When animals cry: The effect of adding tears to animal expressions on human judgment.

At a time of growing interest in and awareness about the relationships between humans and animals, it is of relevance to scientifically analyse the intrinsic nature of these interactions. Reactions to emotional tears show our extraordinary capacity for detecting micro-nuances when judging another human's face. Regarding such behaviour, previous studies carried out in our laboratory have pointed to an adaptive function of emotional tears: i.e. their inhibitory influence on perceived aggressiveness. In the present work we aimed to further explore that hypothesis by extending our investigation from humans to animals, using pictures of five different animal faces (cat, dog, horse, chimpanzee, hamster) to which tears were added digitally. To this end, we conducted an online study of 403 participants recruited from different social networks and academic institutions. We questioned the participants about their perceptions of emotional intensity, aggressiveness and friendliness in the animal faces and analysed the comparisons they made between faces with and without tears. In addition, a latent variable referred to as "passion for animals" was measured using different indicators. By adding the results obtained in each species and breaking them down into different basic emotions, we found that the presence of tears was related to a higher absolute frequency of participants who perceived sadness, which endorsed our previous results obtained using images of humans. Regarding aggressiveness, the presence of tears favoured the perception of less aggressiveness. A structural equation model was also conducted to explore the relations among all the measured variables. The model confirmed that the presence of tears in the animal faces had a significant influence on the perception of higher emotional intensity and friendliness, and of lower aggressiveness.

Introduction

Among primates, the detection of visual signals is crucial to recognize the emotions of others, being the eyes and the gaze the best sources for obtaining such an information [1-3]. Our large sclerae, accompanied by considerable eye mobility, make it easier to recognize the direction of the gaze and the object of attention of other humans, even in infants [4-6]. Such facilitated communication has several advantages for our theory of mind; that is to say, the awareness of what another person is thinking or knows in contrast to our own experience, an essential skill in an ultrasocial species such as ours [7, 8].

The most particular and unique feature of our ocular physiology, and one that allows us to transmit a wide range of information in the form of emotional inference, is the shedding of emotional tears.

There is much scientific evidence to show that humans are extraordinarily apt at detecting micro-nuances that alter their emotional perception when judging another human’s face, and our reaction to emotional tears is the perfect example [9]. Tears running down a cheek are capable of changing our bioelectric brain activity [10], and even alter the typical visual inspection pattern in such a way that tears act as magnets of attention [11].

While research on children’s crying has focused on its acoustic component as a distress call [12], visual stimuli are the most widely used when studying its influence as a signal in adulthood. Interestingly, this distress call maintains its value as a sign of need for help throughout life, helping the crier to obtain the support of the group [13]. As a consequence of the perception of helplessness, people see a crier as more friendly, and are more willing to help them by increasing social connectedness [14]. In the present work, we explore another peculiarity of weeping with strong evolutionary roots.

Hasson hypothesized that tears serve as natural brakes to stop aggression in our conspecifics [15], and evidence obtained since then supports this theory [16]. The mechanisms through which visible tears achieve their appeasing effect on a potential aggressor are unclear, but they appear to be related to at least two different types of inference. On the one hand, (1) they have a significant influence on the perception of the crying person’s vulnerability. The crier is not only perceived as more in need of help [10], thus awakening greater sympathy and empathy [17], but also generates a greater perception of incompetence [18]. On the other hand, (2) appeasement signals are only reliable if they prevent offensive actions in some way; however, by blurring vision, tears are a handicap in the case of having to defend against an attacker [15, 19]. Such appeasement signals are more likely in situations of hierarchical restructuring and other types of intra-group social contexts among equals, in which mortal aggression is unusual, where showing submission is an effective mean of communicating the cessation of agonistic behaviour.

A clever form of manipulation in order to study the function of tears and crying is the use of “poker faces”–faces which show little or no emotion at all [20] while crying (shedding tears); in this way, researchers can observe the effect of tears isolated from the emotional inference that usually accompanies marked emotional expressions. Among these studies, of which there are few, some have used calm crying expressions (see [21] for an in-depth definition) or human-like avatars [22, 23]. In this sense, and regarding the objectives of the present work, we believe that a new experimental way to further explore the effect of emotional tears on the receiver is to add visible tears to the face of different species of mammals, after first providing images in which the animals have a regular expression typical of its species. Animals represent an excellent opportunity to study the effect of tears on human observers, because they do not cry in response to emotional feelings (as commented above, emotional crying is a human feature; see [24] for an evolutionary overview). In this context, it is important to point out that the animal species selected for this study are all physically capable of shedding tears, but do not use this mechanism as an emotional signal to conspecifics [25]. In this way, such emotional signalling, if observed through human judgement, would represent a process of anthropomorphization.

The concept of anthropomorphism refers to the attribution of human-like properties, characteristics, or mental states to real or imagined non-human agents and objects [26]. It is a core concept discussed in HRI (human robot interaction) literature and, in general, in studies describing how humans relate to their computers and other media technologies (see the seminal proposal of the Media Equation theory, which proposes that humans interact with their technological devices “as if they were humans too” [27]). People tend to anthropomorphize robots more than other technology, although this process varies depending on factors such a movement and gestures, verbal communication, embodiment and, most importantly for our interests, the emotional response shown by the robot [28]. In fact, a robot is perceived as more anthropomorphic when it provides emotional feedback, rather than when its feedback is unemotional [29]. Regarding the psychological process of anthropomorphization itself, several comprehensive theoretical accounts have been suggested, especially in the abovementioned HRI context. Among these, a “dual model” has been proposed, in which an implicit, automatic and unconscious process of attributing human-like qualities to non-human agents could coexist with an explicit, cognitive-driven, motivated and conscious process in the same direction. An important implication of this is that the direct measures of anthropomorphism (e.g. questionnaires) can reflect either implicit or explicit anthropomorphism, or a mixture of both, depending on the person [26].

A process of anthropomorphization in humans when describing animals is known to exist, and has a longer history, with Charles Darwin providing an excellent illustration of the human mind’s capacity to see the same kinds of covert emotional states in the behaviour of non-human animals as it does in the behaviour of other humans [30]. This is unlikely to have an exact correspondence with the phenomenon of interaction with robots, given that animals have their emotions in physiological terms (but no feelings, according to their lack of a complex conscious behaviour). However, when we perceive that an animal is experiencing a concrete emotion or feeling we attribute a human feature to the animal, in which case we are indeed applying our anthropomorphism to the animal. Interestingly, this process can provide advantages to the animal, since individual differences in anthropomorphism predict the degree of moral care and concern afforded to an agent [31] and, moreover, empathy towards animals and humans is correlated [27]. In addition, our daily experience interacting with different types of animals has a long history and is much more extensive than the sporadic interactions we have with modern day avatars that express different emotional states, or non-generalized (till now) relationships with robots and other AI devices. Moreover, in line with this interest in the anthropomorphization of animals, some authors have explored its neural basis to show that we use the same neural mechanisms to attribute emotions to the facial expressions of humans and non-human animals [32].

In summary, by exploring how digitally added tears can change our judgments and inferences about animal faces, we have sought to take a step further after our previous work in which we added tears to human faces by opening up new lines of research into the psychological process of anthropomorphization, and increasing understanding of our perception towards other species’ emotions, which partly conditions human-animal interactions. The measurement of some human-like attributions (human emotions or traits, like “sadness” or “aggressiveness”) in regular animal faces can provide information about the process of anthropomorphization towards animals, while the addition of digital tears to their faces (making animals prone to be perceived as if they were crying) can enlighten us about the power of the presence of tears for such a process.

Objectives and hypothesis

In light of the abovementioned literature, we aimed to further explore the putative adaptive function of an inhibitory influence of tears on perceived aggressiveness by using pictures of animal faces with digitally added tears. In general, we expected the perception of emotional intensity in the facial expression to be greater when judging the “crying animals”. More specifically, we hypothesized that adding tears to the face of an animal would arouse that the animal would be perceived as more human-friendly. As a consequence, we expected that, when participants judged aggressiveness, crying animals would obtain a lower score. In addition, we explored whether different species of animals inspired different impressions when seen to be weeping. For example, we speculated that the species which are most popular among humans, such as domestic pets (i.e. dog, cat, and hamster), would have a differential impact on observers when their faces had visible tears. Finally, since we aimed to determine the most relevant variables related to crying, we decided to evaluate our data through a structural equation model that contemplated how previous fondness for animals related to the variables assessed.

Materials and methods

Participants

Four hundred and three participants aged between eighteen and sixty-four years old—94 males (M = 33.95, SD = 10.52) and 309 females (M = 31.61, SD = 11.04)—took part in the study as volunteers, without receiving any reward. To test our hypotheses, we conducted an online study in which participants were recruited from different social networks and academic institutions via different methods (e.g. newspaper ads, social networks and announcements during university classes). Participants were given a link to a survey administered via Google forms and the random redirector allocate.monster. Subjects were treated in accordance with our university’s Ethical Code of Conduct. Informed consent was obtained from all the participants, and all procedures were in accordance with the standards of the Comité de Ética de Investigación en Humanos (CEIH) from the University of Valencia (Spain), which approved the study, and with the 1964 Helsinki declaration and its later amendments.

Visual stimuli

The stimuli consisted of a set of photographs depicting close-ups of the faces of five different animals: a chimpanzee, a horse, a dog, a cat, and a hamster. The pictures were obtained from Google and were free of copyright. These original images were then modified by digitally adding visible tears with Adobe Photoshop Cc 2019. This yielded two groups of five pictures that were identical except for the presence (or not) of tears. An example of the same photograph (dog) with and without tears can be seen in Fig 1. The rest of the photographs can be found in the supplementary material. In addition, in order to evaluate the quality and characteristics of the stimuli employed, an experimental study was carried out (described in detail in the supplementary material; see link below). The added tears were shown to be easily visible, and all the selected animals were perceived to be typically representative of their species. Interestingly, the original image of the dog was perceived as emotional, hinting that, at least in the case of man’s best friend, a certain emotional inference can arise without the presence of tears. These latter results, together with the results of the main study, will be discussed in the context of their relevance to anthropomorphization and emotional inference tasks.

