illustrated. A new set of subjects (n = 104) sorted these snakes and confirmed that the illustrated snakes evoked the same discrete emotions as their photographic counterparts. These illustrations are included in the study and may be freely used as a standardized assessment tool when investigating the role of fear and disgust in human emotional response to snakes.

Introduction

Due to a long co-evolutionary history with snakes, both humans and non-human primates evolved specific neural mechanisms for rapid snake recognition (Isbell, 2006; LoBue and DeLoache, 2008; Öhman et al., 2012; Van Le et al., 2013; Baynes-Rock, 2017). Among evolutionarily irrelevant (neutral) stimuli, snake pictures act as strong distractors (Soares et al., 2009a) and are detected faster (LoBue and Deloache, 2011; Soares et al., 2014) than, for example, flowers and mushrooms, but not faster than stimuli of modern threats such as guns (Fox et al., 2007; Zsido et al., 2018b). Moreover, EEG studies show that neural processing of snake stimuli is prioritized when compared to other animals such as spiders and birds (van Strien et al., 2014).

LoBue and Deloache (2011) provide evidence that it is the distinctive coiled snake morphology that attracts prioritized human attention. However, recent research has shown that the pattern of snake scales is also important. The human brain reacts much faster to pictures of snake skin than similarly, colored bird feathers (van Strien and Isbell, 2017). Moreover, some naturally occurring shapes and patterns are perceived negatively, and processed faster than other patterns, such as sharp edges (Guthrie and Wiener, 1966; Bar and Neta, 2006, 2007), zig zag patterns (Üher, 1991), or strong contrasting patterns in general (Näsänen et al., 2001). Souchet and Aubret (2016) proposed that venomous snakes adopt this phenomenon and use contrasting patterns and morphology with sharp edges as an aposematic signal to deter enemies and communicate their dangerousness. The ability to recognize danger presented by a snake by only a pictured skin detail was also observed in Vervet monkeys (Isbell and Etting, 2017), and Capuchin monkeys were able to recognize whether the presented snake was dangerous or harmless only based on the skin pattern (Meno et al., 2013). On the other hand, Prokop et al. (2018) showed that aposematic coloration did not play a crucial role in eliciting high fear of snakes as both black-and-white and color images of aposematic and cryptic snakes evoked a similar level of increased fear.

Both the typical snake body shape and color pattern seem to be important elicitors of neural mechanisms for rapid snake detection, although it remains unclear what is the typical snake pattern that elicits these reactions. Nowadays, as much as 3 709 snake species from 25 families (Uetz et al., 2018) are described. Many of them differ in ecology, size, skin pattern, behavior, and other aspects. Similarly, snakes vary as to the dangerousness they present to humans, including the efficiency of venom and its delivery system, the snake’s size, aggressiveness, and the probability of human encounters. Deadly venomous species can be found within all snake types, i.e., ground, arboreal, fossorial, and aquatic. However, the highest risk is related to two types of snakes (Kasturiratne et al., 2008). First, the venomous vipers or viper-like species are passive predators that use an ambush strategy for hunting. Their dangerousness lies in the unexpected risk, such as the possibility to step on the snakes by mistake (though some species actively warn potential enemies using demonstrative attacks or various acoustic signals such as hissing or rattling). The second type of snakes posing a high risk for humans are the elapids. No less venomous than the vipers, these active predators are much more mobile. They also rely on acoustic or visual signals (e.g., the typical cobra stance) as well as their high speed, which allows them to actively avoid an unwanted confrontation with humans or other enemies (Valenta, 2008). In Africa, the continent of human origin (Grine et al., 2009), the highest number of death by snakes is caused by vipers (Trape et al., 2001), especially the West African carpet viper (Echis ocellatus) and Gaboon viper (Bitis gabonica; Chippaux, 1998). Each of these snakes is characterized by specific morphological traits that can be easily recognized, e.g., the viperids possess a triangular-shaped head and sharp, visually discrete scales. Many deadly snakes have contrasting (aposematic) skin coloration patterns, by which they make themselves clearly visible. It is possible that humans do not perceive all snake species as one, generally threatening stimulus, but rather distinguish among the snakes that are deadly and react appropriately to the actual threat.

