Multimodal expression of impoliteness in YouTube reaction videos to transgender activism

This study examines the genre of YouTube reaction videos as a distinct form of cultural production and social influence in online communication. Despite its prevalence and popularity, the genre has received limited scholarly attention, particularly with regard to reactions to ideological activism. This paper aims to fill this gap by conducting a social semiotic discourse analysis of videos reviewing the activism of transgender community on TikTok. The analysis demonstrates how intersecting non-verbal and technologically enabled modes, such as gaze, gestures, facial expressions, and various audio-visual effects, contribute to the expression of impoliteness arising from a sense of superiority over the target, shared with non-targeted viewers. Overall, the paper provides insights into the dynamics of online culture wars and the multimodal manifestations of impoliteness in contemporary social media discourse.


Introduction
The rapid growth of different forms of online communication as well as YouTube's role as a dynamic cultural and social platform have led to the development of reaction videosa genre amalgamating commentary and response to media content.Reaction videos can be argued to form a distinct subculture, with unique norms, values, and practices that prevail among both amateur and professional users (Bliss, 2022;Nicoll and Nansen, 2018;Oh, 2017;Palladino, 2016).The genre has both informational and persuasive functions, moulding viewer attitudes and fostering a sense of shared experience and connection between creators and audiences.For this purpose, creators use a variety of techniques that increase visibility while also allowing them to adhere to community guidelines (Cao, 2021;Hopster, 2021;Sandel and Wang, 2022).Importantly, these strategies extend beyond the mere function of attention-grabbing; rather, they figure relationally within the framework of social interaction and (im)politeness (Garcés-Conejos Blitvich, 2022).They can therefore be interpreted within the paradigm of both technological affordances and social relations: not only do they enhance message appeal but also construe a narrative reflecting creators' standpoint, fostering ideological and political alignment with likeminded audiences (Andersson, 2021;Hopster, 2021).
Nevertheless, despite the abundance of reaction videos on YouTube and the wide range of culturally and socially popular and/or relevant themes they typically cover (the literature singles out music, gaming, online challenges, celebrity appearances, and even pornography; e.g., Bliss, 2022;Oh, 2017;Palladino, 2016), the genre has received limited attention in exploratory studies.Van Zoonen et al. (2010) examined reactions to the anti-Islam video Fitna, viewing them as performances of citizenship where individuals enact political and religious identities.Gonzalo Berrocal et al. (2014) explored responses to a political speech, pursuing the concept of social networks as platforms for 'politainment,' where users both consume and produce political content for entertainment, collectively shaping meaning.However, aside from added subtitles, their dataset lacked modifications in other aspects.The present study aims to address this research gap through a social semiotic discourse analysis of different modalities involved in the process of meaning-making within the format.The objective is to investigate how these modalities contribute to the construction of a cohesive narrative centered on the expression of 'fiendish joy' (after Kant, Stohr, 2022: 170) as well as the establishment of a sense of superiority over the target.The analysis focuses on reaction videos targeting the community of transgender activists on the popular video hosting service TikTok, whose engagement tends to be referred to as 'woke' due to their 'overperforming' and even reactionary stance against social injustice and/or what they perceive as such (Sobande et al., 2022: 6).By examining these videos, the paper will provide insights into current societal conflicts known as 'culture wars,' which emerge from the clash between E-mail address: marta.andersson@engelska.uu.se.(Nagle, 2017;Sobande et al., 2022).
To this end, the concept of 'entertaining impoliteness,' as proposed by Culpeper (2011: 233), will be adopted.While impoliteness has traditionally been associated with negative behaviors aimed at demeaning and harming the target's face (i.e., public self-image associated with situationally claimed identities; Garcés-Conejos Blitvich, 2022) and causing social disruption, it has been suggested that it can also serve as a source of entertainment for onlookers (Culpeper, 2011;Dynel, 2013).In such instances, impoliteness manifests through direct attacks on a (potential) target's face, typically executed in a creative manner, to induce sentiments of superiority and 'vicarious pleasure' (Dynel, 2016: 137) in the viewer not directly targeted.Positive responses to impoliteness commonly arise from disparaging others, which can be amusing for the in-group, as posited by 'the superiority theory of humour' (Dynel, 2013: 111ff and references therein).Entertaining impoliteness thus combines fundamental traits of two other relevant concepts, namely ridicule and mockery, both of which aim at 'poking fun' at the victim by scorning their perceived shortcomings.This fosters a sense of superiority and group affiliation among non-targeted participants (Dynel, 2013;Dynel and Zappavigna, 2023: 3).By integrating these concepts into the analysis, the current paper seeks to demonstrate how multimodal forms of expression serve the purpose of disparaging opposing ideologies in the macro-context of online culture wars.
The study adopts a social semiotic approach developed by Halliday (1978) and Kress and van Leeuwen (2006), which is consisted with previous claims that impoliteness is an inherently multimodal phenomenon (Brown and Prieto, 2017;Culpeper, 2011;Jucker et al., 2023).While impoliteness is known to consistently rely on conventionalized forms (e.g., conventional insult; Culpeper, 2011;Van Olmen et al., 2023), as several studies have shown to date, the verbal element is just one of the plethora of means that can be deployed to cause offense.Prosody, gesture, and gaze are among the other means, as are technologically facilitated strategies such as audio-visual effects (Culpeper, 2005;Garcés-Conejos Blitvich, 2022;Jucker et al., 2023;Rancer et al., 2010).Recognizing the diverse conceptualizations of multimodality across various theoretical frameworks, such as those related to meaning-making practices or semiotic modes and their communicative affordances (Stöckl, 2019), this paper explores the utilization and combination of different semiotic means-including gaze, gestures, facial expressions, and audio-visual effects-in reaction videos produced in the United States in 2022-2023.In the following, Sections 2 and 3 offer scholarly and cultural context, Section 4 reviews methods, and Section 5 discusses findings.

