Why We Mimic Emotions Even When No One is Watching: Limited Visual Contact and Emotional Mimicry

Abstract

This article explores interpersonal functions of emotional mimicry under the absence versus the presence of visual contact between the interacting partners. We review relevant literature and stress that previous studies on the role of emotional mimicry were focused on imitative responses to facial displays. We also show that the rules explaining why people mimic facial expressions may be inapplicable when visual signals are unavailable (e.g., people attending an online meeting have their cameras off). Overall, our review suggests that emotional mimicry functionally adapts to whether the perceiver and the expresser can see each other. We, therefore, argue that blocking visual contact between them may provide insight into emotional mimicry's social functions, thereby clarifying its role in fostering affiliation and emotional understanding.

The growing popularity of videoconferencing has encouraged intense debate on the pros and cons of online meetings. One often-mentioned disadvantage of virtual interactions is that they recreate real-life communication to a limited extent (Bailenson, 2021). Nonverbal cues that are present in face-to-face meetings are often missing, especially when participants on both sides have their cameras off, and messages sent by facial expressions are not visible. This limited access to visual information may affect interpersonal communication because perception and response to others’ facial displays play a crucial role in social interactions (Oh Kruzic et al., 2020). One of the most intriguing phenomena observed during face-to-face interactions is that people show an automatic tendency to mirror each other's facial displays, which is manifested by congruent facial activity and referred to as facial or emotional mimicry (Hess & Fischer, 2013; Hess, 2021).

Much evidence shows that emotional mimicry serves two interconnected social functions by facilitating mutual understanding and affiliation: Mimicking helps the perceiver accurately decode emotional displays of the expresser and increases liking and rapport between them (for reviews, see Hess, 2021; Kämpf et al., 2018; Niedenthal et al., 2017). Yet, the majority of previous studies in this area have been focused on emotional mimicry following exposure to facial displays, while mimicry in response to other (e.g., vocal) emotional expressions has rarely been studied. The question thus arises whether mimicry plays similar functions when the perceiver cannot see the expresser. For instance, will facial mimicry be useless when the expresser has their camera off and sends only vocal emotional cues? Or maybe, in this case, facial mimicking of vocally expressed emotions will compensate for the lack of visual information and serve as a tool to understand the expresser's emotional state better? It is also unclear what function mimicry plays when the expresser cannot see the perceiver. For instance, will a person attending an online meeting without a camera on be motivated to mimic smiles for affiliative purposes, knowing that the smiling participant cannot see them?

In the present article, we take a closer look at emotional mimicry and its two often discussed social functions (i.e., fostering affiliation and facilitating emotional understanding) under the absence versus presence of visual cues. For this aim, we provide a brief overview of recent research on emotional mimicry, with a focus on a few studies on facial mimicry in response to vocal emotional expressions and on facial mimicry in blind individuals. We also review studies on the effects of direct vs averted gaze and the awareness of being watched on facial responses to emotional expression. Based on the reviewed evidence, we put forward the notion that emotional mimicry occurs regardless of whether the expresser and the perceiver can see each other, but the relevance of the two functions (that is, fostering affiliation or emotional understanding) changes depending on the availability of visual cues. Put another way, we argue that emotional mimicry serves both functions, yet they are not activated rigidly but depend on visual contact. Finally, we present arguments why studying emotional mimicry under the absence versus presence of visual information may enhance our understanding of what emotional mimicry is and what functions it serves.

Emotional Mimicry as a Social Phenomenon

The term “mimicry,” in its broad sense, refers to the imitation of a variety of others’ nonverbal behaviors, including facial expressions (Hess & Fischer, 2013), gestures (Chartrand & Bargh, 1999), bodily postures (Bavelas et al., 1986; Bernieri & Rosenthal, 1991), speech prosody and nonverbal vocalizations (Neumann & Strack, 2000), or autonomous signals such as heart rate (Feldman et al., 2011) or pupil diameter (Kret et al., 2015). Imitating others’ expressed emotions stands out of this category because facial, vocal, and postural displays of emotion, in contrast to nonemotional nonverbal behaviors (e.g., head scratching, food tapping), are meaningful social signals that inform about the expresser's inner states, dispositions, intentions, and orientation towards the perceiver (Parkinson, 2005; van Kleef, 2017). Accordingly, the tendency to imitate the expresser's emotional display is not simply a mirroring of this display but rather a process that depends on how the perceiver interprets this display in a given social context (Hess et al., 2014). For instance, smiles are generally perceived as intrinsically affiliative social signals and hence, are mimicked more readily than frowns which are considered intrinsically nonaffiliative social signals. Yet, when a smile is shown on the face of a rival or dislikable person, its social meaning may change to non-affiliative, and the tendency to mimic it may be reduced (Hess & Fischer, 2013, 2014). This selectiveness of emotional mimicry suggests that the main goal of imitating others’ emotional displays is to regulate social interactions (Hess & Fischer, 2013; Hess, 2021).