Most studies focusing on snake fear ignore the enormous variability of snakes. However, little is known about the actual effect of particular morphology of snakes on human emotional reactions. Traditional assumptions in psychology research is that the primary emotion involved in ophidiophobia is fear (Öhman and Mineka, 2001, 2003; Soares et al., 2009b), however, our recent studies show that disgust is involved in emotional evaluation of snakes and other vertebrates as well (Polák et al., submitted; Frynta et al., unpublished; see also Prokop and Fančovičová, 2013). Similarly Davey (1994) has shown that snakes are rated as fear-evoking as a result of elevated disgust sensitivity levels, and it is possible that some snake morphotypes may elicit a primary emotion of disgust. Disgust is also of prior interest to clinical researches as increased propensity (i.e., individual tendency to experience disgust) and sensitivity to disgust (i.e., negative appreciation of experiencing disgust) has been demonstrated to play a significant role in etiology of various psychological disorders ranging from specific animal phobia (e.g., arachnophobia: de Jong and Merckelbach, 1998; Woody et al., 2005; Muris et al., 2008; Olatunji et al., 2011) and blood-injection-injury phobia (Sawchuk et al., 2000; Cisler et al., 2009b). Therefore, it may also be involved in the onset and maintenance of snake phobia (Klieger and Siejak, 1997).

Both fear and disgust are considered to be basic emotions, with a universal distinctive facial expression and physiological response among humans and non-human primates (Ekman, 1992). Furthermore, some authors argue that fear and disgust could have opposing effects on sensory perception and attention (Buck et al., 2018). From a biological perspective, the two emotions are similar, as their purpose is to induce an adaptive reaction to life-threatening stimuli, increasing the chances of survival. However, fear arises in cases of immediate threat, which appears suddenly, unexpectedly, and presents a direct risk of injury or death. The resulting physiological reaction leads to activation of the sympathetic nervous system and a cascade of events, including rapid heart and respiratory rates, increased blood pressure, dilated pupils, muscle contraction, and increased perspiration (Barrett et al., 2016). In contrast, disgust is a reaction to stimuli perceived as potential sources of contamination or pathogens (Curtis et al., 2004; Curtis, 2011). The threat is not imminent, but rather takes effect after a prolonged period of time. The physiological reaction to a disgusting stimulus is more variable. Compared with fear, the parasympathetic pathways are activated (de Jong et al., 2011), leading to either an increase or decrease in skin conductance and blood pressure (Stark et al., 2005) and reduced heart rate and respiration (Gilchrist et al., 2016). The reason for this may be the fact that disgust is much more variable emotion than fear; thanks to this pre-adaptation, it evolved into a number of other functional circuits, fundamentally different from its original purpose. Disgust is thus linked to various other triggers such as violation of social norms, political beliefs, sexual orientation, and people from other social groups, etc. (Rozin et al., 1999).

The elicitors of fear and disgust also differ in many aspects. Specific characteristics of an animal can trigger specific emotions, and these characteristics can differ even within an animal group otherwise positively evaluated by humans. A good example is frogs and other amphibians, a vertebrate class which generally is thought to elicit disgust (Angyal, 1941; Davey, 1994; Tomažič, 2011; Prokop et al., 2016). Among them, frogs are listed as one of the animals that are very often objects of specific phobias associated with strong disgust feelings (Doctor et al., 2010). However, only a few frog species with specific characteristics are ranked as disgusting. These characteristics include a round, chunky body shape, small eyes, warts, drab colors, pink color, and white light reflection that gives the impression of a slimy object (Frynta et al., unpublished). Moreover, even mammals, generally a very popular and well-perceived group of animals, contain species with characteristics that make them look dangerous or disgusting. For example, underground-dwellers of a small body size, round shape, and reduced eyes, e.g., mole rats (Bathyergidae), marsupial moles (Notoryctidae), and moles (Talpidae) are often seen as ugly or disgusting (Landová et al., 2018b; see also Frynta et al., 2013 for the rankings of “beauty” and “ugliness”). In contrast, mammals who are thought to be the most dangerous and capable of evoking the highest fear are generally of large body size, such as big cats (although these can also elicit positive emotions at the same time; Landová et al., 2018b; see also de Pinho et al., 2014).