Fig 1

The original photograph of a dog’s face and its tearful version.

All supplementary materials can be freely consulted in https://osf.io/s9yev/?view_only=f91cccb72b5f4369bd7bf185764686c5.

Measures

After being presented with both versions of each photograph (with and without tears) the participant responded to a questionnaire regarding his/her impressions: (1) what type of emotion (if any) was expressed by the animal; (2) the intensity with which the emotion was expressed; (3) the friendliness perceived from the animal (considering “friendliness” to mean the tendency to perform acts of kindness towards others); and (4) the aggressiveness perceived from the animal’s expression.

To assess the perceived emotional expression (question 1), participants were asked to select a basic emotion (i.e. anger; surprise; happiness; fear; disgust; sadness) if they thought the animal was expressing such, or to indicate that the animal’s expression was emotionally “neutral” if they felt that was the case. We then applied a dichotomous classification based on previous works: emotion, if any, versus emotionless [11, 16]. “Emotionless” indicated faces with no perceived expression. To assess the remaining variables (questions 2-3-4), they were scored according to a 10-point scale, where 0 indicated the complete absence of intensity, friendliness or aggressiveness, and 10 the highest degree of these emotions. In addition, prior to the experimental procedure, participants were asked to answer 3 questions concerning their relationship with animals. (1) How much do you like animals? (on a scale of 0 /I don’t like them at all to 10/I am really fond of them; (2) How many animals have you lived with? (a scale of five, from none to one, two, three, or more than three); (3) How much importance have animals had in your life until now? (from 0/absolutely nothing to 10/extremely important). These three questions were used as indicators of a latent variable which we labelled “passion for animals”.

Procedure

After being recruited and obtaining their informed consent, participants answered the three questions about their relationship with animals. Participants were randomly assigned to one of two groups depending on the presence of tears in each animal model. In this way, Group 1 was presented with the tearful chimpanzee, the tearless horse, the tearful dog, the tearless cat, and finally the tearful hamster. For group 2, the presence/absence of tears was inverted. Thus, although all participants were presented with animal faces with and without tears, each participant saw only one version of each picture, which resulted in a between-subjects experimental design. In addition, the order of the pictures was randomly assigned for each participant; subsequently, he/she was presented with a picture of the first of the animal models and then completed an online questionnaire about its face, which was answered without a time limit. This process was repeated four times until the participants had been presented with the five animal pictures. A visual presentation of the procedure can be seen in Fig 2.

Fig 2

Visual outline of the procedure.

Participants answered a general questionnaire about their relationship with animals and were then presented with five photographs of the faces of five different animals. While each photograph was on the screen, participants answered four questions about the face in question, without any time limit being imposed. Participants had been randomly assigned to two different groups depending on the presence of tears in the photographs. Group 1 saw the tearful chimpanzee, the tearless horse, the tearful dog, the tearless cat, and finally the tearful hamster. For group 2 the presence/absence of tears was inverted, so that they saw the tearless chimpanzee, the tearful horse, the tearless dog, the tearful cat, and the tearless hamster.

Statistical data analysis

To assess perceived emotionality, we first used a dichotomic discretization: emotional (if any emotion was detected) versus emotionless. Emotionless indicated faces with no perceived expression at all. Fisher’s exact test for contingency tables [2 (tearless or tearful) x 2 (emotional or emotionless)] was used to calculate the odds ratio of perceived emotion on a face regardless of the species of animal. We also disclosed the odds ratio separately by species. Next, a 6 x 6 matrix of Pearson’s correlation coefficients was used to evaluate separately the strength and direction of the association between the variables measured (intensity of emotion, friendliness, aggressiveness) in the set of original photographs depicting the animals without tears and in the set of photographs with digitally added tears (the five species were pooled for this evaluation). We then carried out separate ANOVAs for each species and calculated post-hoc comparisons when required. Each of the former statistical analyses was carried out with R v.3.6.1 software. Finally, we performed and fitted a structural equation model to explain the causal relationship among variables by applying Mplus version 8 and a maximum likelihood sandwich estimator with robust standard errors [33].

Results

On emotionality and tears in animals

Considering the effect of visible tears on human faces, we expected that adding tears would dramatically raise the perceived emotionality of our models’ faces. First, we used the pooled species data and found that they were in fact associated with emotionality (odds ratio = 1.41, p < .001). Then they were detailed when analysing the effects of tears by species, to find that chimpanzee (odds ratio = 2.15, p < .01), horse (odds ratio = 2.60, p < .001), and dog models (odds ratio = 2.28, p < .01) reached statistical significance when judged as more emotional in the tearful condition. The hamster did not show significant differences (odds ratio = 1.14, p = ns), while the opposite was true for the cat (odds ratio = 0.55, p < .01).

In order to visually explore the above mentioned results to check if such a generally increased perception of higher emotionality was due to any particular basic emotion, we depicted the Fig 3, to find that the presence of tears was related to a higher absolute frequency of participants who perceived sadness (see Fig 3). Moreover, the absence of visible tears was frequently associated with a neutral expression and, interestingly, with an angry emotion (although later visual inspection of the data showed this latter observation was mainly due to the tearless image of the cat).

Fig 3

Cohen-friendly association plot.

For a two-way contingency table, the signed contribution to Pearson’s χ2 for the cell ij is , where f and e are the observed and expected counts corresponding to the cell. In this association plot, each cell is represented by a rectangle whose height is proportional to d and whose width is proportional to , so we can see that the area of the box is proportional to the difference in observed and expected frequencies [34].

On intensity of emotions

All the animals, except the cat, were associate with a higher mean intensity of emotions when tears were visible (Fig 4). In the dog’s case [F (1, 400) = 16.41, p < .001] a significant difference was found in the post-hoc test, with a Tukey HSD of 1.23. In the same way, the chimpanzee [F (1, 400) = 5.80, p = .016] was perceived to be much more emotional when tears were visible (Tukey HSD of 0.65). Regarding gender differences, women rated the cat and the horse models with greater emotional intensity than men [F (1, 400) = 10.10, p = .001] and [F (1, 399) = 6.61, p = .010] (Tukey HSD of 1.29 and 0.5, respectively). Interestingly, in the case of the horse [F (1, 399) = 5.12, p = .024], there was a significant interaction between the presence of tears and the participant’s gender. In the tearful condition, women found the horse to be much more intense than men, while men rated the picture with more intensity than women when tears were absent. Although both genders were sensitive to the tearful effect, the trend was more pronounced among women.

Fig 4

Bar plots per species and presence of tears.

On aggressiveness

Our hypothesis about less perceived aggressiveness is supported by the most consistent result of our experiments. The presence of tears was a significant influence in four of the five animal species when humans judged their faces (Fig 4); namely, the chimpanzee [F (1, 400) = 7.58, p = .006], dog [F (1, 400) = 7.94, p = .005], cat [F (1, 400) = 12.18, p < .001] and hamster [F (1, 400) = 5.08, p = .02], with Tukey’s honestly significant differences of between 0.49 and 0.85. In addition, there were main effects of the participants’ sex on the horse and dog models. Women rated aggressiveness lower than men, with F (1, 400) = 5.52, p < .05 (Tukey HSD of -0.56) and F (1, 400) = 4.50, p < .05 (Tukey HSD of -0.48), respectively.

On friendliness

All animals except the horse showed a higher mean friendliness when tears were visible (Fig 4), though only two reached statistical significance: the dog [F (1, 400) = 6.99, p = .008] and cat [F (1, 400) = 10.01, p = .001]. Post-hoc tests revealed a significant difference in favour of the tearful condition for both animals(crying animals were perceived as more friendly), with a Tukey HSD of between 0.65 and 0.73. Gender also proved to be a significant factor when the horse [F (1, 400) = 6.21, p = .01] and dog [F (1, 400) = 5.22, p = .023] were rated, with a Tukey HSD of between 0.59 and 0.65. Women perceived the tearful pictures of both the horse and dog to be more friendly than did men.

A summary of the means and their standard errors for the five species studied, with and without tears, and the three main dependent variables (intensity of emotion, perceived friendliness, and perceived aggressiveness) is provided in Fig 4.

Correlations between variables

As shown in Table 1, there was a strong correlation between love for animals and the importance of animals in our participants’ lives (r = .726, p < .001), and also between the importance of animals and the number of animals a participant had lived with (r = .493, p < .001). In addition, number of animals and love for them were significantly correlated (r = .422, p < .001), which was expected. Another interesting correlation emerged between the intensity of the emotion perceived and friendliness (r = .222, p < .001), while a significant association was not detected between intensity and judged aggressiveness (r = -.021, p = ns). In this context, one would expect a significant negative correlation between friendliness and aggressiveness (r = -.380, p < .001).

Table 1

Matrix of Pearson correlations with the means of the five species.

	Number	Love	Importance	Intensity	Friendliness	Aggressiveness
Number	-
Love	.422***	-
Importance	.493***	.726***	-
Intensity	-.039	.080**	.087**	-
Friendliness	.068**	.156***	.135***	.222***	-
Aggressiveness	-.020	-.074**	-.084**	-.021	-.380*	-

Note.

* p < .05,

** p < .01,

*** p < .001.

All p-values were corrected using the Holm-Bonferroni method for multiple comparisons.