The unique nature of emotional mimicry is highlighted by recent theories. Most of them hold that emotional mimicry is not a simple reflex-like process based on the perception-action link but a flexible, top-down regulated mechanism that helps navigate social interactions by serving two interrelated social functions. Specifically, emotional mimicry (1) fosters affiliation by signalizing emotional understanding and (2) facilitates emotional understanding by providing insightful cues about the expresser's emotional state and the meaning of their expressed emotion (Hess & Fischer, 2013; Hess et al., 2014; Hess, 2021; Niedenthal et al., 2005; Winkielman et al., 2018; Wróbel & Imbir, 2019). Interestingly, some evidence for the two functions of emotional mimicry is provided by studies in animals, thereby suggesting that the role of mimicry in emotional understanding and social bonding is not confined to humans and may have evolutionary roots (for a review, see Palagi et al., 2020).

Of importance here, however, is that most research focusing on these two functions in humans was limited to imitative responses triggered by exposure to facial displays, while studies on mimicry in response to nonfacial displays have been scarce. Yet, in everyday life, the evaluation of an emotional expression is rarely based on facial cues only. More often, we see a face accompanied by a verbal message, body posture, or vocal expression. It is also not uncommon to hear others talk without seeing their faces. This happens not only when attending an online meeting when the other participant has their camera switched off but also when interacting in complete darkness, talking on the phone, or driving and chatting with a passenger while keeping one's eyes on the road. Although, as we discuss below, the absence of visual information on the perceiver or expresser's side does not eliminate emotional mimicry, the question remains whether the process functionally adapts to situations in which visual cues are not available.

The Role of Visual Cues in Emotional Communication

Two-Way Exchange of Emotional Information via Visual Channel

The visual sensory channel plays a vital role in the perception of the expresser's emotions because it enables the perceiver to observe the expresser's face, posture, and gestures, all of which are powerful sources of emotional information (Hugenberg & Wilson, 2013; Lopez et al., 2017). Out of these sources, facial cues play a dominant role in emotion perception because, in face-to-face interactions, people primarily fixate on the expresser's eye and mouth regions, while bodily emotional expressions are mostly perceived through peripheral vision (Gullberg & Holmqvist, 1999; Hessel, 2020). Studies confirm that the eyes and mouth regions are not only the most expressive parts of the face but also the most diagnostic areas for the identification of distinct emotions (Kim et al., 2022; Smith et al., 2004). Accordingly, facial occlusion with sanitary masks or sunglasses impairs emotion recognition (Kim et al., 2022; McCrackin et al., 2022), which indicates that the visibility of the expresser's face is a critical factor in the perception of their expressed emotion.

At the same time, the interpretation of the expressed emotion does not depend solely on the expresser's emotional display because, as already mentioned, contextual factors may modify the social meaning of the expressed emotion (Barrett et al., 2011; Hess & Hareli, 2016). A wealth of such contextual cues is transmitted via the visual channel. This includes not only a broad visual context, such as the scenery the expresser is in or the clothes they are wearing (Hess & Hareli, 2016), but also the expresser's facial appearance, which provides information about the expresser's personality traits, age, gender, ethnicity, and other characteristics (Hareli & Hess, 2012; Hess et al., 2009). For instance, morphological facial features signal whether the expresser can be trusted, approached, or liked (Todorov et al., 2015), which in turn, can modify the interpretation of the expresser's emotion by making it more or less indicative of affiliative intents.

A powerful visual cue that shapes the social signal value of the expressed emotion is also the expresser's eye gaze. Most importantly, eye gaze informs the perceiver who the target of the expressed emotion is. In consequence, the interpretation of the expresser's emotional display changes depending on whether the expresser's eyes are directed at the perceiver or another target (Bourgeois & Hess, 2008). Eye gaze direction also informs the perceiver whether they are in the expresser's locus of attention, which explains why direct gaze signals approach and social inclusion while averted gaze signals avoidance and social exclusion (Leng et al., 2018; Mason et al., 2005). Accordingly, when combined with emotional expressions of distinct emotions, direct gaze facilitates the processing of approach emotions such as happiness or anger, while averted gaze facilitates the processing of avoidance emotions such as fear or sadness (Adams & Kleck, 2003).

The role of eye gaze, however, goes far beyond transmitting information about the expresser's emotion and its social meaning. The main reason for that is that, in face-to-face interactions, both the expresser and the perceiver use their eye gaze to mutually observe each other (Hessel, 2020; Schilbach, 2015). This dual role of eye gaze not only enables the perceiver to see the expresser (and hence the expresser's emotion in a given social context) but also makes them both aware that they are being watched by each other (Cañigueral & Hamilton, 2019; Hietanen et al., 2019). Accordingly, both the expresser and the perceiver use their interaction partner's eye gaze as instant confirmation of whether the emotional signals they are sending are being received, which is a basis for smooth mutual communication and affiliation (Hessel, 2020; Hietanen et al., 2018). Studies confirm that the experience of being watched is an important factor in eliciting affiliation-related physiological and facial responses in an interacting partner (Hietanen et al., 2018; Kiilavuori et al., 2021, 2022). For instance, when people are aware they are being watched, they are more likely to respond by sending affiliative signals such as smiles (Hietanen et al., 2018; Kiilavuori et al., 2022).

