Multimodal indexicality in Korean: “doing deference” and “performing intimacy” through nonverbal behavior

1 Introduction

One task faced by participants in communicative events is determining whether the other speaker is being polite or impolite. This interpretation is likely to rely on a confluence of multiple factors, including the use of words conventionally associated with (im)politeness, the sound of the speaker’s voice, the use of polite bodily and facial gestures, and various other nonverbal factors. All things being equal, a speaker who employs politeness routines in a quiet tone of voice while smiling and bowing the head is likely to be perceived as polite, whereas a speaker who shouts and produces obscene gestures will be perceived as impolite. Producing and interpreting the appropriate level of (im)politeness is not always so straightforward, however, as evinced, for example, by the well-documented difficulties that second language learners encounter in appropriately producing and interpreting (im)politeness (e. g., Cook 2001).

Despite the obvious multimodal nature of most social phenomena, research on (im)politeness phenomena has traditionally focused mainly only on verbal means to express (im)politeness, such as polite words or honorific markers. Several authors have noted that the role of nonverbal communication is either downplayed or simply not addressed in past research (e. g., Winter and Grawunder 2012; Brown, Winter, Idemaru and Grawunder 2014; McKinnon and Prieto 2014). Culpeper (2011: 151) observes that “non-verbal cues such as gaze, facial expressions, body movements/gestures […] and the spatial positioning of the self play a key role in communication […] yet it is still an area that receives relatively little attention in communication and pragmatic studies.” In the literature on languages with grammatical honorific markers, such as Korean and Japanese, this neglect of the nonverbal domain in politeness-related phenomena is sometimes associated with the view that honorific markers somehow “contain” politeness. For example, Lee and Ramsey (2000: 260) describe the Korean honorific marker -yo as a “simple switching device” that makes an informal sentence into an honorific one. In contrast to this, the fact that honorific markers in Korean can be used in a sarcastic fashion to achieve impoliteness (Brown 2013) shows that the presence of honorific markers is not strictly associated with politeness-related meanings. Instead, the context and the mode of delivery via nonverbal means have to match the intended politeness level.

This study analyzes a corpus of televised Korean interactions in status asymmetrical situations (a status inferior talking to a superior) and status symmetrical situations (two peers talking to each other). In Korean, asymmetrical interactions warrant the use of an honorific register of speech known as contaymal (존댓말) ‘respect speech’ (see Brown 2011). Symmetrical interactions warrant the use of a non-honorific speech register called panmal (반말) ‘half speech’. Whereas much past research in Korean and other languages with grammatical honorific markers has looked at the verbal domain of politeness expression, this study analyzes the various nonverbal behaviors that co-occur together with honorific or non-honorific language. The significance of the paper lies in the range of nonverbal behaviors that we look at, which goes far beyond the small number of gestures that have been considered in previous multimodal accounts of politeness within the pragmatics literature. Our analyses deliberately focus on a whole swath of different nonverbal features, including body orientation and position, gestures, facial expressions, touching behaviors and asymmetries in the initiation of actions. Similar to Winter and Grawunder’s (2012) attempt at characterizing a “phonetic profile” of the Korean vocal expression of politeness, this study aims at characterizing a “nonverbal behavioral profile” of Korean politeness-related phenomena. Overall, our results speak to the importance of recognizing the multimodality of politeness in human interactions.

2 Background on the nonverbal expression of politeness

Some recent studies already demonstrated that vocal and gestural aspects of message delivery play vital roles in interpreting (im)politeness. Brown et al. (2014) showed that Korean speakers could correctly identify the intended politeness level of sentence fragments without verbal honorific markers with about 70 % accuracy, purely via phonetic factors (Korean honorific speech is lower pitched, slower and more monotonous, Winter and Grawunder 2012). Similarly, Nadeu and Prieto (2011) found that Catalan speakers perceptually associated a high pitch range with politeness, but that this also depended on whether the speaker was smiling or had a neutral facial expression. In another study on Catalan, McKinnon and Prieto (2014) found that prosody, as well as gesture, played an important role in differentiating between genuinely offensive politeness and jovial mock impoliteness. Participants in this study were much better at differentiating mock impoliteness from genuine impoliteness if they were provided with videos, rather than just audio recordings alone. Haugh and Bousfield (2012: 1108) made the observation that gestural cues, including laughing and tilting the head backwards may work as cues for mock politeness in Australian English. Orie (2009) observed how certain ways of pointing are perceived as impolite by Yoruba speakers, while Liu et al. (2017) note that fully extending the arm and/or forefinger when pointing at a person tends to be considered impolite.

The role of bodily behavior in politeness-related phenomena has also been studied in sign languages. As discussed in Mapson (2014), speakers of British Sign Language employ a large range of non-manual features to communicate (im)politeness, including tight lips, polite grimaces, raised eyebrows, narrowed eyes, and tilting or lowering the head. George (2011) observed that in Japanese Sign Language, a smaller signing space and a lowered and forward-leaning chin and head position were associated with heightened speech register (comparable to honorific styles in spoken Japanese). George (2011: 113) speculated that the smaller signs and lowered head “signal a non-threatening act display and in turn have an association with a more polite register.” Similar observations have been made for American Sign Language (Cokely and Baker-Shenk 1980; Liddell and Johnson 1989).