Structural equation modelling of perception of aggressiveness

As demonstrated in Fig 5, indicators showed that the model fit the data in a satisfactory way, with a RMSEA of 0.04 (i.e. a close fit according to the guides [35-37], a CFI of 0.988 (well above the recommended value of 0.90) and a SRMR of 0.02 (considering a SRMR < 0.08 to be a good fit), although the χ2 statistic obtained for the model was 20.39 (df = 9, p = .015), which suggested a non-perfect overall fit of the model. We also tested the model by adding the variable Gender, but the indices of fit in said model were substantially lower, so we decided to present the more parsimonious first model, which had obtained the closest fit. The trend towards perceiving animals to be more friendly had an influence on the passion for animals, but the most stand-out finding was that tears had both direct and indirect effects on the perception of aggressiveness.

Fig 5

SEM of perception of aggressiveness with mediational effects.

Standardized estimates with two-tailed p-values. The importance that animals have in our participant´s life (IMP), the love felt for them, and the number of animals with which participants have lived are indicators of the latent variable labelled “passion for animals”. This passion is significantly and positively influenced by the trend towards perceiving animals to be friendlier, while the presence of tears is a significant influence on the intensity, friendliness, and aggressiveness perceived. As illustrated, tears not only have a direct effect on the perception of aggressiveness, but also indirectly, through the friendliness perceived. CFI = 0.988, RMSEA = 0.04, SRMR = 0.02, χ2(9) = 20.39 with p < .05. * p < .05, ** p < .01, *** p < .001.

Note that a table of means and standard deviations (or its bar plot version) per species, experimental condition (tearful vs tearless), and gender can be found as supplementary material at the link provided above (see Visual Stimuli), along with the visual stimuli, mixed effects models (as an alternative to the classical analysis presented), and the database.

Discussion

The main aim of this study was to assess the influence of visible tears on the faces of non-human animals on our perception of aggressiveness, friendliness, and emotional intensity.

As mentioned in the introduction, the shedding of emotional tears is a solely human feature, as non-human animals do not generate tears to express emotions, though they do display their emotional state in other ways [1, 38]. Our first analysis, in which we asked participants if “the animal is showing any emotion” confirmed the initial hypothesis: namely, tearful animals are perceived to express more emotionality. Thus, it was not necessary for the animals to shed emotional tears for human participants to perceive them as emotional, especially in the case of the dog, chimpanzee and horse models. Interestingly, our survey about the quality of the tearless images rendered a median rating of around 5 for every species except the dog, in a 0–10 points scale where 0 was “emotionless” and 10 “absolutely emotional”. Therefore, as expected, adding tears significantly increased the probability of human observers engaging in an anthropomorphization process towards the animals, attributing them an emotional expression they did not really have. The case of the dog was unique, as it was rated emotional even without tears, while the presence of tears increased the attribution of emotionality (this will be discussed below, with regards to domestic animals and their particularities). In terms of the type of basic emotion assigned to the tearful animals, sadness was the most frequently chosen. The fact that the presence of tears is related to an increased frequency of perceived sadness supports previous results obtained using images of humans [10, 21]. In our second analysis, when participants were asked about the intensity of the emotion perceived in the animals’ faces, we observed that the tearful animals were judged to be more intensely emotional, which is again in line with the above mentioned results. An exception was made for the cat, which will be discussed below with regards to domestic animals.

The results of our analyses of friendliness and aggressiveness extend our previous work on tears as appeasement signals [16] by highlighting a clear and widespread trend towards judging tearful animal models to be less aggressive. As we expected, the emotional stimulus (fake teardrops descending from the eyes to the cheeks of the animals) exerted a significant influence on the perception of the observers, and almost every animal model was perceived to be less aggressive when such tears were added to its face. Given that the study about the quality of the images indicated that the tears were perceived as easily visible and mostly realistic, we can argue that the observers judged the animals to be crying naturally, and displayed an emotional reaction to them in the line of what the literature reflects about the effects of tears on humans and their adaptive functions [14]. In addition, the results of our structural equation modelling analysis (SEM) supported our hypothesis: that there is a negative correlation between judging an animal to be more friendly and perceiving it to be less aggressive. In other words, we expected the two perceptions to be antagonists [39]. Indeed, friendliness and aggressiveness are associated with different physical gestures and involve distinct neural circuits [40]. This negative correlation was statistically significant in our study, and the structural equation modelling—which displays the pure effect of friendliness on aggressiveness—confirmed our expectations. Thus, in our causal model it was shown that tears influenced perceived aggressiveness in a direct way, causing the animals with tears to be perceived as less aggressive, and in an indirect way, through the expression of friendliness. Another interesting finding of the SEM analysis is the absence of a significant influence of emotional intensity on the perception of aggression. In fact, the presence of tears was a significant influence on perceived emotional intensity, but the latter did not affect judgment of aggression. Thus, we can affirm that, based on our structural equation model, emotional intensity does not affect perceived aggressiveness in any way, although emotionality is perceived in facial expressions more frequently when tears are present. Ito, Ong, and Kitada [41] found that the mere presence of tears on human faces with neutral expressions makes the viewer more prone to perceiving sadness, a result that we extend to different animal species in our study. In addition, our SEM analysis indicated that the latent variable labelled “passion for animals”, a compound of the relative importance people give to animals, the love they show them, and the number of animals they own, is positively influenced by the friendliness that people perceive in animal faces. In summary, and considering the results as a whole, we can affirm that human observers transfer some of the purported appeasement functions of the non-verbal communication signal of tearing to other species.

Regarding the pictures of the cat and the dog, we think they deserve some words apart, given they are the most common domestic animals. In this sense, both transmitted the appeasement effect of tears and were judged to be less aggressive in the tearful pictures, but we also observed a few interesting particularities which distinguished them. The dog was seen as more intensely emotional and to be less aggressive than the cat, and this was observed even in the tearless picture. The cat was judged to be the most aggressive animal and, surprisingly, to have a less emotional expression (less intense) when depicted with tears. To provide with a speculative reason as to why this difference occurred requires a deep understanding of the particularities of our relationship with dogs and cats. Both species have shown themselves to be capable of flexible learning, which refers to the ability to adapt their repertoire of learned behaviours to circumstances; however, early studies of comparative psychology [42] showed dogs to be much easier to train than their feline counterparts. Naderi et al. [43] believed that this difference was due to the eye contact dogs make with their trainers, since humans have preference for this type of exchange (e.g. gaze contact), and dogs possess the natural advantage of using the same visual signals as humans, with evidence showing they are capable of processing facial emotion [44]. This could explain how experience has resulted in a preference among humans to make intense eye contact with dogs (which dogs return), while interaction with cats is not subject to the same contingencies. In relation to this, it has been proven that domestic cats make use of vocalization, exploiting certain human sensory biases in relation to the acoustic component of babies’ crying to obtain attention and food [45]. Dogs also produce different vocalizations according to the emotional weight of the information they want to transmit [46], but visual communication seems to be favoured, especially in the case of some breeds [43, 47]. We suspect that this routine interaction with certain species that have coexisted closely with humans for thousands of years has sensitized us to a differential communication with animals. In other words, we might be more accustomed to observing the faces of dogs than those of cats or other species to try to infer their emotional state, and thus the process of anthropomorphization may arise more readily in our interaction with dogs (in this context, we feel it is not a coincidence that both the robots Spot, from Boston Dinamycs, and Miro, from Consequential Robotics, are dog-shaped). On the contrary, cat owners have problems identifying the emotional valence of their companions’ faces in an emotional inference task [48]; at best, they can interpret the emotional valence of their own cat’s signals in a reliable fashion, but not those of unfamiliar cats [49]. In any case, the reader should note that this vision of domestic animals together with our interpretation is tentative in nature and calls for caution when applying to actual relations between humans and animals. In addition, the differences observed here could also be due to any remaining particularities of the chosen image (breed, fur, natural expression of the concrete animal in the photograph…).

With respect to a few gender differences observed, women perceived a higher emotional intensity in the case of the cat and horse. There is evidence of a preference for horses among girls, while boys prefer wild animals [50], and females are more likely to label themselves as cat persons [51]. Perhaps the gender effect in terms of perceived emotional intensity can be explained by a training effect, since women are more likely to spend time searching for videos of cats and horses on the internet, and will therefore tend to be more used to interpreting the emotional intensity of these animals [52]. In addition, our women not only perceived the dog and horse to be more friendly, but also found them less aggressive. Such an effect with respect to dogs could be moderated by hormones, as there is a different response in oxytocin levels between men and women when they interact with their dogs [53], but this theory requires more empirical support.

Finally, we noted a few unexpected results (the “colder tearful cat” and the absence of a tear effect on aggressiveness in the horse) that were obtained with the pictures that were rated as less realistic in the quality study. This prompts an interesting reflection on the use of artificial stimuli to depict human emotions in non-human targets (animals and possibly robots), in the sense that, the more artificial and unreal they are perceived to be, the more difficult it is for them to produce empathy in us, and thus an anthropomorphic feeling.

Strengths and limitations

This research demonstrates the advantage of an adequate sample number that provides the desired power. Furthermore, as far as we know, this study is the only one in which animal images have been manipulated in order to explore the effect of visible tears on emotional and moral inference while isolating them from any explicit expression, thus allowing the influence of tears on human observers to be measured. The present findings are strengthened further by the fact that the animal faces were heterogeneous. However, a possible limitation of the present work is the small number of images used, given that a greater number of images of each species could have avoided the potential influence of the particularities of each photograph and species. The study on the quality of the images showed a few limitations like lack of perfect realism for some stimuli (horse, cat) and a variability in the extent to which some of the tearless stimuli were perceived as emotional (especially the dog). On the other hand, an excessive number of photographs could have affected the results, as tears can lead to a rapid habituation of the viewer; in this case the possibility of obtaining a non-significant result would increase. In addition, Prokop and Fancovicova [54] have found that animal species and their colours are important predictors of the response they provoke in human children, and it is possible that tears do not exert an effective influence in the faces of animals that evoke danger and disgust without tears.