Furthermore, when an animal has a combination of characteristics, the animal can elicit mixed emotions. A snake can be rated as beautiful (Marešová et al., 2009a,b; Frynta et al., 2011; Landová et al., 2012), i.e., eliciting positive aesthetic emotion (Schindler et al., 2017), and fear-eliciting at the same time (Landová et al., 2012, 2018a). It is easy to imagine that a mix of negative emotions, i.e., fear and disgust, could be elicited at the same time as well. Thus, to study emotions triggered by snakes correctly, one must choose the right experimental stimuli and mind the differences between various morphotypes.

When studying human-perceived emotions elicited by snakes, one must not forget that respondents with different susceptibility to snake fear or general disgust and/or anxiety may answer differently. A great amount of literature exists that describes differences in behavior of snake-fearful or phobic respondents in comparison to controls (see, e.g., Öhman and Soares, 1994; Öhman et al., 2001; Schaefer et al., 2014). Given all these differences between phobics and healthy controls, one might also expect variation between these groups in fear and disgust evaluation of snake pictures. Therefore, it is necessary to first test the responses of healthy subjects to fear-eliciting and disgust-eliciting snake images and subsequently administer the same images to snake phobics for a comparison. The participation of each emotion and the relationship between them is crucial when we try to better understand the different line of evidence for involvement of these emotions in causing specific animal phobias (Woody and Teachman, 2000; Cisler et al., 2009a).

There is a lot of picture databases with known evaluations of self-reported emotional responses (usually valence, arousal, and dominance, supplemented with emotional categories that include fear and disgust), e.g., IAPS (Lang et al., 1988; Mikels et al., 2005; Libkuman et al., 2007), NAPS (Marchewka et al., 2014; Riegel et al., 2016), SFIP (Michałowski et al., 2017), DIRTI (Haberkamp et al., 2017), or GAPED (Dan-Glauser and Scherer, 2011). However, these databases consist of pictures with very varying properties including background, colors, lightness, focus on detail, etc. The last mentioned database even includes a separate category of snake pictures, but again the pictures portray the animals in very different contexts and environments, including human hands grabbing the snakes. Moreover, the snakes lack taxonomic description and species identification (e.g., the GAPED database even includes one glass lizard among the snakes). And although these pictures may evoke the particular emotions, it is not known which part of the picture is the main elicitor – is it the snake itself, its particular appearance/position, or something else such as its interaction with the background? It is well known that even basic characteristics of pictures such as the color, luminance, and shape structure affect human evaluation of emotion or preference (Üher, 1991; Bar and Neta, 2006; Silvia and Barona, 2009; Lišková and Frynta, 2013; Lišková et al., 2015; Ptáčková et al., 2017; Landová et al., 2018b; Rádlová et al., 2018). Moreover, such stimuli may be suitable for usage in experiments measuring self-reported or physiological response to a single picture, but are inappropriate for studies requiring block-type design (such as fMRI or EEG), because the stimuli are too variable and could elicit different reactions within a block.

Without a precise selection of well-characterized and standardized stimuli, further studies focused on the variability among respondents may result in misinterpretations. Thus, the main goal of this study was to examine the pattern of variability among various snake stimuli and to develop one homogenous set eliciting fear and another homogenous set eliciting disgust. We are very aware that individual differences in snake fear and general disgust propensity may influence the evaluation of snake species. To examine the differences among the respondents in terms of snake-elicited fear and disgust, we collected additional data about the participants including questionnaires such as the SNAQ (Polák et al., 2016, see also Zsido et al., 2018a for its shorter version) and a self-reported response based on a method more suitable for this kind of experiment. However, due to length limitation, these data will be published in a separate article (in prep.) as this paper is mainly focused on the differences among the picture stimuli. In other words, in this paper, we aim to study the effect of different stimuli on the respondents and to create categories of snake stimuli based on exclusive reactions.