Overall, visual contact between the interacting partners plays an indispensable role in emotional communication. As a result, the perception of the expresser's emotion and the perceiver's responses to this emotion depend heavily on the availability of visual cues. Among those cues, the visibility of the expresser's face is the key factor because it conveys information about facial muscles activity, facial appearance, and eye gaze direction, all of which considerably affect the perception of the expresser's emotional state and the social meaning of their expressed emotion. Further, and even more importantly, eye gaze conveys direct, synchronous, and valid information about: (1) The perceiver's attention to the emotional signals the expresser is sending and (2) the expresser's attention to the perceiver's responses to these signals (including the mimicking of the expresser's emotional display) ¹ . This two-way instantaneous exchange of social and emotional information is possible only when they can see each other.

Visual Versus Non-Visual Cues

The dominance of visual cues in emotional communication does not mean that the lack thereof blocks emotional communication. When such cues are unavailable, both the expresser and the perceiver may glean emotional information from non-visual cues (Collignon et al., 2008). These include verbal messages as well as various contextual cues, such as previous knowledge about the interacting partner or pertinent social norms (Hareli & Hess, 2012). Important sources of emotional information are also non-verbal signals such as voice and touch, which, similar to visual cues, communicate distinct emotions (Hertenstein et al., 2009; Laukka et al., 2016; Schirmer & Adolphs, 2017) as well as transmit contextual information (e.g., voice characteristics may be indicative of the expresser's gender, age, or personality; Pernet & Belin, 2012; Ptacek & Sander, 1966; Stern et al., 2021). For instance, the information about the affiliative meaning of the expressed emotion may be gleaned not only from a facial expression (e.g., a smile) or the expresser's facial features indicative of trustworthiness and likability (Todorov et al., 2015) but also from a vocal expression (e.g., a smiling voice; Torre et al., 2022) or voice features indicative of trustworthiness and likabilty (O’Connor & Pat Barclay, 2017; Weiss & Burkhardt, 2010). Research supports the important role of non-visual cues in the perception of emotions—for instance, auditory emotional signals may complement visual ones, thereby facilitating emotional recognition (Paulmann & Pell, 2011). Moreover, information gathered through the auditory channel can alter emotion processing in the visual channel (de Gelder & Vroomen, 2000).

Yet, despite some overlap between visual and non-visual cues, there are two important differences between them that seem particularly relevant here. First, categorizing emotional expressions based on visual cues is easier than when only non-visual (e.g., auditory) cues are available (Collignon et al., 2008; Paulmann & Pell, 2011). This demonstrates visual dominance in emotional processing and explains why people provided with incongruent combinations of facial and vocal expressions (e.g., facial displays of disgust combined with a fearful voice) preferentially recognize emotional expressions based on visual modality (Collignon et al., 2008; Paulmann & Pell, 2011). Second, the visual channel is much better at transmitting information about social attention because, as already mentioned, the dual function of eye gaze makes both the expresser and the perceiver instantly aware that their partner pays attention to the signals they are sending (Cañigueral & Hamilton, 2019; Gobel et al., 2015). When visual contact is limited, similar information is not directly available and can only be accessed indirectly and asynchronously via the interaction partner's verbal and nonverbal vocalizations (e.g., verbal nods). However, such indirect and asynchronous information is much less reliable and depends entirely on whether the interaction partner is willing to share it. Hence, in most situations in which visual contact is limited on both sides, not only the perceiver's access to the expresser's emotion is incomplete, but also the perceiver's knowledge of whether their response to this emotion is received by the expresser is largely restricted. We believe that these crucial advantages of visual cues over non-visual ones may have important consequences for emotional mimicry's social functions, and we elaborate on those consequences below.

Is Visual Contact Necessary for Emotional Mimicry to Occur?

The lack of visual emotional cues on the expresser's side does not eliminate emotional mimicry because, as we already mentioned, when such cues are unavailable, emotional information may be gleaned from nonvisual cues. For instance, a perceiver may rely on their knowledge about the expresser's emotion. A good example is the study reported by Hess et al. (2014), in which participants were presented with neutral facial expressions combined with verbal information on the expresser's emotion. The results showed that participants who believed the expresser was happy or sad reacted with facial activity congruent with the information they received. Put another way, nonvisual cues (verbal information) turned out to be more important than visual input (neutral face), which—as Hess et al. (2014) illustratively put it—indicates that participants mimicked what they knew rather than what they saw (see also Davis et al., 2022).