Current socio-cultural background
This section sheds light on the socio-cultural context of the analyzed videos, which is a vital aspect of nuanced interpretation of multimodality employed to disparage political adversaries.According to Nagle (2017), the wave of ideological conflicts characterised by irreverent mockery and anti-political correctness sentiments that followed Donald Trump's election as US president in 2016, has significantly influenced the political sensibilities of an entire generation, resulting in ideological culture wars.These sentiments have been bolstered by the shift in online user culture from passive consumers to active content producers (Gonzalo Berrocal et al., 2014;Nagle, 2017), which has fostered social justice grassroots activism while deepening ideological divisions and polarization (Hopster, 2021;Nagle, 2017;Sobande et al., 2022).TikTok, the platform under investigation, has emerged as a significant locus for activism, notably recognized for rapid dissemination of political and ideological content.Importantly, while its liberalprogressive atmosphere encourages marginalized groups (e.g., the transgender community) to share their experiences, the platform also facilitates the propagation of populist, sensationalist, and divisive narratives to the public (Abbas et al., 2022;Zhao and Abidin, 2023).
Nonetheless, the rise of social justice activism has given prominence to diverse forms of cultural critique, exemplified by the phenomenon of 'wokeness.'Originating from Black American racial activism, wokeness has undergone radicalization in online culture, becoming associated with virtue signalling and controversial 'cancel culture' rather than genuine engagement (Felaco et al., 2022;Geusens et al., 2023;Sobande et al., 2022).This evolution has been reinforced by the adoption of wokeness as a marketable aspect of progressive corporate identity (Sobande et al., 2022), which has incited conservative commentators, particularly in the United States, to challenge transgender activism and its visibility.1 Most notably, the activism of transgender women, concerning their rights in women's sports, access to female-only spaces, and advocacy for interventions such as puberty blockers, has sparked intense debates across various media platforms (Felaco et al., 2022;Sobande et al., 2022).Discussions of the boundaries to freedom of expression related to initiatives like the censorship of classic literature 2 have engaged a range of political perspectives beyond just conservatives (Felaco et al., 2022;Geusens et al., 2023;Sobande et al., 2022).Notably, strategies involving derision, mockery, and ridicule of ideological opponents have gained traction as effective means of communication, resonating across the entire political spectrum (Sobande et al., 2022).
The present investigation examines reaction videos arising from this political climate, wherein the tactic of discrediting opponents involves emphasizing their insufficiencies (Dynel, 2013(Dynel, , 2016) ) and depicting them as irrational, inauthentic, and exaggerated.Crucially, such portrayal stands in contrast to the perceived 'reasonable and logical' image of both the creators and their cohorts (Sobande et al., 2022: 5).Subsequent sections delve into this aspect in greater detail.