More precisely, the general aims of this study were to (1) examine the full morphological variability of snakes in terms of human emotional responsiveness and to analyse the specific characteristics contributing to the human perception of fear and/or disgust. Then, based on the results, to (2) create two sets of visual snake stimuli that would evoke exclusively fear or disgust in human respondents. Finally, we aimed to (3) confirm the results of Experiment 2 on a reduced set of illustrated stimuli that could be make freely available for download and for usage in experimental studies of human-perceived emotions elicited by snakes.

Experiment 1. Full Morphological Variability of Snakes

Materials and Methods

Selection and Preparation of Stimuli

In case of snakes, subfamilies represent the optimal taxonomical unit that reflect the most distinctive differences between the groups. Thus, we randomly selected one species from each of the extant snake subfamilies (45 in total; see Pyron et al., 2013), inhabiting various parts of the world. To avoid over-representation of a particular genera, we randomly selected a genus and then one of its species.

For each of the selected species, we searched a representative picture of an adult individual on the Internet. It has been shown in our previous study that relative fear rankings of photos and live snakes highly correlate (r = 0.78), thus using picture stimuli may provide valid results regarding human responses to snakes (Landová et al., 2012). Our search criteria included photos that were of a good quality and publicly available, i.e., licensed under the Creative Commons license or a written permission was given to us by the authors (For a full list of included pictures, see Supplementary Material 1). After collecting the pictures, we modified them to create a standardized setting (using GIMP 2.8, Kimball et al., 1995): we placed the snakes on a unified white background, scaled and rotated them to similar relative size and position with each snake facing the same direction (see Figure 1). The total of 45 picture stimuli (one for each of the selected species) was printed on matte 10 × 15 cm cards.

FIGURE 1

Figure 1. Examples of picture stimuli used in Experiments 1 (A) and 2 (B,C). The snake species are as follows, from left to right: (A) Rough-tailed Sand Boa (Eryx conicus) and Madagascar Tree Boa (Sanzinia madagascariensis), both photos by O. Šimková, used with a permission; Coral Cylinder Snake (Anilius scytale), photo by Eduardo Santos via Wikimedia Commons, modified. (B) Fear-eliciting snakes: Orlov’s Viper (Vipera orlovi), photo by Venom Tales via Flickr, modified; Ottoman Viper (Montivipera xanthina), photo by Benny Trapp via Wikimedia Commons, modified; Eastern Diamondback Rattlesnake (Crotalus adamanteus), photo by Pierson Hill, used with a permission. (C) Disgust-eliciting snakes: Southern Blind Snake (Anilios australis), photo by Stephen Zozaya, used with a permission; Northern Rubber Boa (C. bottae), photo by Gary Nafis, used with a permission; Brahminy Blind Snake (Indotyphlops braminus), photo by Alexandre Roux via Flickr, modified.

Testing Emotional Response to Snakes

To test the emotional response to snakes by human respondents, we adopted a widely used method of sorting picture stimuli according to a given scale (e.g., Marešová et al., 2009a,b; Lišková et al., 2015; etc.). The reason for which we chose this method is because the main aim of this paper is to examine the variability among the snake species in terms of human-perceived fear and disgust. In other words, we ask whether different snake species/morphotypes affect the respondents differently. For this kind of study, the sorting method is optimal as it maximizes the differences among species. However, as the acquired rankings for each stimulus are only relative, this method is not a good choice for studies focused on the variability among the respondents (e.g., whether snake-fearful respondents rank the snakes differently). For such kind of study, a different ranking method would be more suitable, such as the VAS, Borg, or Likert-type scale (Likert, 1932; Grant et al., 1999). Note that we also collected this kind of data, the results of which will be published in a separate article.

All respondents (n = 342) were Czechia residents (aged 19 – 67; mean age = 28.18; SD = 9.29). Sample included students and other respondents that volunteered based on informational pamphlets. Other volunteers were recruited using the snowball sampling method (Goodman, 1961; Biernacki and Waldorf, 1981). Our aim was to keep the ratio of respondents with biological and humanities education balanced.