Future research

In this study, the presence of tears on animal faces altered the emotional inference and perception of aggressiveness in our human judges. Previous studies have shown that certain animal species are sensitive to visual information provided by human facial expressions [e.g. 44, 53, 55, 56]. A further step for future research would be to assess whether these animals also respond to the minimal visual signal of emotional tears; specifically, future studies could present dogs with photographs of their owner’s with and without emotional tears. Also, and in order to explore in further depth the emotional inference of animal faces for human participants, it would be interesting to test the effect of animal age on our perceptions. According to Murube [57], emotional tears are the last to appear in the phylogeny and ontogeny of our species (preceded by basal and reflex tears), and our moral judgments may mimic this phenomenon by giving greater social value to tears emitted in adulthood. Human babies do not emit tears in response to affective stimulation until they are at least 6 weeks old [58], and this behaviour does not fully develop until the efferent parasympathetic pathway and supranuclear nerve connections mature [59], which usually occurs at 4 months. The gestures and sounds of emotional crying are more important in the early stages of childhood [60], while the visual component gains stimulus salience as the individual enters adolescence and adulthood. In addition, as Zeifman and Brown [12] reported, the significance of the changes in moral inferences produced by tears increases as we enter adulthood. In this way, babies benefit less from the presence of tears than older children, and the latter benefit less than adults. It would be interesting to determine if age influences human emotional inference with respect to animal faces without tears, as it does when tears are added.

Conclusions

Emotional tears are a special form of non-verbal communication that are unique to humans and which favour the inference of emotional states. They have adaptive advantages; among these, we have hypothesised that tears are an honest biological signal with a clear purpose of inhibiting aggression towards the crier in social contexts. In the present study we have extended the universality of this hypothesis by using animal faces to which we artificially added visible tears. By means of this experimental methodology, and a causal model for explaining the relationships among the variables evaluated, we provide empirical support for the notion that the presence of artificial tears on the face of an animal results in the human observer perceiving it to be less aggressive, possibly through a process of implicit anthropomorphization. In addition, such tears increase the perceived friendliness of the animal, and this influences the perception of its aggressiveness. Moreover, our results show that our passion for animals, a variable which includes how much an individual likes animals, how many animals he/she is living with, and how important animals are for the individual, is clearly affected by our perception of the friendliness of the animal, which increases with the presence of tears. Considering the results as a whole, it seems that the presence of tears improves the social relationship between humans and animals. Given that animals do not cry in natural conditions, the results observed here are presumably a consequence of some anthropomorphization process between the human observer and the animal, which leads us to propose that tears are the main element in this improved social communication. In conclusion, our results further endorse the notion of tears as an important biological signal that is essential for non-verbal communication.

9 Dec 2020

Pécs, Hungary

December 8, 2020

PONE-D-20-35135

When Animals Cry: The Effect of Adding Tears to Animal Expressions on Human Judgment

PLOS ONE

Dear Dr. Gadea,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised by the Reviewers, listed below.

Please submit your revised manuscript by Jan 23 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Joseph Najbauer, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for including your ethics statement:  "The study was under the standards of the Human Research Ethics Committee of the University of Valencia (Spain) and a written informed consent was obtained from all the participants.".

Please amend your current ethics statement to confirm that your named institutional review board or ethics committee specifically approved this study.

Once you have amended this/these statement(s) in the Methods section of the manuscript, please add the same text to the “Ethics Statement” field of the submission form (via “Edit Submission”).

For additional information about PLOS ONE ethical requirements for human subjects research, please refer to http://journals.plos.org/plosone/s/submission-guidelines#loc-human-subjects-research.

3. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

4. Please include a copy of Table 1 which you refer to in your text on page 12.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I have read this contribution on adding tears to animal faces with great interest! Furthermore, I am convinced that it has merit in demonstrating how tears on animal faces are subject to anthropomorphism mechanisms, and that animal tears may influence perceived aggressiveness and friendliness of animals (even if these relationships are not strong). The paper is generally well written, although I spotted several smaller language issues along the way - i.e., the manuscript could still benefit from some thorough proof-reading.

However, while I think there is reason to be excited about this work, I also saw some issues that limit the extent to which the findings can be interpreted. Perhaps most importantly, there was only a single exemplar for each species, and there was no pre-test data reported for the stimulus selection. E.g., it was unclear if all the baseline animal faces were indeed neutral in expression. Likewise, we do not know how typical these examples were. Would the same results be obtained for slightly different expressions, or different fur colors (affecting tear-visibility) among the same species? Conceptually, I felt that the role of "emotional tears" in this context was not so clear. Were these emotional tears at all? The answer seems to be that this was in the eye of the beholder - however, this would need to be introduced and discussed more clearly. Likewise, the reasoning about the difference between adding tears to neutral animal faces vs. neutral human faces (as in prior work) did not become clear to me. I believe there is a point here, but it would need to be better explained.

Methodologically, I was struggling with the decision to obtain forced choice judgments for the discrete emotions. I find it particularly difficult to interpret any kinds of differences between the baseline images (i.e., the images without the tears that were not digitally manipulated) given this question format. In addition, I believe this should be regarded as a mixed design (with repeated measures across a series of exemplars), even if the exemplars themselves were between-subjects. A related concern here is that the Google-based image selection may have turned up examples that already differed with respect to expressions before the modification. I would therefore find it important to include a breakdown of the results per animal species and emotions (instead of or in addition to figure 3). Finally, relative tear-visibility, and tear-realism may have been a concern here, as the animals differed in both fur color, and the number of tears (one or two). This might be addressed by collecting some post test data for typicality (of the animal pictures), tear-visibility, and tear-realism.

Despite these limitations, I overall found this paper to be very interesting and thought-provoking. It does lend support to the notion that humans might react to animal tears as if they were emotional, and that this might include the type of appeasement functions hypothesized by Hasson and others. I hope that the authors may find my more detailed comments below to be of aid in improving their manuscript.

Introduction:

One of the key arguments for conducting this study is the argument that, since animals do not cry in response to emotional feelings, adding visible tears to neutral emotional expressions presents an excellent opportunity to study the effect of tears. I would tend to agree with this general reasoning. However, the claim that because of this, emotional reactions in observers "can easily be attributed solely to the presence of tears" (p.4) may need some refinement and qualification. It would appear to be a bit more complicated than that. How does the fact that humans are the only species that sheds emotional tears influence the responses of participants? Does this require any conscious reflection by participants? E.g., Media Equation theory would seem to suggest that observers should respond to crying animal stimuli in a similar manner *as if* the tears were produced by humans.

Issues to be considered here are (1) Participants may be more likely to be aware that the tears were digitally added, even provided that the tears look very realistic. It may still be that participants did not question or otherwise consciously reflect upon the presence of the tears in these faces - but if they did, then their emotional responses may also have been affected by this realization. (2) What precisely is the key difference here between seeing tears added to a neutral human face, and seeing tears added to the face of an animal? Animal faces can still be perceived as sad, aggressive etc. (even without the addition of tears). To what extent, and in which way precisely, does the manipulation of the animal faces go beyond adding tears to neutral human faces? I believe the authors have an interesting point here - but I think this needs to be further elaborated and clarified. That said, I agree that it is very interesting to study tears on animals for which we have very different stereotypes (you said "constructs" here - what precisely is meant with that in this context?). Also, the point about daily experience and closer interactions with animals helps to differentiate this from work in HRI/HCI.

When raising the argument about anthropomorphism, I would suggest referring to some of the extant literature on zoomorphism as well. This is not an entirely new field, and it would appear to be highly relevant here. As some of the current references additionally relate to human-robot interaction, this could also include zoomorphic robots, such as Spot (Boston Dynamics), or Miro (Consequential Robotics).

Perhaps it would also help to discuss that the kinds of animals selected in this study are all capable of shedding tears - they just do not use this as an emotional signal to conspecifics. This signalling function appears to be what is included in the process of anthropomorphization.

Visual Stimuli

The example of the dog in figure 1, in my view, looks convincing. However, I have been wondering about some of the other animals. E.g., for the cat, the white fur may have presented a challenge for obtaining comparable visibility/salience of the tears. For the hamster, I wonder if the size of the tears relative to the eyes and the rest of the animal's face might have posed a challenge. I trust that all these materials looked well in the end. However, it would be great if they could be made available in full via supplementary materials.

Measures

Why was the basic emotion item presented as a forced choice? This would appear to exclude the possibility of mixed emotions. Also, I think this makes it difficult to argue that there was "greater sadness" (see abstract). Instead, would this measure not only yield a frequency with which sadness appeared to be dominant?

Procedure

It seems that this was a mixed design, rather than a (pure) between-subjects design - with repeated measurements from each observer for intensity, friendliness, and aggressiveness. A GLMM may be a more suitable statistical approach for these analysis, as it could be used to control for effects of observer identity.

Results

Emotionality - Odds ratios

Figure 3: This figure seems to be missing a scale on the y-axis. The figure caption helps somewhat - but I was still wondering why the squares have different widths.

Is there an explanation for why the results for the cat went in the other direction? Given that there are different stereotypes about these different animals, as well as substantial differences in the basic morphology of their faces, it seems likely that even the tearless images may have been perceived as showing a certain amount of non-neutral expressions. Was there any evidence for this? Did expressions for the tearless animals vary between animals? If yes, then it might be useful to include a figure that shows the baseline emotions for each tearless animal before pooling all of this data.