When the expresser is not visible, facial mimicry may also be replaced by or complemented with vocal mimicry, that is, the imitation of vocal emotional expressions. For instance, people who listen to neutral sentences spoken in an emotional manner spontaneously imitate the expresser's emotional tone of voice (Hatfield et al. 1995; Neuman & Strack, 2000; Rueff-Lopes et al., 2015). However, vocal mimicry is not always possible because, in some contexts, the perceiver only passively listens to the expresser talk, which precludes the imitation of speech prosody and nonverbal vocalizations. Such contexts leave room for cross-channel mimicry, that is, matching facial responses to non-facial emotional expressions (Hawk & Fischer, 2016; Sachisthal et al., 2017).

A few studies have already demonstrated that congruent facial responses can be elicited by vocal expressions of emotion (Hawk et al., 2012; Hietanen et al., 1998; Lima et al., 2021; Magnée et al., 2007). Specifically, Hietanen and colleagues (1998) showed that a vocal expression of anger evoked more intense responses of corrugator supercilii (i.e., the “frowning muscle”) and less intense responses of orbicularis oculi (i.e., the “crow's feet muscle” typically associated with authentic smiles) than a vocal expression of contentment. In a related vein, Magnee et al. (2007) found that congruent combinations of facial and vocal expressions of fear increased corrugator activity, while congruent combinations of facial and vocal expressions of happiness increased zygomaticus activity (the “smiling muscle”). Importantly, no increase in zygomaticus and corrugator activity was observed when participants were exposed to incongruent combinations of facial and vocal expressions (that is, a fearful face was combined with a happy voice or a happy face was combined with a fearful voice), which suggests that both channels were important for triggering mimicry response. Arias et al. (2018) used digitally transformed sentences that sounded as if they were spoken with a smile on the speaker's face and found that participants who listened to these sentences responded to them with increased zygomaticus activity. Hawk et al. (2012) demonstrated that hearing and repeating vocalizations of anger, disgust, sadness, and happiness evoked corresponding facial expressions and self-reported emotions in the perceiver. Further and even more importantly, the inhibition of participants’ facial activity impaired their ability to process emotional vocalizations and reduced participants’ subjective experience of corresponding emotional states. Finally, Lima et al. (2021) examined facial electromyographic responses to nonverbal vocalizations and found that authentic laughs, as compared to posed laughs, evoked stronger activity of the muscles associated with positive affect (zygomaticus and orbicularis). Additionally, more intense orbicularis activity predicted higher perceived laughter authenticity, whereas more intense corrugator activity predicted lower perceived laughter authenticity, which suggests that facial responses were related to how participants interpreted the vocal expressions they were exposed to.

Further and even more compelling evidence for cross-channel mimicry comes from research on facial mimicry in the blind (Arias et al., 2021; Woloszyn et al., 2021). For instance, Arias et al. (2021) demonstrated that congenitally blind participants imitated smiles heard in digitally transformed spoken sentences with their facial muscles (zygomaticus major and corrugator supercilii), despite having never seen a facial expression of happiness. Importantly, their responses did not differ from the responses of sighted participants. In a similar vein, Woloszyn et al. (2021) showed that social-emotional sounds such as laughter evoked congruent facial responses (as indicated by the zygomaticus and corrugator muscles activity) to a similar extent in both sighted and blind participants.

Taken together, these findings show that vocally expressed emotions evoke congruent facial muscle activity in perceivers, thereby suggesting that, despite rather modest empirical interest in cross-channel mimicry, facial responses to vocal expressions of emotion are remarkably similar to mimicry of facial expressions. Moreover, studies on facial mimicry in blind individuals, similar to findings on the role of verbal cues in triggering mimicry response (Hess et al., 2014), indicate that the tendency to facially imitate expressions of emotion may be independent of the availability of visual cues or even previous visual experience. In addition, Hawk et al.'s (2012), as well as Lima et al.'s (2021) studies indicate that cross-channel mimicry may contribute to emotional understanding in a similar way unimodal mimicry does. This implies that both types of imitation—in response to facial and vocal emotional displays—should be equally important in regulating social interactions (see also Hawk & Fischer, 2016).

However, some indirect evidence challenges the notion that facial mimicry to vocal displays and facial mimicry to facial displays are functionally the same. More specifically, exposure to vocal displays of emotion eliminates the wealth of visual cues which, as noted above, significantly facilitate emotion recognition (Collignon et al., 2008; Paulmann & Pell, 2011). Therefore, when the perceiver cannot see the expresser, emotional mimicry will possibly compensate for the absence of visual cues typically gleaned from the expresser's face, thereby helping understand the expresser's internal state.

We should also add that the visibility of the perceiver's face, although rather neglected in previous studies, also seems important for emotional mimicry's functions because the perceiver may have much less motivation to signal affiliation by imitating the expresser's smile when they are not aware whether the expresser can see them. Hence, even though the absence of visual cues does not eliminate facial mimicry, it seems that the social functions of this process seem to be contextually regulated by the visual contact on the expresser and perceiver's sides. Below we review and discuss evidence supporting this claim.