Previous studies
Until recently, research into the role of nonverbal communication in conveying and intensifying (im)politeness has primarily focused on prosody (for a review, see Brown and Prieto, 2017), leaving other aspects largely unexplored.One of the first more comprehensive attempts to address this gap is a recent collection of papers discussing multimodal (im)politeness across spoken, written, and signed languages (Jucker et al., 2023).Of relevance to the current focus are Lehmann's (2023) findings regarding the differentiation of multimodal ironic marking (e. g., gaze aversion or raised eyebrows) based on the function of ironic constructions, including impoliteness.Brown et al.'s (2023) study on Korean and Catalan languages, which discovered that gaze aversion and head shaking can be employed for indexing and modulating social distance (common facets of impoliteness), tackles related means of expression.Other studies include Culpeper's (2011) comprehensive analysis of conventionalized nonverbal impoliteness in British culture, including behaviours such as spitting, eye-rolling, and the use of the middle finger.Investigations have also delved into indicators of nonverbal aggression, such as disgusted or contemptuous looks, head shaking, or other paralinguistic cues like a deep sigh (Infante et al., 1988).Lastly, Rancer et al. (2010: 271ff) empirically categorized 51 nonverbal communicative behaviors perceived as highly 'hurtful' by North American undergraduates (crucially, impoliteness has been described as 'hurtful behaviour/communication, ' Culpeper, 2011: 71ff).These behaviors were classified into five categories: aggressive facial expressions (e.g., frowning), eye behaviors (e.g., angry stare), vocalic behaviors (e.g., mocking tone), kinesic behaviors (e.g., slamming doors), on_nike-budlight-backlash/.

M. Andersson
and disconfirming behavior (e.g., walking away).As the authors concluded, these modalities can serve various purposes, including seeking reciprocity, expressing anger, being derogatory, or displaying disdain (Rancer et al., 2010).Many of these objectives are relevant to the current dataset.Nonetheless, investigations of multimodal expressions of impoliteness in online discourse are scarce.Andersson (2023) explores several 'modal affordances' of the written mode (e.g., hashtags or blank spaces) of user commentary employed for impoliteness across social media platforms, demonstrating their efficacy and creative potential for differentiation between online participants.In a comprehensive analysis of 'polyvocal' Tweets combining different voices to ridicule Covid19-related conspiracy theories, Dynel and Zappavigna (2023) refer to the multimodal potential of image/video-text-tag for the expression of mockery.Finally, Garcés-Conejos Blitvich (2022) examines impoliteness in videos from the Instagram page 'Karen's gone crazy.'Having adopted Norris's (2019) basic unit of analysis, that is, a mediated action (i.e., social actions performed with/through mediational means or cultural tools), the paper identifies prosodic elements of anger, such as high pitch and loudness, emphatic deictic gestures (e.g., pointing at the target), aggressive haptics (e.g., hitting), etc.As a result, the study emphasizes the importance of further research into the role of nonverbal modalities in impoliteness expression/perception, with a particular focus on gaze and facial expressions.