In contrast to pragmatics and politeness research, the literature on nonverbal communication in social psychology extensively deals with nonverbal behavior in politeness-related phenomena. Most of this research has focused on the psychological notions of “power”, “status” and “intimacy”. “Power” is commonly defined as the ability to influence others; “status” commonly involves one’s positioning within a social hierarchy and concomitant expectations and social rules (see Ellyson and Dovidio 1985). Even though these psychological constructs do not directly map onto the categories used in politeness research, they are still relevant because power, status and intimacy (or “social distance”) are recognized as key factors embedded in the communication of politeness. Indeed, in their seminal work on purported politeness universals, Brown and Levinson (1987) include power (defined as “an asymmetrical social dimension” - p. 82) and distance (“a symmetric social dimension of similarity/difference” - p. 81) as two of the three factors that determine the level of politeness required in a given situation. Subsequent studies found that power is a fairly reliable predictor of politeness-related behavior (see Goldsmith 2007: 227 for overview). For example, Holtgraves and Yang (1992) found that Korean and American respondents use more polite request strategies when addressing higher status interlocutors (professors), with Koreans varying their responses according to power more strongly than the Americans. Distance (intimacy) has also been shown to interact closely with politeness, although the relationship appears to be more context-specific: Holtgraves and Yang (1992) found that respondents made the most polite requests when addressing complete strangers. In contrast, Baxter (1984) found that students used more polite strategies towards close friends when making certain highly face-threatening requests (e. g., asking a fellow student to redo their part of a group project). These studies serve to show that issues of politeness are closely intertwined with issues of power and intimacy, which means that social psychological studies of nonverbal behaviors of these dimensions are relevant to politeness research.

Research in social psychology reveals that power is communicated non-verbally through behaviors implying strength, comfort-relaxation, and fearlessness, whereas submissiveness is communicated with behaviors implying weakness, smallness, discomfort, tension and fearfulness (Mehrabian 1971: 47). Superiors adopt large, relaxed postures, such as sitting with legs splayed and/or arms akimbo and head tilted back, and they may perform large or intense movements and changes in facial expressions (Burgoon and Dunbar 2006: 289). They use more arm gestures, and also touch their own faces or bodies more frequently. In contrast, inferiors adopt stooped and rigid postures, with a downturned head and/or forward lean (Harper 1985: 34). Whereas superiors orient themselves away from or even turn their backs on low status interlocutors (Mehrabian 1971), inferiors maintain direct bodily orientation (Mehrabian 1968; Jorgenson 1975; Burgoon and Saine 1978). Status superiors also enjoy more freedom in the employment of eye contact (Dovidio and Ellyson 1982): they have the power to stare at status inferiors (Snyder and Sutker 1977), but also the power to break eye contact first (Strongman and Champness 1968). Inferiors, on the other hand, display attentiveness by maintaining gaze on the status superior. Superiors furthermore enjoy the privileges of “precedence” (i. e., “going first” when entering a space, starting to eat, etc.) and of “prerogative” (i. e., freedom to initiate and control behaviors associated with intimacy, such as touch) (Burgoon and Dunbar 2006: 289). Whereas status equals may reciprocate each other’s nonverbal behavior, status superiors have leeway to “meet another’s smile with a blank expression” (Burgoon and Dunbar 2006: 291). It is the status superior rather than the status inferior who will initiate interactional patterns, to which others are expected to orient.

Whereas nonverbal behaviors that communicate power work to enhance the potency and control of the status superior over the status inferior, nonverbal behaviors associated with intimacy create involvement and shared positive affect. These behaviors tend to convey immediacy, expressiveness, and altercentricism (i. e., showing attention to the interlocutor) (Burgoon and Newton 1991). Intimacy is signaled by assuming closer conversational distances, leaning the body forward and orienting oneself directly towards one’s interlocutor, as well as through sharing of the same physical plane (i. e., both sitting, both standing) (Andersen, Guerrero and Jones 2006: 265). Direct body orientation is associated with “involvement and immediacy”; side-to-side or back-to-back orientation is associated with being “cold, unavailable and uninvolved” (Andersen 1999: 194). Intimate relationships are also marked by haptics (i. e., physical touch), and through smiling, displaying expressiveness in the face, maintaining eye contact, and gesturing frequently (see Guerrero and Floyd 2006: 87-88).

The specific politeness contrast at the focus of this study involves the distinction between two interrelated relational practices in Korean: “doing deference” towards elders and superiors (i. e., in +power contexts) and “performing intimacy” towards status equals (-power, -distance). We follow Haugh, Chang, and Kádár (2015) in defining doing deference as “submitting to or showing regard to a superior or someone else deserving of respect.” In South Korea, doing deference is associated with pervasive politeness-related neo-Confucian social slogans including kyenglosasang (경로사상) ‘respecting the elderly’ (Yoon 2004: 198) and cangyuyuse (장유유서) ‘the old and the young know their place’ (Lee and Ramsey 2000: 267), which are “promoted very actively and widely at the national level” (Yoon 2004: 198). As argued by Brown (2011: 80), the idea of elder respect in the Korean context is qualitatively different to corresponding Western notions in that disagreeing with, contradicting or causing discomfort to elders is largely taboo. Performing intimacy is also associated with Korean mottos and slogans that promote intimacy and affection between status equals such as ceng (정) - “a feeling of psychological solidarity, sympathetic emotion or affection which grows over time between people who are close” (Brown 2011: 84). Importantly, the idea of performing intimacy is orthogonal to doing deference in Korean society since friendship tends to connote equality and similarity, whereas deference is strongly associated with status inferiors showing respect to superiors and elders.