Intensity

What does it mean that the dog and the chimpanzee "benefited" from the presence of the tears? Do you mean that they were anthropomophized more (more intense perceived emotions)? For friendliness, this is more straight forward.

The gender effects for the horse condition are interesting and seem to align with the notion that personal relevance of the animal should matter.

Figure 4: This figure is much too small and unreadable in the manuscript. The downloadable version was readable, however.

Correlations between variables:

I unfortunately could not find table 1 anywhere in the manuscript, nor in manuscript central. However, the most relevant results appear to have been reported in the text.

SEM:

I found the effects of tears on aggressiveness examined in this analysis to be very interesting. However, it seems that the direct effect, albeit significant, was rather weak. Likewise, the link to perceived friendliness also did not appear to be very strong. Nevertheless, it would seem worth reporting. Perhaps this might be further substantiated in future work, as it could help to make the point that human observers might transfer even some of the purported appeasement functions of this signal to other species.

Unfortunately, I could not access the supplementary materials on OSF without logging in with my ID to request permission from the authors.

Discussion:

The discussion takes up the point again that emotional tears are a solely human feature that non-human animals are unable to produce. I found the argument about extending the work on appeasement signals to be very interesting, and perhaps more clearly formulated here than in the introduction.

However, I found the distinction between emotional and non-emotional tears to be rather fuzzy here. I.e., on the one hand, the argument is made that non-human animals cannot produce emotional tears. On the other hand, shortly thereafter (p.13), it is argued that domestic animals gained benefits from "the emotional stimulus". Please be clearer on how these arguments go together to avoid confusion about when you are talking about emotional tears, and when you are talking about non-emotional tears. Note that the argumentation in the introduction and stimulus design was that tears were added to neutral faces of animals. Where does the emotional stimulus come from then? If the stimuli looked realistic (which I assume they all did), should these tears then not have been non-emotional tears? I think that this potential source of confusion could be addressed - i.e., it may not be necessary for the animals to shed emotional tears for human participants to perceive them as such. Nevertheless, I think would still need to be made more explicitly. I would see the finding that participants perceived these ostensibly non-emotional tears as noteworthy.

The discussion mentions that some of the tears consisted of one teardrop, and some consisted of two teardrops. Given that there was only one exemplar per species, could this distinction have confounded some of the findings - e.g., concerning apparent differences between species? You might refer to prior work in this context that has looked at tear-intensity to address this issue at least partially (i.e., the amount of tears may not matter much). Still, this appears to be a relevant limitation that would call for further research.

On dogs and cats: While I found the discussion about differences in how we relate to cats versus dogs interesting, I felt concerned here that this might be over-interpreting things, given that there was only a single exemplar - a single image, for each. Again, I think it would be important to include in the results also the emotions that were perceived for the unmodified stimuli. Might this specific dog face have been perceived as sadder to begin with, and the cat face as emotional in some other way? Were they both regarded as completely neutral without the tears? I would suggest expanding a bit more on the basic literature about facial expressions in cats and dogs before raising the point about familiarity. E.g., this might start from the DogFACS and CatFacs (both https://www.animalfacs.com/).

SEM: On p.14, it is claimed that the absence of tears is associated with the perception of anger. Perhaps clarify that this discussion point is no longer about the SEM but about the forced-choice emotion rating task. Again, I was also not entirely convinced of this task. Participants saw (purportedly) neutral animal faces with or without tears. I would therefore expect any sense of "anger" in these faces to be rather subtle, and I iwonder if this apparent anger might have been due to slight biases in the selection of images. Were these images pre-tested? -> If the source animal faces were intended to be neutral, then this result would appear to suggest that they were not (entirely) neutral after all. Furthermore, the manipulation involved the addition of tears. Therefore, how does the absence of this manipulation indicate something about the effect of tears in this case? Here, I would be more convinced if there was a set of pre-tested neutral images and, e.g., a Likert-type or continuous scale for each of the discrete emotion. Then, if the addition of tears leads to a reduction of perceived "anger", this would seem to lend further support to Hasson's hypothesis.

Strengths and limitations:

I understood that this is the first study that systematically manipulated tears on animal images, and I would see this as one of the main contributions of this work. However, as in the introduction, I think it would help to elaborate a bit more about how tears on neutral animal faces might differ from examining tears on human faces.

As concerns the limitations, I would see the limited number of exemplars as one of the crucial points. I agree that an excessive number of stimuli might result in habituation effects. However, perhaps there could be two or three exemplars per animal, possibly with somewhat fewer different animals? It could also help to present some additional pre-test or post-test data showing that the chosen exemplars in this study were perceived as both typical and "neutral" exemplars of their species. This is perhaps even emphasized by the findings discussed about animal species and their colors. Here, it would seem that cats, and dogs, and likely also hamsters and horses could be found that have an entirely different fur color. Would, e.g., a black (or dark brown) cat have elicited a different pattern of responses?

Overall, I think that this paper makes a relevant contribution to the literature by showing that humans respond systematically to tears added to animal faces. The majority of the findings appears to show a rather consistent pattern to this effect, and this is strengthened even further by the fact that these animal faces were so heterogeneous overall. However, I would be more cautious when interpreting apparent differences between the species, or the apparent gender effects. The explanations raised for these effects appear plausible, but the present state of the findings seems rather tentative in this regard (mostly because of the very limited number of exemplars).

Future Directions and conclusions

I fully agree with the authors here, that this line of work opens up a lot of interesting possibilities for further research. Likewise, the conclusions very nicely summarize the main points and merits of this work.

Reviewer #2: In this manuscript, the authors attempt to extend the theoretical and empirical literature on emotional tearing in humans to non-human animals (cat, dog, horse, chimpanzee, and hamster). Participant’s gave self-reported ratings of regarding perceptions of emotional intensity as well as individual variables that might affect these perceptions. Similar to the work on tearing in humans, it was found that the presence of tears affected perceptions of in non-humans. Here are my comments/concerns:

1. Lines 109-111 “… however, by blurring vision, tears are a handicap in the case of having to attack or defend against an attacker, as they blur vision.” I think it’s only necessary to mention the blurring of vision once here.

2. The authors mention that emotional tears are uniquely human. I’d like more of a discussion on why there would be expected to be a “tear effect” in non-humans in light of this fact. Of course, the data suggest that there is a reason, but I’d like to see more of a discussion on this.

3. The authors begin the Participants section with a number. I believe this should be written out if they are writing in APA style.

4. Do we have any reason to believe that the digitally added tears were seen as authentic? If not, this could be seen as a limitation.

5. The “F” should be italicized when presenting the F statistic.

6. On line 283, the authors state that “the dog and chimpanzee benefited from the presence of tears…” I think the term “benefited” should be replaced with a less judgmental term. A term that doesn’t imply anything positive or negative. This happens again on line 307.

7. The “r” should be italicized when presenting correlations.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

24 Feb 2021

JOURNAL REQUIREMENTS

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

Answer: The style requirements are fulfilled

2. Thank you for including your ethics statement: "The study was under the standards of the Human Research Ethics Committee of the University of Valencia (Spain) and a written informed consent was obtained from all the participants.".

Please amend your current ethics statement to confirm that your named institutional review board or ethics committee specifically approved this study.

Answer: The statement is amended

Once you have amended this/these statement(s) in the Methods section of the manuscript, please add the same text to the “Ethics Statement” field of the submission form (via “Edit Submission”).

Answer: This is done too

3. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

Answer: Please consult the cover letter

4. Please include a copy of Table 1 which you refer to in your text on page 12.

Answer: This is done too.

RESPONSES TO REVIEWERS

REVIEWER 1.

We will respond to your comments point by point. First of all, we would like to thank you for your review. We now believe that our work is better because of your review.

Reviewer #1: I have read this contribution on adding tears to animal faces with great interest! Furthermore, I am convinced that it has merit in demonstrating how tears on animal faces are subject to anthropomorphism mechanisms, and that animal tears may influence perceived aggressiveness and friendliness of animals (even if these relationships are not strong). The paper is generally well written, although I spotted several smaller language issues along the way - i.e., the manuscript could still benefit from some thorough proof-reading.

Answer: Thanks very much for your interest in our manuscript and for your encouraging words. We have made deep changes through the whole of the manuscript, taking into account the most of your clever suggestions. Moreover, the new manuscript is now better supported by a new study and its analyses, designed to answer several points from your review. We hope to have clarified also our theoretical concepts, through some new references and background. Please, note that there are many changes in the manuscript (highlighted text). We took note of your suggestion regarding the language, and an expert in scientific English has rechecked our work in order to make it more accurate.

Reviewer #1: However, while I think there is reason to be excited about this work, I also saw some issues that limit the extent to which the findings can be interpreted. Perhaps most importantly, there was only a single exemplar for each species, and there was no pre-test data reported for the stimulus selection. E.g., it was unclear if all the baseline animal faces were indeed neutral in expression. Likewise, we do not know how typical these examples were. Would the same results be obtained for slightly different expressions, or different fur colors (affecting tear-visibility) among the same species?

Answer: The stimuli we used have been submitted to an additional new study to answer the questions you raised. Please check supplementary materials and new arguments in the main text of the manuscript (including the discussion). Anyway, a legitim doubt (a wondering indeed) is your last question (“Would the same results be obtained for slightly different expressions, or different fur colors (affecting tear-visibility) among the same species?”) which still remains and which has been commented in the limitations of the study.

Reviewer #1: Conceptually, I felt that the role of "emotional tears" in this context was not so clear. Were these emotional tears at all? The answer seems to be that this was in the eye of the beholder - however, this would need to be introduced and discussed more clearly.

Answer: We believe that this point, which connects clearly with the anthropomorphization conceptual line in the manuscript, has been introduced and discussed much better (please check the Introduction and Discussion sections).