Current analysis
The examination of impoliteness through modalities extending beyond language is justifiable for several reasons.First, YouTube creators may encounter social and economic ramifications linked to explicit offense and the breach of social order.Given their dependency on commercial revenues and sponsored partnerships, the outcomes of such events can be detrimental (Tucker, 2018).Non-verbal expressions thus provide a convenient avenue to adhere to social norms while mitigating potential algorithmic repercussions.Second, the distinctive 'visual frame' (Zappavigna, 2019) characteristic of reaction videos, which presents a close-up view of the creator's facial expressions and their upper body, juxtaposed with the reviewed footage, not only renders this genre conducive to analyzing diverse modalities but necessitates their explorationas creators semioticize communicative strategies to convey meaning.Third, within the fiercely competitive landscape of social media, the strategic utilization of one's creative abilities becomes imperative in order to establish a distinct status (Andersson, 2023).Successful creators therefore often push boundaries, experiment with formats, and innovate to engage with their audience and cultivate a devoted and enthusiastic follower base.Achieving this objective invariably entails the utilization of various semiotic modes.
The current study adopts a multi-perspective methodology rooted in the social semiotic framework introduced by Halliday (1978) to investigate the primarily pragmatic phenomenon of impoliteness.Social semiotics posits that language and other semiotic means are organized based on their 'metafunctions'.These include experiential and logical (i.e., ideational) meaning, which structures our understanding of the world; interpersonal meaning, which establishes social relations and attitudes towards events and entities; and textual meaning, which organizes experiential, logical, and interpersonal meanings into coherent texts (Halliday, 1985).Originally applied to language, the metafunctional principle has since been reinterpreted and expanded to encompass visual images (including moving images) and other semiotic modes (Kress, 2009;Kress and van Leeuwen, 2006).From a social semiotics perspective, digital technologies create novel communication channels, where semiotic products such as websites or video clips are considered communicative acts (Kress, 2009;Schmidt and Marx, 2020).Considering the intermodal nature of meaning construction in the context of such acts, a multi-perspective approach is essential (Chovanec, 2019).Therefore, Rancer et al.'s (2010) classification of nonverbal offensive behaviors in North American culture has been deemed relevant for the current purpose.The analysis of gestures will draw on McNeill's (1996) system, distinguishing between concrete/abstract concepts (iconics vs. metaphorics) and deictic/emphatic movements (e.g., pointing).For YouTube's technological affordances, the application of the Visual Grammar framework (Kress and van Leeuwen, 2006) will help explore the visual principles underpinning impoliteness expression in the videos.Further, Relevance Theory (Sperber and Wilson, 1996) will aid in interpreting the collective contribution of various modes to this expression.The integration and interplay of diverse modalities will also be examined in terms of recontextualization, defined as the transfer of discourse elements from one context to another (e.g., Bauman and Briggs, 1990;Gruber, 2019).Finally, the analysis will draw on Culpeper's (2011) concept of impoliteness formulae to disambiguate connotations of the scrutinized semiotic modes, as suggested by Abdel-Raheem (2021).These formulae encompass both conventionalized expressions (e.g., personalized negative references: 'he's a nutjob') and non-conventionalized 'implicational impoliteness', such as sarcasm, which require particularized inferences for interpretation (Culpeper, 2011).