Doing deference and performing intimacy directly map onto distinct verbal styles in Korean. When doing deference, speakers normatively use an honorific register of speech known in Korean emic terms as contaymal ‘respect speech’. This style of speech involves the use of honorific verbal suffixes such as -yo and -supnita, which prototypically appear in all utterances addressed to elders, status superiors and adult strangers, although these forms may also be used for other functions such as sarcasm (Brown 2013) and modulating stance (Strauss and Eun 2005; Brown 2015). Performing intimacy, on the other hand, typically involves using a non-honorific speech register called panmal ‘half speech’, which features the verbal suffixes -e or -ta.

Recent research has shown that “doing deference” and “performing intimacy” in Korean not only involve the verbal domain, but also phonetics. Winter and Grawunder (2012) found that contaymal ‘respect speech’ addressed to a superior was spoken with lower pitch, smaller pitch range and lower pitch variability. Furthermore, contaymal was slower, quieter and less breathy. These acoustic differences were demonstrated to be perceivable in the absence of any verbal honorific markers in Brown et al. (2014).

Although these studies already show that “doing deference” and “performing intimacy” are indexed through multiple communication channels (i. e., speech acoustics and lexical and grammatical markers), the visual modality (body orientation, gesture etc.) has not yet been investigated with respect to these behaviors. The current paper extends this line of research by exploring the bodily behaviors that speakers of Korean engage in when interacting with status superiors (+power) and intimate status equals (-power, -distance), focusing on the behaviors identified as interacting with politeness, power and intimacy in the research reviewed earlier in this section. To this end, we collected clips from three Korean televised dramas of the same characters interacting in (+power) and (-power, -distance) settings and analyzed various nonverbal behaviors, detailed in the next section.

3 Methodology

In order to compare the different nonverbal behaviors used when “doing deference” and “performing intimacy”, we extracted clips from the three Korean television dramas, Bad Guy (SBS 2010; Korean title: 나쁜 남자), Pinocchio (SBS 2014; 피노키오), and Two Outs in the Ninth Inning (MBC 2007; 9회말 2아웃). These dramas were selected because all three featured prominent male and female characters who interacted with same-sex superiors and same-sex intimate status equals. Televised materials have been widely used in (im)politeness research (e. g., Culpeper 2005; Bousfield 2008), including in studies of Korean and Japanese honorifics (e. g., Barke 2010; Brown 2013). These shows generally feature a rich range of (im)politeness-related contexts, including conflict and aggression, which can be difficult to capture in recordings of authentic interactions. Although televised interactions may be scripted, and therefore simplified and idealized, the use of nonverbal communication in these materials must still relate to that found in “real-world” language use in order for it to be recognizable to viewers. Moreover, whereas scripts often specify the details of verbal interactions, the actual in-the-moment execution of nonverbal behavior is often left to the actor’s interpretation.

In each of these three dramas, we focused on one leading female and one leading male character. In Bad Guy, we focused on Jae-in, a gallery art consultant in her early 30s, and Gun-wook, a company employee also in his early 30s. In Pinocchio, we selected In-ha, a news reporter in her mid-twenties, and Dal-po, who works at the same news department as In-ha and is also in his mid-twenties. In Two Outs in the Ninth Inning, we concentrated on two 30 year olds: female Nan-hui who works in a publishing company, and male Hyeong-tae, who works at an advertising company. All of these characters represent young professionals in their mid-twenties or early thirties. In the analysis that follows, we refer to the main characters as “protagonists”, and to the characters they interacted with as “interlocutors”. When referring to both the protagonists and the interlocutors together, we use the term “participants”.

For each of these six characters, we identified a same-sex superior and a same-sex intimate status equal with whom they had regular interactions.^[1] We then extracted video clips of all of dyadic interactions between these characters. This produced a dataset of 92 asymmetrical interactions between status inferiors and status superiors (1:54:33, 5,682 words), and 62 symmetrical interactions between intimate friends (1:01:17, 4,960 words). In the description below, we variously refer to the asymmetrical interactions as “honorific settings” or as indicating the behavior of “doing deference”, and we refer to the symmetrical interactions as “intimate settings” or “performing intimacy”. We then transcribed all of the interactions and annotated them for the non-verbal categories shown in Table 1.^[2] In order to create comprehensive profiles of deferential and intimate behaviors, we looked at a wide range of categories, all of which are identified in the previous literature as interacting with politeness, power and intimacy (see Section 2).

4 Results

4.1 Body positioning

This section looks at the ways in which protagonists arrange their bodies (standing or sitting) and posture when doing deference with status superiors and when performing intimacy with peers. The emphasis is on establishing the normal, static or “resting” position of the body when people perform these two distinct relational practices.

When performing intimacy, participants almost always interacted on the same physical plane (see Table 2). In other words, both participants stood or both participants sat. This characterized 86 % of all intimate interactions. When participants interacted with superiors, there was still a preference for interactions on the same plane, but a much smaller one (68 %). In many cases of doing deference, the inferior protagonist stood, while the superior was seated (26.4 %). This pattern hardly ever occurred when performing intimacy (5.7 %). A Chi-Square test^[4] reveals that there are reliably more symmetrical body plane orientations in the intimate settings compared to honorific settings (χ²(3) = 61.92, p < 0.0001). Conversely, honorific settings were characterized by a higher proportion of asymmetrical body positioning where the protagonist and the interlocutor assumed different body planes.