Reviewer #1: Likewise, the reasoning about the difference between adding tears to neutral animal faces vs. neutral human faces (as in prior work) did not become clear to me. I believe there is a point here, but it would need to be better explained.

Answer: Again, we feel that this explanation is improved now.

Reviewer #1: Methodologically, I was struggling with the decision to obtain forced choice judgments for the discrete emotions. I find it particularly difficult to interpret any kinds of differences between the baseline images (i.e., the images without the tears that were not digitally manipulated) given this question format.

Answer: We understand that perhaps items in Likert scale format could have provided more information, but our choice of an item with discrete emotions was due to our previous work. As we wanted to extend the conclusions of that work to animal faces, it was better not to change the methodology. On the other hand, we do think that it would be necessary to investigate emotions from a dimensional approach, and this could be a great advance that we will consider in our next research.

Reviewer #1: In addition, I believe this should be regarded as a mixed design (with repeated measures across a series of exemplars), even if the exemplars themselves were between-subjects. A related concern here is that the Google-based image selection may have turned up examples that already differed with respect to expressions before the modification. I would therefore find it important to include a breakdown of the results per animal species and emotions (instead of or in addition to figure 3). Finally, relative tear-visibility, and tear-realism may have been a concern here, as the animals differed in both fur color, and the number of tears (one or two). This might be addressed by collecting some post test data for typicality (of the animal pictures), tear-visibility, and tear-realism.

Answer: We are happy to answer your questions about methodology below. Regarding the choice of the figure, since we had reported the odds ratios by species, we believed that it should provide new information that did not appear in the text in an explicit way. However, those plots can be found in the supplementary material along with the other information.

Reviewer #1: Despite these limitations, I overall found this paper to be very interesting and thought-provoking. It does lend support to the notion that humans might react to animal tears as if they were emotional, and that this might include the type of appeasement functions hypothesized by Hasson and others. I hope that the authors may find my more detailed comments below to be of aid in improving their manuscript.

Answer: We can assure you that your review has been very useful to us, and we believe that thanks to it our work is now better.

Introduction:

Reviewer #1: One of the key arguments for conducting this study is the argument that, since animals do not cry in response to emotional feelings, adding visible tears to neutral emotional expressions presents an excellent opportunity to study the effect of tears. I would tend to agree with this general reasoning. However, the claim that because of this, emotional reactions in observers "can easily be attributed solely to the presence of tears" (p.4) may need some refinement and qualification. It would appear to be a bit more complicated than that. How does the fact that humans are the only species that sheds emotional tears influence the responses of participants? Does this require any conscious reflection by participants? E.g., Media Equation theory would seem to suggest that observers should respond to crying animal stimuli in a similar manner *as if* the tears were produced by humans.

Answer: We are very grateful for pointing out clues towards a deepening of the scientific literature on anthropomorphization. We have revised and added several references to be considered for the Introduction (some of them regarding Media Equation Theory), so that our argument is now more precise and better supported. Please check the new Introduction section, especially pages 5-6.

Reviewer #1: Issues to be considered here are (1) Participants may be more likely to be aware that the tears were digitally added, even provided that the tears look very realistic. It may still be that participants did not question or otherwise consciously reflect upon the presence of the tears in these faces - but if they did, then their emotional responses may also have been affected by this realization. (2) What precisely is the key difference here between seeing tears added to a neutral human face, and seeing tears added to the face of an animal? Animal faces can still be perceived as sad, aggressive etc. (even without the addition of tears). To what extent, and in which way precisely, does the manipulation of the animal faces go beyond adding tears to neutral human faces? I believe the authors have an interesting point here - but I think this needs to be further elaborated and clarified.

Answer: We elaborated an argument in which we propose that the presence of the tears (provided they were perceived as realistic from the new study, in supplementary materials) significantly increased the probability for human observers to engage in an anthropomorphization process towards animals, attributing them an emotional expression they did not really have (please check the Discussion section). The extent of whether this process was conscious or unconscious in our observers has not been discussed, mainly because direct measures of anthropomorphism (i.g. questionnaires like ours) can reflect either the implicit or explicit anthropomorphism, or a mixture of both, depending of the person [Złotowski J, Sumioka H, Eyssel F, Nishio S, Bartneck C, Ishiguro H. Model of Dual Anthropomorphism: The Relationship Between the Media Equation Effect and Implicit Anthropomorphism. International Journal of Social Robotics. 2018;10(5):701-714.]. In addition, to deepen in such aspect it was a bit far from the main objectives of our study.

Reviewer #1: That said, I agree that it is very interesting to study tears on animals for which we have very different stereotypes (you said "constructs" here - what precisely is meant with that in this context?).

Answer: This phrase was in fact very confusing and it has been reformulated.

Reviewer #1: Also, the point about daily experience and closer interactions with animals helps to differentiate this from work in HRI/HCI.

When raising the argument about anthropomorphism, I would suggest referring to some of the extant literature on zoomorphism as well. This is not an entirely new field, and it would appear to be highly relevant here. As some of the current references additionally relate to human-robot interaction, this could also include zoomorphic robots, such as Spot (Boston Dynamics), or Miro (Consequential Robotics).

Answer: Please check the new references and the mention to such dog-shaped robots in the Discussion section, and we would like to thank you for bringing this to our attention.

Reviewer #1: Perhaps it would also help to discuss that the kinds of animals selected in this study are all capable of shedding tears - they just do not use this as an emotional signal to conspecifics. This signalling function appears to be what is included in the process of anthropomorphization.

Answer: This key point has been added to the Introduction section, together with the reference of Frey’s (1985) book «The mystery of tears»

Visual Stimuli

Reviewer #1: The example of the dog in figure 1, in my view, looks convincing. However, I have been wondering about some of the other animals. E.g., for the cat, the white fur may have presented a challenge for obtaining comparable visibility/salience of the tears. For the hamster, I wonder if the size of the tears relative to the eyes and the rest of the animal's face might have posed a challenge. I trust that all these materials looked well in the end. However, it would be great if they could be made available in full via supplementary materials.

Answer: As you can see in the new version, all materials are available as supplementary material. Following your suggestion of a post-hoc study on the stimuli used, as far as tear visibility is concerned, you can see from the results provided that it was extremely high in all species.

Measures

Reviewer #1: Why was the basic emotion item presented as a forced choice? This would appear to exclude the possibility of mixed emotions. Also, I think this makes it difficult to argue that there was "greater sadness" (see abstract). Instead, would this measure not only yield a frequency with which sadness appeared to be dominant?

Answer: After reviewing that part of the paper we can only say that we completely agree with you. Our conclusions regarding emotions should be expressed in terms of frequency. We have changed the text to reflect this. Regarding why we present basic emotions as a forced choice, this is because this paper is a continuation of our line of research in which in our previous papers we have included that item in exactly the same way.

Perhaps we should consider opting for a different methodology that allows more flexibility of response, and we intend to be guided by your suggestions for future studies.

Procedure

Reviewer #1: It seems that this was a mixed design, rather than a (pure) between-subjects design - with repeated measurements from each observer for intensity, friendliness, and aggressiveness. A GLMM may be a more suitable statistical approach for these analysis, as it could be used to control for effects of observer identity.

Answer: We have carefully read this comment. As supplementary material we have provided the results of the same study using a fixed and random effects analysis. As can be seen from these results, the variability among participants is minimal, and in essence, the conclusions drawn about the effect of tears on changing perceptions are identical. We opted at first for a simpler analysis because the methodology was the same as that of one of our previous published studies (where the data were analyzed using an ANOVA like the present one), and on the other hand, because we followed the principle of parsimony.

If we were to interpret intensity, friendliness, and aggressiveness as repeated measures, the scores obtained across species should be considered experimental pseudoreplicates (see Lawson, 2015). What we did in this other approach was to introduce species as another random factor along with participants. We could also have opted for a hierarchical model (according to our statistician, an appropriate model) in which the random effects would have been nested. It would not change the conclusions and we would lose ease of interpretability.

Reference

Lawson, J., 2015. Design and Analysis of Experiments with R. 1st ed. Chapman and Hall/CRC.

Results

Emotionality - Odds ratios

Reviewer #1: Figure 3: This figure seems to be missing a scale on the y-axis. The figure caption helps somewhat - but I was still wondering why the squares have different widths.

Answer: In the current version of the manuscript we have added an explanation of how this figure is constructed. Now our readers will be able to know how to obtain the height and width of the bars.

Reviewer #1: Is there an explanation for why the results for the cat went in the other direction? Given that there are different stereotypes about these different animals, as well as substantial differences in the basic morphology of their faces, it seems likely that even the tearless images may have been perceived as showing a certain amount of non-neutral expressions. Was there any evidence for this? Did expressions for the tearless animals vary between animals? If yes, then it might be useful to include a figure that shows the baseline emotions for each tearless animal before pooling all of this data.

Answer: As indicated above, these plots can be found as supplementary material.

Intensity

Reviewer #1: What does it mean that the dog and the chimpanzee "benefited" from the presence of the tears? Do you mean that they were anthropomophized more (more intense perceived emotions)? For friendliness, this is more straight forward.

The gender effects for the horse condition are interesting and seem to align with the notion that personal relevance of the animal should matter.

Figure 4: This figure is much too small and unreadable in the manuscript. The downloadable version was readable, however.

Answer: We agree with you and have changed the graph. We hope that this time the size is right and the colors allow you to see the differences in more detail. In addition, and as we have commented below, the word “benefited” has been removed since it was confusing.

Correlations between variables:

Reviewer #1: I unfortunately could not find table 1 anywhere in the manuscript, nor in manuscript central. However, the most relevant results appear to have been reported in the text.