Visual Contact and the Social Functions of Emotional Mimicry

Visual Contact and Emotional Mimicry as a Tool for Fostering Affiliation

The first function of emotional mimicry, that is, its role in fostering affiliation, is supported by much evidence showing that mimicry increases liking and rapport between interaction partners (for a review, see Hess, 2021). In general, this evidence is consistent with studies on behavioral mimicry, which show that imitating others’ gestures, postures, and other motor movements acts as a social glue that bonds individuals together (for a review, see Lakin et al., 2003). Importantly, the “social glue” function of emotional mimicry is often linked to the already mentioned selectiveness of the process, that is, its tendency to occur more (vs. less) readily when the context of the interaction is perceived as affiliative (vs. non-affiliative) (Hess & Fischer, 2013; Hess et al., 2014; Hess, 2021). It seems probable that the selectiveness of emotional mimicry is independent of whether the perceiver can see the expresser because, as noted above, the overall interpretation of the expresser's emotion as affiliative or not may be shaped by not only visual but also nonvisual cues. To the best of our knowledge, however, the top-down modulation of emotional mimicry by social cues has been demonstrated only for facial mimicry following exposure to facial displays (for reviews, see Hess & Fischer, 2013; Hess et al., 2014; Hess, 2021). We are unaware of similar evidence for the social modulation of vocal or cross-channel mimicry. The only study we know, which addressed this issue directly, tested the influence of the expresser's group membership on participants’ facial responses to facial or vocal displays of emotion but reported null effects of this social factor on both facial and cross-channel mimicry (Sachisthal et al., 2017). Hence, future research is needed to examine these effects more thoroughly.

The question also remains whether the perceiver's evaluation of the interaction context as affiliative is sufficient to motivate them to mimic the expresser's emotion for affiliative purposes. We think this is a valid question, given that the facilitating influence of mimicry on social bonding is often attributed to the fact that imitating others communicates attention and emphatic understanding (Hess et al., 2014; Lakin et al., 2003). For example, Hess et al. (2014) stress that mimicking others’ expressed emotions is not just imitation but a communicative signal by which the perceiver shows the expresser that they know how the expresser feels. This suggests that people should mimic others when they are motivated to send such affiliative communicative signals (Hess & Fischer, 2013; Hess, 2021). Of course, this motivation may be triggered by the perceiver's perception of the interaction context as affiliative, but it seems that considering emotional mimicry as a communicative act goes far beyond that perception and requires the perceiver to be additionally aware that the expresser can see their imitative response (and hence, pick up the affiliative message the perceiver is sending). This suggests that the awareness of being watched by the expresser is an important factor in activating (or at least facilitating) the affiliative function of emotional mimicry.

This reasoning is indirectly supported by studies showing that both behavioral and emotional mimicry is modulated by eye gaze direction. For instance, the expresser's direct (as opposed to averted) eye gaze enhances mimicry of their hand movements (Wang et al., 2011; Wang & Hamilton, 2014). A similar pattern of results has been found for emotional mimicry in response to happy faces, that is, the expresser's direct gaze was shown to evoke higher zygomaticus activity following exposure to happy expressions than averted gaze (Kuang et al., 2021; Marschner et al., 2015; Rychlowska et al., 2012; Schrammel et al., 2009; Soussignan et al., 2013). Animals have also been reported to respond to the direct gaze of others. For instance, mimicry of playful facial expressions (which communicate positive emotions and are strongly linked to affiliation) is often preceded by direct eye-gazing (for a review, see Palagi et al., 2020). When it comes to emotional mimicry to less affiliative emotional displays (i.e., angry faces), the findings are less consistent, with some studies showing that direct gaze evokes higher corrugator activity than averted gaze (Schrammel et al., 2009; Soussignan et al., 2013), while other studies report no such effect (Kuang et al., 2021; Marschner et al., 2015).

The fact that eye gaze direction consistently modulates responses to faces expressing affiliative emotions suggests that eye contact may be particularly important when it comes to the affiliative function of emotional mimicry. This is not surprising given the already mentioned tendency to regard direct vs. averted gaze as a signal of social approach vs. avoidance, respectively (Leng et al., 2018; Mason et al., 2005). Put another way, the expresser's direct gaze, when combined with displays of affiliative emotions, boosts the perceived affiliativeness of the interaction context, and hence, the positive influence of such gaze on mimicry responses seems a logical consequence of mimicry's affiliative function. For instance, Mauersberger et al. (2022) showed that when the perceivers were able to see the expresser's direct eye gaze (relative to the condition when the expresser's eyes were covered with sunglasses) they were more likely to mimic affiliative emotions (e.g., happiness), but for antagonistic emotions (e.g., disgust), the opposite was true (that is, mimicry was stronger when eyes were invisible).