Sampling procedure and data
The following examination of multimodal expressions of impoliteness comprises 18 videos by six American creators posted to the platform between April 2022 and April 2023.The videos, ranging from 06:44 to 16:11 min in duration, focus almost exclusively on transgender activism (other themes excluded).Notably, only clips from male creators are considered, based on the observation that female YouTubers who produce reaction videos do not actively engage with the technologically enabled semiotic modes, which this study intends to explore.Further research could investigate the preferences and motivations of female creators in utilizing different modalities, as well as the influence of gender on the multimodal expression of (im)politeness suggested in the literature (e.g., differences in facial expressions and gestures; Brown and Prieto, 2017).
The dataset comprises videos by two independent creators (who own the channel and produce the content) and four professional creators (who collaborate with production teams).Their fanbases range from 25,000 to 5.7 million followers (as of April 2023).This diversity supports the claim that both professional and amateur YouTube creators employ comparable multimodal techniques (Nicoll and Nansen, 2018).Regarding ethical considerations, the current data are sourced from public discourse, which presently encompasses political or social debates disseminated through various mediums such as newspapers, television, films, radio, music, and web-mediated forums (Marlow, 2017).While public discourse is accessible to researchers, adhering to the Association of Internet Researchers' 2019 guidelines 3 on ethics in Internet research and prioritizing objectivity, the creators discussed in the following are referred to using labels such as Creator1, rather than their actual names or aliases.The images of the reviewed TikTokers have been anonymized using black censor bars.Lastly, the primary criterion to classify the clips as reaction videos was the interaction between the creators and a TikTok footage displayed on the screen complete with their commentary and/or reaction to the content.The videos were found through searches for titles/tags containing the string: 'reaction/woke/ TikTok.'
analysis grounded in social semiotics, the study of signs and symbols in communication, exploring diverse modalities beyond language--including gestures, imagery, auditory elements, and other non-verbal means of communication (Kress and van Leeuwen, 2006).As a first step, the paper undertakes a bottom-up analysis of diverse semiotic means and their interpretations as manifestations of impoliteness.Subsequently, attention turns to the interaction among these resources and their collective contribution to impoliteness expression within the dataset.
The (qualitative) data analysis involved identifying and manually coding strategies employed by creators to convey mockery and impoliteness.Table 1 lists both the semiotic means and their aspects that were annotated.Following Zappavigna's (2019) approach, features were explored until saturation, where no new data contributed to the analysis.Coding involved the examination of both non-verbal and digital audio-visual means of expression, encompassing well-documented features such as gestures, alongside less familiar elements like sound embeddings.This necessitated context-sensitive evaluation for relevance and symbolic interpretation (Kress and van Leeuwen, 2006), as understanding and interpretation of nuanced and potentially overlapping expressions, such as 'angry/scornful stare' (see Table 2), may not be straightforward.However, as discussed in the following (see Section 5.3), the process of meaning decoding operates both within single modes and across modes, thereby enabling contextual evaluation (Stöckl, 2019).For instance, the utilization of a flatulence sound in response to a TikToker's utterance was interpreted in terms of interpersonal dynamics, serving the dual purpose of audience entertainment and mockery of the reviewed footage (Section 5.2).Nonetheless, by drawing upon research on impoliteness perception (Rancer et al., 2010) and incorporating the analyst's own perspective (impoliteness2; Culpeper, 2011), the current study offers an empirically derived picture of previously identified multimodal expressions of impoliteness while introducing novel ones through its findings.

Non-verbal pragmatic units
Table 2 outlines the specific expressions of impoliteness found in the dataset.Categories not identified by Rancer et al. (2010) are highlighted in boldface.

Gesture
Many gestural behaviours found in the data can be categorized as 'emblems,' that is, intentionally performed culturally established symbols with their own meanings (Ekman, 1965;Kendon, 2004).Examples include shaking one's head from side to side or throwing hands up, commonly used to convey disaffiliation and/or contempt (Brown et al., 2023;McNeill, 1996;Rancer et al., 2010).However, in videos intending to establish the creator's superiority through disparagement of the target, even ambivalent gestures may convey impoliteness.For instance, while covering one's face commonly signals dismay, exaggerated face-palming was interpreted as condescension (i.e., 'that is X′, in line with Culpeper; 2011) in the dataset.Similarly, head tilts exhibit multifaceted communicative potential, possibly indicating curiosity or interest, while often deployed to convey irony or sarcasm (Lehmann, 2023).A distinctive feature observed in the dataset is sequences of gestures conveying a narrative ('pantomime';McNeill, 1996: 117), with the creator turning their head away from the reviewed footage while simultaneously covering their forehead or eyes (Fig. 1; top).Such sequences form sets of 'defensive gestures' (Kress and van Leeuwen, 2006: 118), often expressed exaggeratedly for entertainment purposes, indicating a deliberate avoidance of watching and conveying disgust.While disgust is not a typical expression of impoliteness, as an 'other-targeting' emotion (Haidt, 2003: 855), its gestural manifestation can be interpreted as an act of mocking and belittling the target.Similar principle applies to other gestures that figuratively represent the creator's negative emotions.For instance, the gesture of clutching the head to signify a headache (McNeill, 1996) can be interpreted as an expression of pointed criticism (cf.Culpeper, 2011).