Answer: In the previous version we made the mistake of not including table 1. This error has now been corrected and Table 1 can be found inserted in the manuscript.

SEM:

Reviewer #1: I found the effects of tears on aggressiveness examined in this analysis to be very interesting. However, it seems that the direct effect, albeit significant, was rather weak. Likewise, the link to perceived friendliness also did not appear to be very strong. Nevertheless, it would seem worth reporting. Perhaps this might be further substantiated in future work, as it could help to make the point that human observers might transfer even some of the purported appeasement functions of this signal to other species.

Unfortunately, I could not access the supplementary materials on OSF without logging in with my ID to request permission from the authors.

Answer: We have fixed that problem and supplementary material can now be accessed anonymously.

Discussion:

Reviewer #1: The discussion takes up the point again that emotional tears are a solely human feature that non-human animals are unable to produce. I found the argument about extending the work on appeasement signals to be very interesting, and perhaps more clearly formulated here than in the introduction.

Answer: We hope that such argument is now clearer in the Introduction and also the Discussion sections.

Reviewer #1: However, I found the distinction between emotional and non-emotional tears to be rather fuzzy here. I.e., on the one hand, the argument is made that non-human animals cannot produce emotional tears. On the other hand, shortly thereafter (p.13), it is argued that domestic animals gained benefits from "the emotional stimulus". Please be clearer on how these arguments go together to avoid confusion about when you are talking about emotional tears, and when you are talking about non-emotional tears. Note that the argumentation in the introduction and stimulus design was that tears were added to neutral faces of animals. Where does the emotional stimulus come from then? If the stimuli looked realistic (which I assume they all did), should these tears then not have been non-emotional tears? I think that this potential source of confusion could be addressed - i.e., it may not be necessary for the animals to shed emotional tears for human participants to perceive them as such. Nevertheless, I think would still need to be made more explicitly. I would see the finding that participants perceived these ostensibly non-emotional tears as noteworthy.

Answer: We noted that the word “benefited” created confusion to the main argument through the text, so we eliminated it. Please, check also that we incorporated the following phrase (from you) to the text in the Discussion section: “it may not be necessary for the animals to shed emotional tears for human participants to perceive them as such”. We think such statement can work as a key point for the final argumentation.

Reviewer #1: The discussion mentions that some of the tears consisted of one teardrop, and some consisted of two teardrops. Given that there was only one exemplar per species, could this distinction have confounded some of the findings - e.g., concerning apparent differences between species? You might refer to prior work in this context that has looked at tear-intensity to address this issue at least partially (i.e., the amount of tears may not matter much). Still, this appears to be a relevant limitation that would call for further research.

Answer: This was just a linguistic figuration to give emphasis to the text, but created confusion, so it has been eliminated from the text. Please check the supplementary materials and see that the tearful images barely differ in the digital manipulation performed (in all animals the tearing was added to both eyes).

Reviewer #1: On dogs and cats: While I found the discussion about differences in how we relate to cats versus dogs interesting, I felt concerned here that this might be over-interpreting things, given that there was only a single exemplar - a single image, for each. Again, I think it would be important to include in the results also the emotions that were perceived for the unmodified stimuli. Might this specific dog face have been perceived as sadder to begin with, and the cat face as emotional in some other way? Were they both regarded as completely neutral without the tears? I would suggest expanding a bit more on the basic literature about facial expressions in cats and dogs before raising the point about familiarity. E.g., this might start from the DogFACS and CatFacs (both https://www.animalfacs.com/).

Answer: Our argument about domestic animals has changed a little in order to be clearer, and, on the other hand, we have indicated that it has to be understood as tentative (thus, as a good line for future research). About the concern on the number of images used to illustrate species, it is commented in the limitations (“However, a possible limitation of the present work is the small number of images used; a greater number of images of each species could have avoided the potential influence of the particularities of each photograph and species.”), and, finally, we didn’t deepen in DogFACS and CatFacs because, though it could be an inspiration for future studies, it was also a bit further from our main objectives in this work.

Reviewer #1: SEM: On p.14, it is claimed that the absence of tears is associated with the perception of anger. Perhaps clarify that this discussion point is no longer about the SEM but about the forced-choice emotion rating task. Again, I was also not entirely convinced of this task. Participants saw (purportedly) neutral animal faces with or without tears. I would therefore expect any sense of "anger" in these faces to be rather subtle, and I iwonder if this apparent anger might have been due to slight biases in the selection of images. Were these images pre-tested? -> If the source animal faces were intended to be neutral, then this result would appear to suggest that they were not (entirely) neutral after all. Furthermore, the manipulation involved the addition of tears. Therefore, how does the absence of this manipulation indicate something about the effect of tears in this case? Here, I would be more convinced if there was a set of pre-tested neutral images and, e.g., a Likert-type or continuous scale for each of the discrete emotion. Then, if the addition of tears leads to a reduction of perceived "anger", this would seem to lend further support to Hasson's hypothesis.

Answer: All the paragraph this comment refers to has been changed and/or deleted in the new Discussion section. Please check also the supplementary study on the quality of the images used.

Strengths and limitations:

Reviewer #1: I understood that this is the first study that systematically manipulated tears on animal images, and I would see this as one of the main contributions of this work. However, as in the introduction, I think it would help to elaborate a bit more about how tears on neutral animal faces might differ from examining tears on human faces.

Answer: The new manuscript takes deepen this last aspect or the study.

Reviewer #1: As concerns the limitations, I would see the limited number of exemplars as one of the crucial points. I agree that an excessive number of stimuli might result in habituation effects. However, perhaps there could be two or three exemplars per animal, possibly with somewhat fewer different animals? It could also help to present some additional pre-test or post-test data showing that the chosen exemplars in this study were perceived as both typical and "neutral" exemplars of their species. This is perhaps even emphasized by the findings discussed about animal species and their colors. Here, it would seem that cats, and dogs, and likely also hamsters and horses could be found that have an entirely different fur color. Would, e.g., a black (or dark brown) cat have elicited a different pattern of responses?

Answer: This concern has been analysed through the new study on the quality of the images, which we invite you please to consult. However, note that, as we answered above, a legitimate question would still remain about intra-species diversity (in color, fur etc), and such limitation is noted in the manuscript. In this sense, adding more intra-species images would be interesting, but then why not more species? (and so on) … So, we think, taking into account that our main objective goes around the tear effect, the present materials are appropriate enough to reach our goals.

Reviewer #1: Overall, I think that this paper makes a relevant contribution to the literature by showing that humans respond systematically to tears added to animal faces. The majority of the findings appears to show a rather consistent pattern to this effect, and this is strengthened even further by the fact that these animal faces were so heterogeneous overall.

Answer: Thanks very much and note that, again, one of your statements has been added to the text (“is strengthened even further by the fact that these animal faces were so heterogeneous overall”).

Reviewer #1: However, I would be more cautious when interpreting apparent differences between the species, or the apparent gender effects. The explanations raised for these effects appear plausible, but the present state of the findings seems rather tentative in this regard (mostly because of the very limited number of exemplars).

Answer: We totally agree with you and both suggestions have been lowered in intensity and stated as tentative.

Future Directions and conclusions

Reviewer #1: I fully agree with the authors here, that this line of work opens up a lot of interesting possibilities for further research. Likewise, the conclusions very nicely summarize the main points and merits of this work.

Answer: Thanks very much for your kindly words and for your time.

REVIEWER 2

We would like to thank you for all your comments. We will answer to your suggestions point by point.

Reviewer #2: In this manuscript, the authors attempt to extend the theoretical and empirical literature on emotional tearing in humans to non-human animals (cat, dog, horse, chimpanzee, and hamster). Participant’s gave self-reported ratings of regarding perceptions of emotional intensity as well as individual variables that might affect these perceptions. Similar to the work on tearing in humans, it was found that the presence of tears affected perceptions of in non-humans. Here are my comments/concerns:

1. Lines 109-111 “… however, by blurring vision, tears are a handicap in the case of having to attack or defend against an attacker, as they blur vision.” I think it’s only necessary to mention the blurring of vision once here.

Answer: We fully agree with you. The sentence has been rewritten so that it is not redundant and the text is clearer.

Reviewer #2:

2. The authors mention that emotional tears are uniquely human. I’d like more of a discussion on why there would be expected to be a “tear effect” in non-humans in light of this fact. Of course, the data suggest that there is a reason, but I’d like to see more of a discussion on this.

Answer: This suggestion has been of great help to us. We have rewritten the discussion so that it is clearer and is an extension of our new introduction. We hope that the reader will understand our explanations in which we focus our hypotheses on the process of anthropomorphization. This topic, along with how visible tears favor emotional inference, has been discussed in greater depth.

Reviewer #2:

3. The authors begin the Participants section with a number. I believe this should be written out if they are writing in APA style.

Answer : We agree with you and have corrected that part of the text to conform to Vancouver standards.

Reviewer #2:

4. Do we have any reason to believe that the digitally added tears were seen as authentic? If not, this could be seen as a limitation.

Answer: After reading this commentary and some of the comments made by reviewer 1, we found it necessary to conduct a complementary study on the stimuli used. Those results can be found as supplementary material and demonstrate that in general tears were seen as realistic stimuli.

Reviewer #2:

5. The “F” should be italicized when presenting the F statistic.

Answer: We have corrected that part of our results and thank you for your comment.

Reviewer #2:

6. On line 283, the authors state that “the dog and chimpanzee benefited from the presence of tears…” I think the term “benefited” should be replaced with a less judgmental term. A term that doesn’t imply anything positive or negative. This happens again on line 307.

Answer: Thank you very much for this comment. We have realized that we had abused the use of this word. Upon further reading, we have rewritten part of the discussion and replaced the term "benefited" with more appropriate ones.