The question remains whether mimicking others to build liking and rapport is fostered by the affiliative meaning of direct eye contact, the perceiver's awareness of being watched by the expresser, or the joint influence of these two mechanisms. An indirect answer to this question may be found in studies showing that the awareness of being watched is a strong moderator of perceivers’ responses to faces with direct vs averted gaze (Jarick & Bencic, 2019; Hietanen et al., 2020; Myllyneva & Hietanen, 2015). For instance, Myllyneva and Hietanen (2015) showed that greater skin conductance, heart rate, and cortical P3 responses to direct vs averted gaze were observed only when participants were led to believe that the expresser was able to see them through an electronic shutter. When participants believed that this person was unable to see them, no effect of the expresser's gaze direction on participants’ reactions was found. In a similar vein, Jarick and Bencic (2019) demonstrated that autonomic arousal, as indicated by skin conductance, was enhanced when interaction partners made direct eye contact, that is, turned their heads towards each other (in contrast to turning their heads away). This, however, was evident only when the contact was possible; once gaze signals became obscured or blocked (i.e., participants wore sunglasses or were blindfolded, respectively), arousal significantly decreased and was no longer modulated by sending or receiving direct gaze. Taken together, these studies suggest that direct eye gaze may play its functions not only due to its affiliative meaning but also due to the experience of being watched by the expresser.

However, the only study we are aware of that directly tested the influence of the awareness of being watched on facial mimicry led to inconclusive findings. Specifically, Hietanen et al. (2019) measured participants’ zygomatic and corrugator responses to a live model's smiling and neutral faces and manipulated participants’ awareness of being watched by using an electronic shutter. The results revealed that the difference in zygomatic activity following exposure to smiling versus neutral faces was more pronounced when participants believed they were being watched by the expresser (as compared to when they believed they were not being watched). This pattern, however, was mainly influenced by participants’ responses to neutral rather than smiling faces. Specifically, zygomaticus response to a neutral face was inhibited when participants believed that they were being watched, whereas zygomaticus response to a smiling face was not modulated by this belief. These findings, according to the authors, may mean that smiling in response to another person's smile is a strong automatic reaction that is relatively independent of whether the perceiver believes that they are being watched by the expresser.

We should also note that in most studies on emotional mimicry, the expresser is not actually present but shown in a picture or a video. Thus, even when the expresser's eyes are visible to the perceiver, the experience of being watched is either nonexistent or imaginary. Thus, the question remains whether the physical presence of the expresser is important to make the perceiver aware that their imitative responses are received. An indirect answer to this question has been given by Hietanen et al. (2020), who measured psychophysiological responses to another person's direct versus averted gaze, using a mere video recording, live interaction, and bidirectional video call. The authors showed that direct gaze, relative to averted gaze, increased perceivers’ autonomic arousal in the live and video call conditions, but not in the pre-recorded video condition. However, smiling (as manifested by increased zygomaticus activity) in response to another person's direct gaze was observed in all three conditions, thereby suggesting that the physical presence of the other person was crucial for the autonomic reactions but not for the facial responses elicited by direct gaze.

Overall, current evidence suggests that the perceiver's visibility to the expresser might be important for triggering the affiliative function of emotional mimicry. This function should be activated when the context is affiliative, and the expresser's eyes are directed at the perceiver (regardless of whether the expresser is physically present or not). However, the question remains whether the facilitating effect of direct eye gaze on mimicking others results from the fact that such gaze, when accompanying affiliative expressions, provides additional support that the context of the interaction is affiliative (as other contextual factors do) or from the fact that it makes the perceiver aware that they are being watched by the expresser. Future research is needed to address this question and examine the social role of emotional mimicry when the perceiver is not visible to the expresser.

Visual Contact and Emotional Mimicry as a Tool for Facilitating Emotional Understanding

The idea that emotional mimicry plays a role in understanding the emotions of others is based on the assumption that mimicking others’ expressed emotions provides bodily feedback, which helps the perceiver to re-experience these emotions in themselves and, as a result, decode these emotions more accurately (Arnold & Winkielman, 2020). The rationale underlying this assumption can be found in theories of emotional embodiment, according to which people process emotional information by reactivating neural states associated with their own emotional experiences (Niedenthal, 2007; Winkielman et al., 2018). Therefore, when imitating others’ emotional expressions, perceivers automatically reactivate corresponding states in themselves, which provides a basis for emotional understanding (Niedenthal, 2007; Niedenthal et al., 2017).

However, research addressing the effects of imitating others’ expressed emotions on the ability to recognize these emotions has yielded an inconsistent pattern of findings, with some studies showing that blocking facial mimicry impairs the recognition of emotions (Oberman et al., 2007; Neal & Chartrand, 2011; Niedenthal et al., 2001), while others report no relationship between mimicry and emotion recognition accuracy (e.g., Blairy et al., 1999; Hess & Blairy, 2001). Importantly, research that employs facial blocking, often taken as evidence for the link between mimicry and emotion recognition, sparked an intense debate on the reliability of this method (Coles et al., 2019). Moreover, a recent meta-analysis of correlations between facial mimicry and facial emotion recognition provided no support for a significant relationship between the two phenomena (Holland et al., 2021). This suggests that emotional mimicry may not necessarily serve as a tool to infer emotions from other people's emotional displays.