Out of the 33 scenes in honorific settings that featured asymmetrical body plane orientation, 26 (79 %) involved the status inferior visiting the office or approaching the desk of the status superior, with the inferior remaining standing during the interaction and the superior remaining seated (see Figure 1). On top of the status difference between protagonist and interlocutor, the difference in plane may be associated with entering somebody’s territory, such as the office of a superior. Burgoon, Buller and Woodall (1996: 307) note that “people entering someone else’s territory become more deferent and submissive” whereas “those on their ‘home court’ gain confidence from the familiarity of their surroundings”. Moreover, it is customary for status inferiors in Korean culture (and in many other cultures) to not sit down unless being requested to do so or unless being offered a seat, which might explain why protagonists often remained standing for the entirety of the interaction.

Table 2:

Number of scenes featuring standing and sitting body positions

		Symmetrical body planes		Asymmetrical body planes
		Both standing	Both sitting	Protagonist standing; Interlocutor sitting	Protagonist sitting; Interlocutor standing
1	Doing deference	53* (42.4 %)**	32 (25.6 %)	33 (26.4 %)	7 (5.6 %)
2	Performing intimacy	34 (48.6 %)	26 (37.1 %)	4 (5.7 %)	6 (8.6.%)

*Scenes that feature two different body positions (i. e., participants changing position during the scene) are counted as two separate scenes for this analysis.

**Percentage of clips in that relational practice that featured this behavior, i. e., percentages correspond to row-wise proportions.

Figure 1.

Status inferior Dal-po remains standing when interacting with his status superior Gyo-dong. (Pinocchio, Episode 7)

We now look more closely at the specific postures that participants adopted while they were seated or standing. Doing deference was characterized by an erect and compact body position. The protagonist never slouched (i. e., sitting or standing with hunched shoulders and/or with a bent or leaning posture) in honorific settings, whereas slouching did occur in the behavior of the superior and in intimate situations (Figure 2). A simple binomial test indicates that protagonists were more likely to slouch in intimate (8 instances) than in honorific settings (0 instances) (p = 0.008). In honorific settings, the superior interlocutors slouched in 14 instances, which a binomial test indicates to be reliably more often than inferiors slouched (0 instances, p = 0.0001). In status equal situations, protagonists were as likely to slouch as their interlocutors (also 8 times), with no statistically reliable difference (p = 1.0). The asymmetry in slouching behavior is present regardless of whether one looks at seated or standing slouches.

Figure 2.

Jae-in (left) adopts a slouched body position as she interacts with friend Ju-yeon. (Bad Guy, Episode 8)

Sitting or standing with the knees together (a marker of a compact body position) was strongly associated with doing deference. This was often accompanied by a lowering of the head. Sitting with the knees together particularly characterized the behavior of female status inferiors, as exemplified in Figure 3a. In honorific settings, protagonists adopted a “knees together” position 26 times in total (summing over both standing and seated contexts), which was more often than superiors adopted this posture (11 instances) (binomial test, p = 0.02). In status equal interactions, protagonists adopted a “knees together” posture only 6 times in total, less so than in honorific settings (26 instances, p = 0.0005). In intimate settings, protagonists (6 instances) and interlocutor (8 instances) showed no statistically reliable difference for their knee positioning (p = 0.79). There were a total of 12 occurrences of superiors adopting a cross-legged position, whereas inferiors adopted this position only 3 times (a statistically reliable difference, p = 0.035). This suggests that the possibility of assuming a cross-legged position is associated with high status. In honorific settings, inferiors furthermore clasped their hands in front of their body (when standing) or in their lap (when seated). This behavior occurred 17 times in total^[5] when doing deference, and never appeared when performing intimacy (p < 0.0001) (see Figure 3b).

Figure 3a.

Status inferior In-ha (left) adopts a compact body position with knees and feet together and hands in lap when doing deference to status superior Cha-ok. (Pinocchio, Episode 5)

Figure 3b.

Status inferior Jae-in (right) clasps her hands in front of her body in this interaction with status superior Mrs. Shin. (Bad Guy, Episode 7)

4.2 Body orientation

Protagonists exhibited distinct patterns of bodily orientation when doing deference compared to when performing intimacy. Despite an overall preference for the protagonists to face the interlocutor and maintain gaze, these tendencies were stronger when protagonists were involved in honorific interactions (see Table 3). The protagonists maintained direct bodily orientation towards the interlocutor 89.5 % of the time when interacting with superiors, and they also maintained gaze on the line of sight (i. e., made or sought eye contact) 84.3 % of the time. When interacting with status equal intimates, the protagonists maintained direct bodily orientation only 81.6 % of the time, and gaze only 65.9 % of the time. We performed a simple paired Wilcoxon signed-rank test on the cumulative time for intimate versus deferential setting (each protagonist contributes one data point, N = 6), which showed that protagonists orient their body more to superiors than to inferiors (V = 21, p = 0.03). The same test also showed that protagonists look reliably more towards superiors (V = 21, p = 0.03).

Table 3:

Body orientation and eye contact (time)*

	Body Orientation				Gaze
	Protagonist		Interlocutor		Protagonist		Interlocutor
	Towards Interlocutor	Away from Interlocutor	Towards Protagonist	Away from Protagonist	On Interlocutor’s line of sight	Away from Interlocutor’s line of sight	On Interlocutor’s line of sight	Away from Interlocutor’s line of sight
Doing deference	1:23:06 (89.5 %)	9:44 (10:5 %)	1:19:52 (86 %)	12:58 (14 %)	1:00:40 (84.3 %)	11:16 (15.7 %)	57:34 (80.5 %)	13:57 (19.5 %)
Performing intimacy	37:01 (81.6 %)	8:21 (18.4 %)	39:02 (86.3 %)	6:11 (13.7 %)	23:29 (65.9 %)	12:08 (34.1 %)	25:57 (72.2 %)	10:00 (27.8 %)

*Figures give the total time in hours, minutes, and seconds for which the behavior occurred, but only include times when participant orientation or eyes are visible.