Reviewer #2:

7. The “r” should be italicized when presenting correlations.

Answer: We have corrected that part of our results and thank you for your comment.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

16 Mar 2021

Pécs, Hungary

March 15, 2021

PONE-D-20-35135R1

When Animals Cry: The Effect of Adding Tears to Animal Expressions on Human Judgment

PLOS ONE

Dear Dr. Gadea,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised by the Reviewer, listed below.

Please submit your revised manuscript by Apr 30 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Joseph Najbauer, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Overall, I would like to thank the authors for having addressed my previous points so thoroughly! Taking this into account, I find the present version of this manuscript to be substantially improved. In particular, I appreciated the inclusion of the additional post-test and analyses included in the supplementary materials that, I think, are very helpful with respect to better understanding the strengths and limitations of the present work. Likewise, I would like to thank the authors for having performed additional fixed and random effects analyses. The finding that these results converged with the previous ANOVA-results suggest that these were indeed robust. It also makes sense to me to stick with presenting the simpler ANOVA results in this case.

As illustrated by the post-test, a few limitations still remain: the limited number of exemplars; the lack of realism for a few of the stimuli (horse, cat); the variability in the extent to which some of the tearless stimuli were perceived as emotional (e.g., horse vs. dog). I therefore still find some parts of the discussion regarding differences between animal species (e.g., between cats and dogs) to perhaps be a bit premature at this point. Nevertheless, it is of course interesting to speculate about such differences, and this is being done more cautiously now (and in view of the limitations of this initial study). The discussion has furthermore been substantially improved overall, and the main findings are presented more clearly now.

Results of the new rating study:

I believe that the new post-test has helped to sufficiently strengthen this already very interesting paper. Since these materials are openly available, it should be easy to replicate/extend the current work. Tears were overall judged as very well visible in the new rating study. This, together with being able to view all the pictures in full helps to address a lot of the concerns about the technical soundness of this work. However, as shown by the study of the stimuli, some limitations do remain. First, perceived emotionality of the tearless images appears to have been quite variable. In particular, the tearless horse picture seems to have been perceived as substantially less emotional than the dog picture (which was already perceived as very emotional even without tears). This suggests that some of the effects of tears on specific animal images may indeed be more difficult, and perhaps more interesting, to interpret (see discussion). Second, it seems that the tears on the Horse and Cat may not have been perceived as sufficiently realistic. I.e., here, it would seem that the answer to this question from the response letter: "provided they were perceived as realistic from the new study" - may have to be that this is not fully supported by the post-test. Nevertheless, neither of these limitations would seem to invalidate/threaten the conclusion that the addition of tears generally increased perceived emotionality.

I only have a few minor additional comments beyond the above-mentioned limitations of the post-test:

Introduction:

I was not sure if quite as much background on human eye gaze is needed here. The first paragraph might be shortened somewhat to move more quickly towards tears, or animal crying.

Regarding reference [9], i.e., responses to tears within 50 milliseconds - here, recent work has so far failed to replicate this finding (Gračanin et al., 2021).

Reference:

Gračanin, A., Krahmer, E., Balsters, M., Küster, D., & Vingerhoets, A. J. (2021). How weeping influences the perception of facial expressions: The signal value of tears. Journal of Nonverbal Behavior, 45(1), 83-105.

Minor points:

p.5, l.134: there is an opening "(" here that is missing the closing ")".

p.5, l.154: replace "according their lacking of..." with "according to their lack of..."

p.6, l.174: replace "for such process" with "for such [a] process"

p.16, l.450: replace "confirmed our hypothesis" with "supported our hypothesis"

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

12 Apr 2021

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

RESPONSE: The reference list is revised. No retracted papers. A reference was changed after reading a comment from the reviewer (please check the response below)

Comments to the Author

Reviewer #1: Overall, I would like to thank the authors for having addressed my previous points so thoroughly! Taking this into account, I find the present version of this manuscript to be substantially improved. In particular, I appreciated the inclusion of the additional post-test and analyses included in the supplementary materials that, I think, are very helpful with respect to better understanding the strengths and limitations of the present work. Likewise, I would like to thank the authors for having performed additional fixed and random effects analyses. The finding that these results converged with the previous ANOVA-results suggest that these were indeed robust. It also makes sense to me to stick with presenting the simpler ANOVA results in this case.

As illustrated by the post-test, a few limitations still remain: the limited number of exemplars; the lack of realism for a few of the stimuli (horse, cat); the variability in the extent to which some of the tearless stimuli were perceived as emotional (e.g., horse vs. dog). I therefore still find some parts of the discussion regarding differences between animal species (e.g., between cats and dogs) to perhaps be a bit premature at this point. Nevertheless, it is of course interesting to speculate about such differences, and this is being done more cautiously now (and in view of the limitations of this initial study). The discussion has furthermore been substantially improved overall, and the main findings are presented more clearly now.

RESPONSE: Regarding the dog, we are in agreement with you in that it was seen as emotional even in the original tearless image, which we described as «unique » in the discussion, given the amount of emotionality rated on it. At this regard, we think it could be because a variety of reasons, including our relationship with dogs in general (as especulated in the Discussion) or also due to the chosen breed for the photography, or even to the expression of this concrete dog/photo. In any case, this illustrates the need for a careful selection of the images, and also the need of a quality study like the one we performed, following your suggestion. Please check a couple of new phrases included in the dog&cat argument as advices for caution with interpretations, and also a new phrase in the « Strenghts and limitations » section, in line with your comments.

Results of the new rating study:

I believe that the new post-test has helped to sufficiently strengthen this already very interesting paper. Since these materials are openly available, it should be easy to replicate/extend the current work. Tears were overall judged as very well visible in the new rating study. This, together with being able to view all the pictures in full helps to address a lot of the concerns about the technical soundness of this work. However, as shown by the study of the stimuli, some limitations do remain. First, perceived emotionality of the tearless images appears to have been quite variable. In particular, the tearless horse picture seems to have been perceived as substantially less emotional than the dog picture (which was already perceived as very emotional even without tears). This suggests that some of the effects of tears on specific animal images may indeed be more difficult, and perhaps more interesting, to interpret (see discussion). Second, it seems that the tears on the Horse and Cat may not have been perceived as sufficiently realistic. I.e., here, it would seem that the answer to this question from the response letter: "provided they were perceived as realistic from the new study" - may have to be that this is not fully supported by the post-test. Nevertheless, neither of these limitations would seem to invalidate/threaten the conclusion that the addition of tears generally increased perceived emotionality.

RESPONSE: The phrase you cite -copied from the rebuttal letter- is certainly not fully supported by the post-test (surely due to an excess of abstract for the letter), and in fact it was not included in the text of the manuscript. Instead, in the anterior manuscript it can be read: “Finally, we noted a few unexpected results (the “colder tearful cat” and the absence of a tear effect on aggressiveness in the horse) that were obtained with the pictures that were rated as less realistic in the quality study. This prompts an interesting reflection on the use of artificial stimuli to depict human emotions in non-human targets (animals and possibly robots), in the sense that, the more artificial and unreal they are perceived to be, the more difficult it is for them to produce empathy in us, and thus an anthropomorphic feeling” We think that this paragraph is expressing the limitations you note in a constructive manner.

I only have a few minor additional comments beyond the above-mentioned limitations of the post-test:

Introduction:

I was not sure if quite as much background on human eye gaze is needed here. The first paragraph might be shortened somewhat to move more quickly towards tears, or animal crying.

RESPONSE: The first paragraph has been adequately shortened.

Regarding reference [9], i.e., responses to tears within 50 milliseconds - here, recent work has so far failed to replicate this finding (Gračanin et al., 2021).

Reference:

Gračanin, A., Krahmer, E., Balsters, M., Küster, D., & Vingerhoets, A. J. (2021). How weeping influences the perception of facial expressions: The signal value of tears. Journal of Nonverbal Behavior, 45(1), 83-105.

RESPONSE: We thank you for this reference, which we have read with great interest. After carefully reading, we have decided to change a bit our sentence and to eliminate the commented reference (number 9: Balsters et al., 2013). We noted that both the Balsters paper and the attempt to replicate its findings by Gračanin et al. assess reaction times and not the influence of a preattentional stimulus. Given that our aim was not the measurement of exact reaction times, and that this paragraph of the Introduction was written just to generally point that we (and our brain) in fact react to emotional tears, we now cite a reference we found in the literature on bioelectrical changes caused by the sight of tears: Krivan S, Caltabiano N, Cottrell D, Thomas N. I'll cry instead: Mu suppression responses to tearful facial expressions. Neuropsychologia. 2020;143:107490.

Minor points:

p.5, l.134: there is an opening "(" here that is missing the closing ")".

p.5, l.154: replace "according their lacking of..." with "according to their lack of..."

p.6, l.174: replace "for such process" with "for such [a] process"

p.16, l.450: replace "confirmed our hypothesis" with "supported our hypothesis"

RESPONSE: All these minor points have been observed and corrected.

Submitted filename: Response to reviewers (2nd).docx

Click here for additional data file.

20 Apr 2021

Pécs, Hungary

April 19, 2021

When Animals Cry: The Effect of Adding Tears to Animal Expressions on Human Judgment

PONE-D-20-35135R2

Dear Dr. Gadea,

We’re pleased to inform you that your manuscript (R2 version) has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Joseph Najbauer, Ph.D.

Academic Editor

PLOS ONE

-------------------------------------------------

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Dennis Küster

26 Apr 2021

PONE-D-20-35135R2

When animals cry: The effect of adding tears to animal expressions on human judgment

Dear Dr. Gadea:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Joseph Najbauer

Academic Editor

PLOS ONE