We should note, however, that the number of studies that examined the role of emotional mimicry for emotion recognition accuracy is low, and most of these studies were focused on a narrow set of prototypical emotional expressions (Holland et al., 2021). Such prototypical emotional expressions (i.e., stereotypical configurations of facial movements indicative of so-called basic emotions, Ortony & Turner, 1990) differ from spontaneous, multimodal emotional expressions encountered in real life (Barrett et al., 2019; Israelashvili et al., 2019). Most importantly, they are often much easier to decode than natural, more complex emotional expressions (Fridenson-Hayo et al., 2016; Krumhuber et al., 2021). As a result, real-life emotional expressions are often ambiguous and highly susceptible to the influence of contextual information (Hassin et al., 2013). Accordingly, the null effects of emotional mimicry on basic emotion recognition accuracy observed in previous studies may stem from the fact that mimicking is less useful when the emotion categorization task is easy to perform. These null results, however, may not preclude the usefulness of emotional mimicry when more complex, real-life emotional expressions are being decoded.

Notably, the idea that emotional mimicry plays a crucial role in emotional understanding when emotional information is relatively difficult to decode is often mentioned in the context of the Simulation of Smiles (SIMS) model (Niedenthal et al., 2010). The model holds that smiles are not unequivocal signals of joy (as suggested by the basic emotion account), but their social meaning is much more complex and dependent on the circumstances in which smiles appear. As a result, differentiating between different types of smiles (for instance, reward and affiliative smiles) is far more difficult than a simple categorization of a smile as such and thus, may require the involvement of emotional mimicry. In line with this assumption, a recent study by Orlowska et al. (2021) has shown that, when verbal contextual information suggested that the affiliative smile appeared in a reward-related context, participants qualified such a smile as a happy (reward) smile, which indicates that they did not recognize it correctly. However, of importance here, this effect was more pronounced when participants’ mimicry was restricted. For participants who could freely mimic the observed smiles, this confusing influence of verbal contextual information on the smile interpretation was reduced. Put another way, restricting emotional mimicry made smile discrimination more difficult and thus increased participants’ reliance on additional cues (i.e., verbally presented contextual information). In a similar vein, Maringer et al. (2011) found support for the role of emotional mimicry when participants differentiated between fake and real smiles. It is reasonable to assume that such a task was difficult because people are not very accurate at differentiating genuine and fake emotional expressions (Ekman & O'Sullivan, 1991; Hess & Kleck, 1994; Zloteanu et al., 2021). This probably explains why emotional mimicry was of help here. Finally, Kowallik et al. (2021) showed that when people with higher levels of autistic traits were instructed to imitate emotions during a basic emotion recognition task, the recognition accuracy increased (in contrast to the condition in which there was no imitation). It seems likely that the helpful role of emotional mimicry here resulted from the fact that the task was quite demanding, given that autism is associated with emotion recognition difficulty (Uljarevic & Hamilton, 2013).

Taken together, these results suggest that emotional mimicry may be useful when the recognition and interpretation of others’ emotions are challenging. This seems highly probable in situations when the perceiver cannot see the expresser's face because, as already mentioned, categorizing emotional expressions based on nonvisual cues is more difficult than when visual modality is involved (Collignon et al., 2008; Paulmann & Pell, 2011). Accordingly, situations such as talking on the phone or participating in an online meeting where the expresser has their camera turned off make emotional recognition quite difficult and therefore, may call for the engagement of additional tools, such as emotional mimicry. This reasoning is indirectly backed up by the above-mentioned study by Hawk et al. (2012), which demonstrated that inhibiting facial mimicry limited the processing of vocal emotional expressions. More specifically, participants were exposed to audio recordings of two contrasting emotion vocalizations—laughing and crying (in a counterbalanced order)—in which an initial sound gradually decreased in volume while the second sound gradually increased in volume (so that both sounds overlapped for a short period). Participants’ task was to identify the moment at which the sounds of laughing and crying overlapped. The results showed that participants who could freely move their facial muscles were slower to respond to the shift between the sounds than participants whose mimicry was restricted. This, according to the authors, suggests that facial mimicry facilitated a stronger focus on the first sound, thereby resulting in ignoring the second (incongruent) sound. Put differently, facial mimicry promoted deeper emotional processing of the first sound, which provides support for the role of cross-channel mimicking in emotional understanding.

We should also note that, unlike the affiliative function of emotional mimicry, which is activated when the context of the interaction is considered affiliative, the role of mimicry in emotional understanding should be relatively independent of the perceived affiliativeness of the context. Here the strategic use of emotional mimicry as a tool that facilitates emotional understanding should depend on the relative difficulty of emotion recognition rather than the affiliative versus nonaffiliative social meaning of the expressed emotion. Importantly, the difficulty of emotion recognition varies across different sensory channels. For instance, when exposed to vocal expressions, people are more accurate in decoding anger and sadness relative to fear and happiness (Hawk et al., 2009; Juslin & Laukka, 2003), but this pattern is different for facial displays, where the highest accuracy rates are observed for happiness, whereas fear and surprise are often confused (Hawk et al., 2009; Olszanowski et al., 2015). As such, the limited visibility of the expresser may boost the role of emotional mimicry in the recognition of emotional expressions which are more ambiguous when presented solely via nonvisual channels. This observation is important because it suggests that both functions of emotional mimicry may be differently activated by diverse emotions depending not only on these emotions’ affiliative versus nonaffiliative social meaning but also on the relative difficulty in their recognition.