The analyses just reported compare the protagonist’s body/gaze orientation across intimate and honorific settings. How does the protagonist’s body/gaze orientation compare to the orientation of their superior interlocutors in honorific settings? Overall, inferiors maintained higher direct bodily orientation and gaze than their superior interlocutors (89.5 % versus 84.3 % for bodily orientation; 86 % versus 80.5 % for gaze). Many encounters had the inferior maintaining direct bodily orientation and/or gaze, which the superior did not reciprocate (see Figure 1, Section 3.1), however, this numerical trend was not statistically reliable (both: V = 15, p = 0.44).

Inferiors also took explicit measures to maintain direct bodily orientation and/or gaze. For example, in a scene from Episode 9 of Pinocchio, status inferior In-ha and a male colleague meet status superior Cha-ok as she is about to board an elevator (Figure 4). As Cha-ok walks into the elevator, In-ha and the colleague deliberately shuffle their feet around so that their bodies are oriented roughly in the direction of the superior. Throughout the encounter, Cha-ok maintains orientation away from In-ha and the colleague.

Figure 4a.

Status inferior Jae-in (far right) and male colleague bow to status superior Mrs. Shin as she prepares to go into the elevator. Their bodies are oriented towards the left of the picture (i. e., the current location of Mrs. Shin). (Bad Guy, Episode 9).

Figure 4b.

As Mrs. Shin approaches the elevator, they shuffle their positions round to the right. Note the difference is the Jae-in’s torso and feet positions.

Although the general tendency was for status inferiors to maintain direct bodily orientation and gaze, there were exceptions. Notably, status inferiors avoided gazing at the superior’s line of sight when put in a position of embarrassment, being blamed, or accepting wrongdoing. Although some interactions showed status inferiors aiming their gaze at the chest of status superiors (see Figure 3a), which is claimed in previous studies to be polite in Asian culture (Koo and Jeon 2007: 76), such behavior was not widely represented.

4.4 Manual gestures

The data showed an overall low frequency of manual gestures. This could be taken as evidence of the lower rates of manual gestures claimed for Asian societies (Koo and Jeon 2007: 76). However, it could equally well be due to conventions of Korean TV dramas whereby acting tends to focus on facial expression and voice quality, especially because many shots show the actors’ faces from up close, rather than showing the entire body. Despite overall low gesture rates, there were noticeable differences between the two types of interactive situations (Table 5). Overall, there were only 54 gestures (regardless of which type) in the honorific settings, fewer gestures than when performing intimacy (102 gestures, p = 0.0001). However, it was the case that protagonists performed fewer gestures than their interlocutors in both honorific and intimate settings, which may be part of a TV drama convention. In honorific settings, protagonists performed 16 gestures (29.6 %), compared to 38 gestures performed by status superiors (70.4 %). In intimate settings, protagonists performed a similarly low number of gestures (36 gestures, 35.3 %), much fewer than their interlocutors (66, 64.7 %). It is, however, noteworthy that waves (i. e., waves of the hand) were entirely absent in honorific settings. Figure 5 shows a nice contrast between how bows and waves have similar functions, but one occurs when doing deference and the other occurs when performing intimacy.

Table 5:

Frequency of manual gestures

	Points		Waves		Other manual gestures
	Protagonist	Interlocutor	Protagonist	Interlocutor	Protagonist	Interlocutor
Doing deference	raw	7	16	0	0	9	22
	per minute	0.06	0.14	0	0	0.08	0.19
Performing intimacy	raw	14	34	5	6	17	26
	per minute	0.23	0.55	0.08	0.10	0.28	0.42

Figure 5a.

Status inferior Hyeong-tae bows to status superior Pwucang. (Two Outs …, Episode 2)

Figure 5b.

He then waves over his shoulder to friend Jin-seung. (Two Outs …, Episode 2)

4.5 Self-touches

There were numerically more self-touches when performing intimacy (32 intimate as opposed to 26 deferential), but this difference was not statistically reliable (binomial test, p = 0.51). In honorific settings, there was a numerical trend for superiors to use self-touches more frequently (17 instances, 65.4 %) than inferiors (9, 34.6 %) (Table 6), although this was also not statistically reliable (p = 0.17). Protagonists and interlocutors used self-touches about equally (15 vs. 17) in intimate settings. Of a total of 58 self-touches in the data, exactly half (29 occurrences) involved touching the face, with the mouth being the most frequently touched body part (10 occurrences).

The self-touches in intimate settings at times functioned as speech-accompanying gestures (Figure 6a) or as a component of a casual sitting or standing posture (Figure 6b). In one particular incident in Pinocchio, after female friend Yoo-rae smelled In-ha’s hair to confirm that she had washed it, In-ha responded to this infringement of her personal space by flicking her own hair backwards and in the direction of Yoo-rae (Figure 6c). These incidents of self-touching all equate with behavior that is too animated or casual to be freely used in interactions with superiors.

Figure 6a.

Chan-soo pats his own chest as a speechaccompanying gesture while interacting with friend Dal-po. (Pinocchio, Episode 19)

Figure 6b.