First, each of the respondents filled a personal questionnaire about their age and gender, and signed an informed consent. In addition, each respondent provided us with a rank of their self-perceived affiliation toward snakes on a seven-point Likert scale (1 = I like snakes very much, I would like to keep one as a pet; 4 = neutral affiliation; and 7 = I dislike/hate snakes very much, I fear them; Wuensch, 2015). This variable is further referred to as “affiliation.”

Then, we distributed all 45 snake cards on a well-lit table in a random order. The respondents were asked to imagine the depicted snakes as real animals. Their task was to pick up the picture of a snake that was most fear-evoking, then to pick up the second most fear-evoking snake, until they picked up the least fear-evoking snake on the table. In total, 97 women and 78 men sorted the picture set according to perceived fear. Additional 167 respondents (94 women, 73 men) sorted the same picture set according to perceived disgust. Then, because the highest rank obtained this way (i.e., the 45th snake) corresponded to the lowest emotional response, we multiplied the ranks by -1 to obtain more intuitive results.

Explanatory Variables

To analyze the effect of morphometric characteristics of the snake species on the human responses, we measured the following variables on the snake photos using the ImageJ 1.40 g (Rasband, 1997): total body length, body width (also referred to as body size), head size (length from the tip of the nose to the end of the jaw), eyes size, neck size (in the thinnest part), and tail width (width of the tail base next to the cloaca). The number of pixels were transformed to centimeters relative to the picture size printed on a 15 × 10 cm card.

To examine the effect of colors on the respondents’ ranking, we used Barvocuc (Rádlová et al., 2016) to extract specific information about hues, lightness, and saturation of each of the stimulus pictures converted to the HSL colorspace. For a detailed description of the Barvocuc software, see Lišková and Frynta (2013) and Lišková et al. (2015). Colors were pre-defined to visually correspond to the snake pictures as accurately as possible: red < 330°; 15°), orange (corresponding to brown in all of the snakes) < 15°; 37°), yellow < 37°; 60°), green < 60°; 200°), blue < 200°; 240°), and violet-rose < 240°; 330°). Further, the red color was divided into two colors as in some snakes it corresponded to reddish brown, while in others, it corresponded to bright red. Three achromatic colors were defined on the basis of the saturation (S) and lightness (L) values, which covered the interval 0–1: black (L < 0.27), white (L > 0.8), and gray (S < 0.1). Mean S and L were also included as explanatory variables. The white background of the stimuli was set to transparent and was excluded from the calculation.

Additionally, we included two more variables extracted from Barvocuc: the pattern, computed using the edge detection method (Sobel, 1978), and opaque pixels, which is the sum of all non-transparent pixels. The latter variable also corresponds to the overall robustness of the depicted snake species. In order to improve normality, the portion of colors, mean S, pattern, and opaque pixels were square-root arcsine transformed prior to the analysis.

Statistical Analysis

To quantify and test the congruence in species ranking provided by different respondents, we adopted the Kendall’s Coefficient of Concordance. Significance of differences in mean rank among species was calculated by post hoc Friedman–Neményi test as implemented in R-package PMCMR. Prior to analyses, the raw ranks were transformed as follows: each value minus 1 was divided by the number of evaluated species minus 1 (44) and square-root arcsine transformed to achieve a normal distribution. A Principal Component Analysis (PCA) was performed to visualize the multivariate structure of the data sets. A Mann–Whitney U test was used as a non-parametric alternative for variables deviating from normality (raw sores). MANCOVA was applied to test the effects of independent explanatory variables. Contribution of the explanatory variables (constrains) to the rankings of the snakes was examined in the Redundancy Analysis (RDA) as implemented in the R package vegan (Oksanen et al., 2017). RDA is a multivariate direct gradient method. It extracts and summarizes the variation in a set of response variables (subjective evaluation of fear and disgust evoked by snakes) that can be explained by a set of explanatory variables. This analysis permits to plot both response and explanatory variables to a space defined by the extracted gradients and enables detection of redundancy (i.e., shared variability) between sets of response and explanatory variables. Statistical significance of the gradients was confirmed by permutation tests. Calculations were performed in R Development Core Team (2010) and Statistica 9.1 (StatSoft Inc., 2010).