Summing up, current evidence indirectly shows that limited visibility of the expresser makes emotion processing more challenging and, therefore, may activate the second function of emotional mimicry, that is, its role in emotional understanding. Future research is needed to address this possibility directly.

Future Directions in Research on Emotional Mimicry Under the Absence versus Presence of Visual Contact

Overall, the reviewed literature provides some initial support for the notion that emotional mimicry occurs regardless of visual contact, but the social functions of this process strategically adapt to whether the perceiver can see the expresser and whether the expresser can see the perceiver. Yet, the mechanisms behind the use of emotional mimicry for different social purposes might differ depending on whether the visibility is reduced on the expresser or perceiver's side. For instance, when the perceiver cannot see the expresser, the increased use of emotional mimicry as a tool for improving emotion understanding may be attributed to the fact that the recognition of the expresser's emotion becomes more challenging. In such a case, mimicry may provide additional information about the expresser's inner state. When the expresser cannot see the perceiver, the use of mimicry to understand the expresser's internal state is also possible, but here it can result from the fact that the affiliative function of mimicry may become less relevant. This may happen because the perceiver cannot see the expresser's gaze, and hence, it is more difficult to tell whether the context of the interaction is affiliative and/or whether the expresser can see their imitative responses.

Currently, the support for these claims remains largely indirect, and future research is needed to provide insight into the mechanisms behind the role of visual contact in using mimicry as a tool for affiliation or emotional understanding. We believe that using an alternative means to transfer emotional information (e.g., vocal expressions instead of facial displays) as well as manipulating the visibility of the perceiver and the expresser may shed light on these mechanisms. An important direction here would be to test whether vocal or cross-channel mimicry is modulated by social contextual factors in a similar way facial mimicry is. If it turns out that emotional mimicry is more likely to occur in affiliative than non-affiliative social contexts, regardless of whether the expresser can see the perceiver or not (that is, regardless of whether the expressed emotion is transmitted via the visual or vocal channel), this will mean that the experience of being watched is not necessary to trigger the affiliative function of emotional mimicry. Therefore, research testing how the functions of emotional mimicry adapt to whether emotional signals are sent (1) only via a visual channel; (2) only via a vocal channel, or (3) simultaneously via both channels may help fill in an important gap in the literature.

Future research would also benefit from including emotional displays of different emotions. As already stated, different emotional expressions vary in (1) the extent to which they convey affiliative versus nonaffiliative information and (2) how easily they are recognized via vocal and visual channels. Therefore, the inclusion of various emotional displays of the expresser, who will be either visible on invisible to the perceiver, may shed light on how challenging emotional understanding is and whether emotional mimicry may improve it.

Another important future direction is to test the role of mimicry in face-to-face, real-life interactions rather than relying on videos or pictures of basic emotions. This direction is important not only because it is more ecologically valid, but also because real interactions might provide a better insight into the effect of the awareness of being watched on the “social glue” function of emotional mimicry. Put another way, the physically present expresser may make the perceiver aware that they are actually being watched by the expresser. Further, emotions spontaneously expressed in real life are typically more difficult to recognize, which may shed light on the role of mimicry in emotion understanding. Finally, real interactions (e.g., video calls) allow for flexibility when arranging situations in which one-way and two-way visual contact between the interacting partners is possible or not. Hence, they seem to be a promising research pathway for studying the role of visual cues on the expresser and perceiver's sides for mimicry's social functions.

Conclusion

Despite a huge interest in emotional mimicry, most studies in this area were focused on mimicking facial displays. Here, we argue that research addressing the imitation of emotional expressions when the expresser's face is invisible (e.g., focusing on vocal or cross-channel mimicry), while still being in its infancy, may be crucial for understanding whether emotional mimicry functionally adapts to the absence of visual cues on the expresser's side. We also propose that reducing visual contact on the perceiver's side (e.g., by making them invisible to the expresser) is important for understanding the effects of eye gaze and the awareness of being watched on emotional mimicry's social role.

We believe that getting insight into how the functions of emotional mimicry are modulated under different circumstances may clarify the role of this process in fostering affiliation and understanding of others’ internal states. Over the years, these two functions have been accentuated somewhat differently by different theoretical approaches. Generally speaking, the embodied cognition theories have primarily emphasized the influence of mimicry on emotional understanding (e.g., Arnold & Winkielman, 2020; Winkielman et al., 2018), while more socially oriented models have focused on mimicry's affiliative role in human interaction (Hess, 2021; Hess & Fischer, 2013). We believe that examining the role of facial mimicry when the expresser cannot see the perceiver or the perceiver cannot see the expresser may bridge the gap between these two frameworks, by showing that the realization of these two social functions does not occur rigidly but follows flexible situation-dependent rules.