Dal-po holds his head cupped in his hands in this interaction with Chan-soo. (Pinocchio, Episode 7)

Figure 6c.

In-ha flicks her hair towards Yoo-rae. (Pinocchio, Episode 7)

Table 6:

Frequency of self-touches

	Protagonist	Interlocutor
Doing deference	raw	9	17
	per minute	0.08	0.15
Performing intimacy	raw	17	15
	per minute	0.28	0.24

4.6 Haptics

As shown in Table 7, other-directed touches were associated with the situation and whether the touch was initiated by the protagonist or the interlocutor (χ²(1) = 5.93, p = 0.014). In honorific settings, there were more other-touches initiated by superiors (33 instances, 70.2 %) than by inferiors (14 instances, 29.8 %, binomial test: p = 0.008). In intimate settings, there was no statistically reliable difference (p = 0.57), with about equal proportion of protagonist-initiated touches (42, 53.8 %) and interlocutor-initiated touches (36, 46.2 %). Examples of haptic behaviors found only when performing intimacy included embracing, putting arms around each other’s shoulders, walking arm in arm (females only) and holding hands (females only; Figure 7).

Table 7:

Frequency of haptic behaviors

	Protagonist	Interlocutor
Doing deference	raw	14	33
	per minute	0.12	0.29
Performing intimacy	raw	42	36
	per minute	0.69	0.59

Figure 7.

Female friend Ju-yeon (right) holds Jae-in’s hands. (Bad Guy, Episode 3)

Protagonists rarely touched superior interlocutors (only 14 instances in total). When this happened, this often accompanied unusual plot developments. For example, in Pinocchio, status inferior Dal-po embraced his superior Gyo-dong when the latter found out a contact number for his long lost older brother, who he had not seen since childhood. The marked nature of Dal-po’s behavior is confirmed by the surprised reaction of onlookers in the office who stare at the spectacle. More commonly, any physical contact occurring in honorific settings is instigated by the superior, with the inferior tending not to reciprocate the behavior. Superiors are seen touching the shoulders, arms and legs of the inferiors. In Figure 8, superior Mr. Hong greets Gun-wook by patting his upper arms with both hands, while Gun-wook stands motionless.

Figure 8.

Status superior Mr. Hong (left) greets Gun-wook by patting his upper arms. (Bad Guy, Episode 7)

4.7 Prerogative

The data revealed different patterns for doing deference and performing intimacy in terms of who had the prerogative to initiate certain actions. We looked at actions that were reciprocated by both participants, but which were asynchronous, i. e., they were initiated by one of the participants before being picked up by the other. The specific actions that we analyzed included walking, sitting down, standing up, touching the other participant and drinking alcohol and/or proposing a toast. The overall frequency of these asynchronous reciprocated actions was similar across honorific (22 instances) and intimate settings (13 instances) (binomial test, p = 0.18) (Table 8). However, the honorific settings showed a pattern not found in the intimate interactions: it was almost always the interlocutor who initiated the action, 19 as opposed to 3 times (p = 0.02). Allowing the interlocutor to go first therefore appears to work as a display of deference.

Table 8:

Frequency of initiation of asynchronous reciprocated actions

		Protagonist	Interlocutor	Total
Doing deference	raw	3	19	22
	per minute	0.03	0.17	0.19
Performing intimacy	raw	8	5	13
	per minute	0.13	0.08	0.21

The drama data featured only two scenes in which status inferiors drank alcohol with status superiors. In these two scenes, we see examples of alcohol related etiquette, which are well-known rituals for doing deference in Korea. As shown in Figure 9, the status superior raises her glass to propose a toast first, and holds her glass higher than that of the inferior when chinking glasses together. These practices metaphorically index the primacy and heightened status of the superior. After chinking their glasses, the status inferior turns her head to the side while drinking as if to hide the act of drinking from the superior (this relates to the fact that drinking alcohol in front of superiors was traditionally taboo in Korean culture). Note that throughout these rituals the status inferior holds the cup with two hands, whereas the superior uses only one. Using two hands when drinking as well as when giving or receiving something is a clear signal of deference in Korean culture (cf. Dennison and Bergen 2010).

Figure 9a.

When chinking toast, status superior Choo-ja (left) raises her glass higher than that of status inferior Nan-hui. (Two Outs …, Episode 16)

Figure 9b.

Nan-hui (right) turns her head slightly to the side as she drinks.

5 Discussion

The analysis has revealed distinct patterns of non-verbal behavior associated with two interrelated relational practices in Korean: doing deference and performing intimacy. Doing deference was characterized by nonverbal behavior that can be interpreted as reflecting decreased animatedness. This was evidenced by the fact that fewer manual gestures were used overall. Moreover, in honorific settings involving superiors, the protagonist actively curtailed their level of activity by assuming a relatively erect but constrained position, marked by knees together or by the hands being crossed (in particular for females). This position essentially means that protagonists cannot as readily move as when their hands are free. Thus, protagonists move less, but they also actively signal that they have less affordance for movement in deferential situations. Although maintaining erect body positions is low in terms of animatedness, it requires a relatively high degree of effort to maintain such rigid body positions. Thus, showing effort through maintaining rigid postures, as well as direct bodily orientation and eye contact, is associated with deference.