Gaze
Analysing gaze presents potential challenges due to the visual frame of the videos, wherein impolite eye behaviour is usually directed at the reviewed footage (and hence the target), except when the creator intentionally turns their gaze towards the audience to ensure comprehension and establish a sense of shared viewpoint.At times, this may manifest as a sequence of gazes interactionally addressing the viewer (Kress and van Leeuwen, 2006).
Nonetheless, in line with Rancer et al. (2010), the current analysis identified behaviours such as angry stare (intense and piercing), bugged out/squinted eyes, and eye rolling.These are often accompanied by an intense face expression (e.g., signalling confusion/disgust; Creator2, Fig. 1, bottom left).Scornful stares, which usually involve raised eyebrows, as well as instances of closing of the eyes/averting the gaze, as if not wanting to watch, seem to be distinctive of the dataset.These behaviours were found to be frequently combined with the gesture of turning the head away from the reviewed content, all of which conveys derision, contempt, anger, etc.Finally, while not mentioned by Rancer et al. (2010) and not precisely measured here, a commonly observed prolonged hard gaze targeting the reviewed footage, was interpreted as signifying sarcasm and/or condescension (Culpeper, 2011).

Facial expressions
According to Rancer et al. (2010), the top four most hurtful facial expressions are: giving a look of disdain (narrowing the eyes), frowning, scowling, and tightened/pursed lips.These behaviours, which were all observed in the current dataset, can be considered emblematic (like gestures), as they possess meaning on their own (e.g., frowning).In some instances, facial emblems are accompanied by a verbal component, such  as clenching one's teeth while uttering the phrase 'Something is really wrong here,' which can be interpreted as Culpeper's (2011) condescension.The dataset also features various grins, such as exaggerated 'wide' grins indicative of mockery, and more subdued expressions conveying sarcasm, like lopsided smiles implying insincerity.Another observed feature is dropping the jaw, which can be associated with mockery and condescension, as well as raised eyebrows, which imply negative assessment (Lehmann, 2023).Naturally, facial expressions may include a variety of combined gestures and gazes, as seen in Fig. 1 (bottom right), which depicts Creator3 closing his eyes, raising his hands, and simultaneously pursing his lips while exhaling loudly (evoking associations with an unpalatable question such as, 'What the hell…?'; cf.Culpeper, 2011).

Digital audio-visual units
Table 3 lists the main technologically enabled audio-visual expressions of impoliteness found in the dataset.

Incorporation of static images
The degree of utilization of static visual elements in the videos varies greatly, ranging from affixing a sarcastic emoji onto the TikTok footage, to more sophisticated techniques such as embedding the Tik-Toker or the creator in a different contextual environment.This is demonstrated in Fig. 2, where the video content is integrated within the context of the popular 'This Is Fine' meme, portraying a dog nonchalantly sipping coffee in a burning room.Considering the primary sense of the meme, wherein an individual asserts that the situation is 'OK' despite indications to the contrary, alongside the situated context of the image-specifically, a conservative creator's reaction to a TikTok addressing the rights of transgender women to utilize women's restrooms-it becomes evident that the visual element foregrounded in the video serves as a sarcastic commentary on this social development (cf.Abdel-Raheem, 2021).