Results

Agreement Among Respondents

The overall agreement among respondents on the snake fear (Kendall’s Coefficients of Concordance W = 0.254) and disgust rankings (W = 0.293) was somewhat low. To correct for potential incongruence caused by an unequal effect of snakes that were ranked as the top, we retested the responses after we divided the set into two halves, i.e., the first half containing 11 snakes rated as the least and 11 as the most fear/disgust-evoking (mean values), and the other half containing 23 snakes in the middle results showed a slight improvement in the agreement on the position of these species (fear-scale extremes W = 0.332, middle W = 0.091; disgust-scale extremes W = 0.409, middle W = 0.119). Post hoc tests revealed high proportion of significant comparisons among mean ranks of examined stimuli (see Supplementary Material 2a). This confirmed that the congruence among respondents was high enough to extract a reliable order of stimuli according to the examined emotions.

RDA analysis of the fear ranks, which included gender, age, and affiliation toward snakes, revealed that these variables explained only 3.1% of the full variability. Sequential “Type I” ANOVA (n permutations = 10 000) confirmed that only the effect of affiliation was significant: affiliation: F_1,171 = 2.272, p = 0.0188; gender: F_1,171 = 1.728, p = 0.0724; and age: F_1,171 = 1.461, p = 0.1315). A reduced model, which included only the affiliation, explained 1.29% of the full variability (ANOVA: F_1,173 = 2.257, p = 0.0214). Similar results were found in case of the disgust ranks, although the full model explained more variability (7.57%): affiliation: F_1,163 = 10.715, p < 0.0001; gender: F_1,163 = 1.2665, p = 0.2091; age: F_1,163 = 1.3744, p = 0.1553. Reduced model explained 6.08% variability; ANOVA affiliation: F_1,165 = 10.673, p < 0.0001. Since there was no significant effect of age nor gender, the data from both genders were pooled for all further analysis.

Factors Determining Fear and Disgust Rankings

We employed RDA to examine the contribution of various explanatory variables to the rankings of fear evoked by snakes. We used the automatic model-building feature based on both the Akaike criterion and permutation P-values. Both methods agreed on inclusion of the following variables into the reduced model, which were confirmed as significant by the sequential “Type I” test (n permutations = 10 000): head size (F_1,39 = 12.095, p < 0.0001), red color (F_1,39 = 4.778, p = 0.0029), pattern (F_1,39 = 3.952, p = 0.0070), body size (F_1,39 = 4.096, p = 0.0038), and blue color (F_1,39 = 2.615, p = 0.0306). The RDA model has generated five constrained axes, which explained 41.39% of the full variability (Figure 2A).

FIGURE 2

Figure 2. Redundancy Analysis (RDA) of the characters determining the ranks of fear and disgust elicited by snake stimuli used in Experiment 1. (A) Analysis of the fear ranks. The model explained 41.39% of the full variability, RDA1 = 0.2784% of variability and RDA2 = 0.0714% of variability. (B) Analysis of the disgust ranks. The full model explained 44.77% of the variability, RDA1 = 24.75% of variability, RDA2 = 10.76% of variability. For a better visualization of the fear/disgust ranks polarity, 10 species with the largest and lowest mean ranks of fear/disgust are mapped onto the pictures. For the description of the snake species abbreviations, see Supplementary Material 1.

The same procedure was used to find out which variables affect disgust rankings. The final model included the body size (F_1,38 = 6.871, p = 0.0002), green color (F_1,38 = 6.353, p = 0.0002), eyes size (F_1,38 = 5.316, p = 0.0009), red color (F_1,38 = 4.570, p = 0.0015), pattern (F_1,38 = 4.549, p = 0.0019), and yellow color (F_1,38 = 3.148, p = 0.0150). The six constrained axes explained 44.77% of the full variability (Figure 2B).