Compared to the status inferior, the status superior has relatively more freedom in the nonverbal domain. Whereas inferiors tended to repress manual and head gestures as well as haptic behaviors, superiors were more expressive in their use of gestures and had the freedom to touch the status inferior. Inferiors maintained direct bodily orientation and eye contact, whereas superiors enjoyed the freedom to withhold such behavior. Doing deference also entailed allowing the superior to go first, including in the performance of alcohol drinking rituals. The fact that inferiors follow the actions initiated by superiors directly relates back to the psychological notion of power, and what the social psychology literature refers to as “resource control” (Burgoon and Dunbar 2006: 289). Power, defined as the ability to control other peoples’ behaviors (see above), is here associated with the ability to curtail the behavior of inferiors. However, this curtailing is not done actively, it happens through the expectations that come with status differences.

Also related to power, the results show that inferiors index deference through the adoption of non-threatening, non-confident and submissive postures, whereas superiors engage in “power posing” (Carney, Hall, and LeBeau 2005), i. e., open and wide body positions that communicate dominance and confidence. By adopting constrained body positions with knees together and hands crossed, inferiors create postures that are small and which signal that they are not in a position to threaten the power of the superior in any way. In this way, power and the indexing of relative status works according to the principal of “physical potency” (Burgoon and Dunbar 2006: 287-288): bodily behaviors that mimic prototypical threat displays on the part of the superior, and fright displays on the part of the inferior. Previous research notes important interactions between the adoption of powerful and submissive postures with gender (see Burgoon and Dunbar 2006: 284-285). Some evidence was found for this in our data too, with female status inferiors adopting higher frequencies of knee-together postures and hand clasps, and also using smiles to mark deference. The gender differences in Korean deferential behavior warrant further investigation in future research, including in cross-gender dyads which were not explored in the current study.

Performing intimacy was characterized by animatedness, casualness and frequent reciprocated movements. There was overall more symmetry of nonverbal behaviors in intimate situations, with similar frequencies across several different nonverbal actions. Speakers preferred to interact on the same physical plane, and both the protagonist and the interlocutor could adopt relaxed or even slumped postures. Manual and head gestures were used more frequently by both participants, and there was prolific self-touching and haptic behavior, some of which was gender specific (walking arm-in-arm and holding hands were only attested for females). The constraints that are imposed on protagonists in honorific settings do not apply to them in symmetrical, intimate situations.

To a large extent, the behavior exhibited reciprocally by friends in intimate situations mirrors the kind of behavior used nonreciprocally by status superiors in honorific interactions. However, it is worth noting that in general, friends interacting in intimate settings displayed even more freedom and lack of inhibition in their actions than the status superiors did in honorific settings. This finding generally mirrors claims in the previous literature on nonverbal behavior in intimate interactions, which claim that increased intimacy results in increased nonverbal involvement in general (Andersen et al. 2006: 263). In relation to the South Korean context shown in the current data, it may be the case that status superiors are becoming increasingly constrained in their use of “power posing” towards status inferiors due to the gradual erosion of traditional hierarchical social structures. Indeed, this is reflected in changing patterns of honorifics use whereby status superiors such as college professors now often reciprocate contaymal (honorific language) with their students, and parents reciprocate panmal (non-honorific language) with their children. Kim-Renaud (2001: 42) points out that being overtly power-conscious, subservient or self-effacing is falling out of favor in modern day South Korea, and is being replaced by what she refers to as a “friendlier” and “nicer” mode of politeness.

The findings summarized so far mirror previous research in the social psychology literature that have looked at power, status and intimacy. However, the paper did uncover some behaviors that do not appear to be reported in the previous literature, perhaps because they are specific on some level to Korean culture. Such behaviors include the alcohol-related rituals described above, as well as the handclasp used by female inferiors when doing deference. In addition, the findings of the paper call into question some previous claims and assumptions regarding nonverbal behavior in Asian societies. We found that physical contact was very frequent and also richly varied in its expression, which stands against the findings of other previous studies that Asian people engage in less physical touching than other cultures (e. g., McDaniel and Andersen 1998). The frequency of physical contact in our data may be due to the contexts that we examined, or it may be due to Korea differing from other Asian societies investigated with respect to the social use of haptic behavior. We also found little evidence for previous claims (e. g., Koo and Jeon 2007: 76) that Korean status inferiors avoid eye contact with status superiors as a way of doing deference. Rather, our data shows that Korean status inferiors prefer to maintain gaze on the superior’s face in order to show attentiveness, which matches observations about Western interactions in the social psychology literature (e. g., Burgoon and Dunbar 2006: 290). It appears that any cultural-specific preference for avoiding eye contact may apply only to specific contexts, such as when an inferior is being scolded by a superior. It might also be the case that eye contact avoidance only occurs when interacting with the elderly, who tend to be treated with an extra level of deference in Korean society.

Although South Korea may be gradually becoming a less hierarchical society (Kim-Renaud 2001), the consistency and strength of the findings in this paper show that vertical power dimensions still play an important role in determining appropriate nonverbal behavior in the workplace settings represented in our data. We believe that these findings have important implications for intercultural communication, including potentially the training of second language learners and/or business people from other countries looking to work in Korea. It can easily be imagined that the need for status inferiors to constantly exhibit behaviors that are associated with submission and a lack of confidence may be difficult to acquire by Americans, Europeans and other Westerners where being submissive and appearing unconfident are often viewed in a negative light. On this point, the social psychology literature on nonverbal behavior (produced in the Western context) suggests several negative results of unconfident or “low status” behavior, including that it will decrease one’s employability (see Remland 2006: 507), and that it will result in the exhibitor of the behavior experiencing more negative emotions (Anderson and Berdahl 2002). Although these claims may also apply in large part to the Korean context, what they perhaps obscure is the fact that in some contexts and in some cultures displaying confidence-related behavior may be socially inappropriate. For example, although a junior businessman may be praised for giving a confident presentation in an American or European business setting, displaying the same level of confidence in a Korean business setting could be viewed as impudent or immature. For successful interaction in Korean society, status inferiors may require adeptness at displaying unconfident (as well as confident) behavior.