Incorporation of moving images
While the analyzed videos exhibit varying degrees of integration of static images, incorporation of moving images (video clips) prevails  (with Creator2 being less engaged in this aspect).Videos enable creative incorporation of relevant cultural references, including clips from popular cartoons, TV shows, commercials, viral footages, and authentic utterances by politicians, celebrities, journalists, and others.A scene from the animated sitcom King of the Hill (Fig. 3), in which the conservative protagonist states, 'You've just said the stupidest thing ever, you keep talking,' is a pertinent example, as is a clip from the black comedy Bad Santa with the (in)famous line, 'Are you off your fucking meds or something?'Both utterances serve as a creative form of commentary/ reaction to the assertions made by TikTok activists regarding the absence of biological differences between women and men.Authentic utterances include Donald Trump's 'Everything woke turns to shit' quote, which is used as an introductory line in some videos, or a footage with Justin Trudeau's tongue-tied pronouncing of the LGBTQ+ acronym, 4 deployed by some creators in various contexts to mock the targeted community's raison d'être.In each instance, the verbal stratum of such utterances comprises conventionalized explicit impoliteness (Culpeper, 2011); however, their semantic import can be interpreted in terms of recontextualization.Section 5.3 provides more insight into how the individuals quoted serve as vehicles for the creator's expression of  impoliteness, leading to the development of a complex participatory framework in the videos, embracing multiple voices (Gruber, 2019).

Other visual effects
The present dataset encompasses not only the integration of images but also specific visual alterations of the visual frame, such as blurring, shaking, rotating, or inverting the screen orientation.These techniques are more prevalent among professional creators, presumably due to their collaboration with professional production teams.Such effects, leveraging available technologies, aim to highlight the underlying message, which in this case is mockery of the transgender community's values (Kress and van Leeuwen, 2006).Crucially, they can function independently as nonverbal commentary or enhance the verbal component.For example, the top section of Fig. 4 below illustrates the use of a blurred and unstable screen effect as a backdrop for TikTok content, utterly devoid of verbal input.Thus, Creator5's disinterested demeanour and engagement in other activities, serve as the sole form of commentary.According to Culpeper (2011), a lack of interest or response within a social context implies impoliteness.The employed effect of a wobbly screen, emblematic of chaos and confusion, effectively communicates a pervasive sentiment of derision toward the targeted content.In a similar vein, the bottom part of Fig. 4 exemplifies Crea-tor6's act of mocking the TikToker's perspective on the subordinate role of women in the Bible.This goal is accomplished by verbally echoing what the creator claims to be a feminist belief, namely, 'Back then women had absolutely no value,' and by his own depiction as a left-wing propaganda conduit, complete with multiple iterations of his head image and an echo sound effect (discussed below).The involved visual effect can be interpreted as an insinuation of transgender activists' perceived lack of critical evaluation of the ideas they repeat exemplifying implicational impoliteness (Culpeper, 2011).
Finally, advanced techniques such as incorporating the creator/ TikToker within a cartoon clip (Fig. 3, as performed by a nonprofessional creator), animated effects such as the 'demonic head spin' (as a commentary implicating spirit possession inspired by The Exorcist), a liberal protester breathing fire (a metaphor for anger), or digitally enhanced bulging eyes/facial expressions (symbolizing mockery) were also identified in the dataset (prevalent in professionally edited videos).Section 5.3 will cover these strategies in more detail.