Scores of the first axes on the RDA visualization of fear (Figure 2A) and disgust (Figure 2B) rankings correspond to the mean rankings of the respective emotions; RDA1 scores x mean fear: r² = 95,4%; RDA1 scores x mean disgust: r² = -95,8%. The results show that similar characteristics affect the rankings of fear and disgust, but in the opposite directions. Big brightly red-colored snakes with an expressive head and complex scale pattern were ranked as the most fear-evoking, but the blue color present on some snakes in form of a light reflection caused them to be perceived as non-fearful. On the contrary, bright green, red, and yellow colors, large size, big eyes, and a complex pattern characterized snakes perceived as the least disgusting.

Discussion

Most participants agreed on the position of snakes that were ranked as the most fear- or disgust-eliciting, and also on the snakes who are least fear- or disgust-eliciting. The reason behind this may be that the snakes “in the middle” of the scale may either evoke none of the studied emotions at all, or evoke both emotions at the same time, hence the disagreement. Contrastingly, the respondents agreed on the snakes placed at the extreme ends, which suggest that these snakes unambiguously either evoke the given emotion, or do not evoke it at all. The RDA analysis revealed the characteristics specific for snakes ranked as the most fear-eliciting – these included the large body size, bright red coloration, body shape with a distinctive head (as compared to the non-fear eliciting snakes with a small, almost indistinguishable head), and complex pattern. This pattern may include both contrasting ornaments and/or distinctive, raised scales, typical for some groups such as the vipers.

The results suggest that the same characteristics determine whether the snake will be rated as highly fear-evoking or disgusting, only in the opposite direction. This really applies to a number of characteristics: bulky, red, textured snakes are fear-evoking, while thin snakes with low red coloration and smooth texture are disgusting. This overlap of characteristics leads into a partial overlap (correlation) of the fear and disgust axes (r² = 38.39%, p = 0.00001). However, each emotion can also be elicited by other specific characteristics, e.g., small eyes and dull coloration in the case of disgust-eliciting snakes. In Landová et al. (2018a), certain snake characteristics are associated with high fear rankings, including a short body length, wide head, and overall darkness. Moreover, the characteristics specific for disgust and fear elicitation are not mutually exclusive and can be present on one snake, e.g., the Northern Eyelash Boa (Trachyboa boulengeri) obtained high ranks of both fear and disgust. This snake has dull, brown coloration and a chunky body with sharp, distinctive scales, and represents a good example showing that when specific characteristics mix, the snake can evoke a combination of emotions.

Our findings concur with recent descriptions that European respondents fear the viper and viper-like snakes (Landová et al., 2018a), in which we showed that the evaluation of vipers as the most dangerous snakes is shared across people from Europe and Near East. As the only dangerous snakes in Europe (and adjacent Asian areas) are the vipers, these results may not be that surprising. However, we obtained similar results from studies in sub-Saharan Africa (Frynta, in preparation), where people also mainly fear vipers, even though there are deadly elapids present. One explanation can be that the vipers, being ambush predators, want to warn others of their presence and thus use specific characteristics, such as sharp edges and contrasting patterns, as aposematic signalization to deter enemies (Souchet and Aubret, 2016). Another explanation may be that the vipers are rather very cryptic and unseen (wanting to merge with the background to ambush a prey), but humans, being often unintended victims of encounters with vipers, perfected attentional mechanisms to effectively recognize and avoid these snakes. More research is needed to resolve this. Moreover, position of the snake plays role in human emotional response. Snakes in a striking posture with a coiled body and S-shaped contracted neck ready to attack are detected faster (Masataka et al., 2010) and trigger more intensive fear (Schaefer et al., 2014).

In most cases, the snakes ranked as the most disgusting were small, harmless species of blind snakes (Typhlopidae) living underground. As these species pose no danger to humans (both in terms of bite and possible contamination; O’Shea, 2018), we can assume that they evoke disgust only due to their close resemblance to phylogenetically unrelated, but morphologically very similar species (such as maggots or parasitic worms causing diseases; Muller and Wakelin, 2002). These animals share slimy look, pale pink coloration, little to no eyes and visually undefined (indistinguishable) head and tail, i.e., characteristics identified as disgust-evoking in this study.