The findings from our paper play an important role in bringing the tradition of studying nonverbal behavior from the social psychology literature into the emerging field of multimodal (im)politeness. Crucially, the results of the current study correspond in notable ways with previous findings on acoustic dimensions of Korean politeness, as well as morphological characteristics. The current paper found that doing deference was characterized by low animatedness, as evidenced by the restricted range of movements and facial expressions displayed by inferiors in honorific settings. This corresponds to the acoustic properties of Korean honorific contaymal speech, which is slower, quieter, lower pitched, and less varied, and thus sounds less animated overall (Winter and Grawunder 2012). Similarly, the increased effort found in the rigid postures of nonverbal behavior used by status inferiors in the current study is mirrored by the increased effort of the acoustic domain of contaymal, which is characterized by speech that is more tensed (Winter and Grawunder 2012).

In addition, the decreased freedom experienced by status inferiors in what nonverbal behaviors they can use in honorific settings is reflected in the variation in the use of morphological honorifics. Just as status inferiors have fewer options in terms of nonverbal behavior, they also have less freedom to choose which honorific verb endings to use. For inferiors, use of honorific contaymal is practically obligatory according to normative social convention, whereas superiors enjoy more degrees of freedom as to whether they use contaymal or non-honorific panmal, or a combination the two registers. Here, it is worth noting that the growing literature on speech style switching in Japanese and Korean (e. g., Strauss and Eun 2005; Dunn 2010; Brown 2013; Cook 2011; Brown 2015) rarely features examples of inferiors switching to non-honorific levels when directly addressing superiors (for one exception, see Lee 1996: 216-220).

We see therefore that doing deference towards superiors is indexed by a variety of channels, including verbal honorifics as well as multiple different acoustic markers and multiple different nonverbal behaviors. One view of behaviors in politeness-related phenomena is to see them as a large set of rules, all of which have to be followed simultaneously by the speaker. However, from the side of the observer of communicative acts, each nonverbal behavior can be seen as a cue for the intended politeness reading of an utterance. Mason, Domínguez, Winter and Grignolio (2015), in a discussion of redundancy in communication systems, make the argument that given the significance of politeness in Korean society, having multiple cues for honorification is essential. Just as linguistic contrasts are “redundantly” encoded via multiple channels (e. g., voicing contrasts are signaled simultaneously through voice onset time, pitch, duration and other acoustic cues; see Winter 2014), social contrasts are also simultaneously signaled through multiple channels. In the case of Korean, politeness-related phenomena are simultaneously signaled through verbal markers, vocal features (Winter and Grawunder 2012; Brown et al. 2014) and - as this study investigated - through a large and variegated set of bodily behaviors. This essentially helps to make politeness meanings something that can be robustly and efficiently encoded and decoded by interactants. And it serves to show that politeness does not exclusively lie “in” any of these channels but is an inherently multimodal phenomenon.

6 Conclusion

This paper has important implications for politeness research, which until recently, has focused on characterizing only the verbal dimension of politeness. When coupled with the results of previous studies looking at speech acoustics (Winter and Grawunder 2012; Brown et al. 2014), the results show that the politeness-related relational practices of doing deference and performing intimacy in Korean are inherently multimodal in nature.

This multimodal approach to honorifics now needs to be applied more broadly into the growing body of literature on Korean as well as Japanese honorifics. One particularly fertile area of recent research on honorifics has been the study of honorific shifting (e. g., Strauss and Eun 2005; Dunn 2010; Brown 2013; Cook 2011). This research has shown that speakers use shifts between honorific levels, as well as the mixing of honorific and non-honorific speech as important resources for indexing various speaker stances. Brown (2013), for example, shows how status equal intimates create sarcastic meanings through mixing honorific morphemes with non-honorific morphemes and other casual language. We can imagine however, that if research were to consider multiple cues for deference including acoustics and nonverbal behavior, this would show a more complex picture of how honorific shifting works in real world interactions. Due to the existence of multiple cues for doing deference, interactants probably deliberately create mismatches between the different social signals and potentially create even more nuanced social meanings than those captured in the previous literature, that has focused mainly on verbal honorification. For example, a speaker may choose honorific markers in speech while at the same time engaging in power posing to convey mockery and sarcasm (see Brown, 2013), thus creating a mismatch between the different politeness-related channels. Taking into account what the current paper has found about the nonverbal features of doing deference, future studies will need to look at the roles that multimodal signals such as power posing, frowns, or loud and high-pitched speech play in the complex meanings produced by honorific shifting (see also Mehrabian 1972).

More broadly, the robust multimodal nature of deference and intimacy in Korean that we have discovered in our own research underlines the need for more widespread adoption of multimodal analysis in politeness research. Analyzing verbal forms of politeness is enlightening, but when these verbal signals are studied in isolation, this potentially obscures the roles of other modalities, which may or may not correspond with the verbal signals. Going forward, research will need to fully embrace the inherently multimodal nature of politeness displays in different cultural contexts.