Sound effects
Most sound effects found to express impoliteness in the analyzed videos are digitally manipulated alterations to either the reviewed TikToker's speech or the creator's own voice.These include high tones and creaky voice, as well as echo-like effects and slow-motion speech.While many of such strategies align with the markers of impoliteness identified in prosody studies (Brown and Prieto, 2017), their specific purposes vary significantly across contexts.For instance, Creator5 employs an echo sound effect that obscures the TikToker's speech, implying a lack of significance and boredom with the reviewed content (presumably a lack of interest; Culpeper, 2011).In contrast, Creator6 utilizes a deep echo sound for his own voice to mimic a propaganda machine.Another pertinent example is the application of a slow-motion effect to the creator's voice, which typically signifies a sense of boredom, whereas when applied to the TikToker's speech, it becomes a token of mockery.
These observations corroborate the view that no particular prosodic or sound effect can be unequivocally linked to (im)politeness in all contexts (Brown and Prieto, 2017), which suggests that the task of disambiguation of the identified effects is not straightforward.Culpeper (2011) used Relevance Theory, a framework that stresses the optimization of cognitive affects in comprehension (Sperber and Wilson, 1996), to analyze various non-conventionalized impolite statements (including mimicry).According to the relevance principle, subtle cues, such as strategically incorporated music snippets, can grab viewers' attention and guide comprehension.In the genre under investigation, interpreting such cues as mockery or ridicule reduces cognitive effort while offering cognitive benefits and cultural acknowledgment.For example, the incorporation of music files or videos, a seldom-utilized tactic attributable to copyright restrictions on YouTube, typically entails the muting of the original audio content coupled with the introduction of supplementary material, such as Julius Fučík's composition, Entry of the Gladiators.As a cultural reference known from circuses, this selection aligns with the goal of disparaging the target while amusing the non-targeted audience.
Having said that, the creators commonly employ sounds that can be viewed as 'emblems' with their own meanings (recall emblematic gestures; Section 5.1).Some examples include screams of horror; the synthetic 'error' sound deployed as ironic commentary; or flatulence sound effects, either in isolation, or as a replacement for the word shit within the explicit expression of pointed criticism (i.e., 'This is shit; ' Culpeper, 2011).These strategies, like many others mentioned above, commonly contribute to the effective interplay of various semiotic means in impoliteness expression, which will be discussed in the next section.

Interplay of semiotic means
The remainder of the paper will focus on the integration of various semiotic modes, their cumulative rhetorical impact, and their collective influence on meaning (for a discussion see Kress and van Leeuwen, 2006), the core idea of multimodality (Stöckl, 2019).This step is essential due to the 'modal complexity' (i.e., interrelationship of modes; Stöckl, 2019: 52) of the dataset and the nuanced nature of impoliteness, which requires assessing numerous subtle cues comprehensively.An illustrative example is gestures, which often comprise a wide range of expressive movements, extending beyond hand and arm involvement (McNeill, 1996).This is seen in Fig. 1 above, where the creators were captured to engage in complex sequences of hand, face, and eye movements to express mockery, disgust, and contempt.Similarly, the integration of diverse digital effects/resources will lack coherence unless a meaningful cultural connection can be established between the original sense/source and the new, recontextualized meaning in context (Bülow and Johann, 2023).Needless to say, digital and paralinguistic resources can effectively combine to generate meaning, frequently on par with verbal input, corroborating the need for holistic interpretation.
A pertinent example is the memetic scenario in Fig. 2 (Section 5.2), where Creator4 does not engage in gaze interaction with the audience, indicating that the image 'offers' information for the viewer to interpret, in accordance with Visual Grammar principles (Kress and van Leeuwen, 2006).The ideational component of the scenario is the state of affairs involving a male-to-female transgender activist discussing problems using female bathrooms, conveyed through the meme template with an interpersonal appeal, as evidenced by the way the graphic dominates the frame -it is large, prominently positioned, vibrant in colour, and compositionally salient (Kress and van Leeuwen, 2006;O'Halloran et al., 2017).This strategy not only captures the viewer's attention but also results in an interesting congruence with the verbal stratum of the video, as Creator4 utters the phrase 'This is fine,' exemplifying sarcasm in this context (Culpeper, 2011).Indicative of sarcasm/mockery is also the low pitch of the utterance (Brown et al., 2023 and references therein) and the creator's facial expressionthe restrained grin (different from a broad smile with a full mouth extension) and the slightly raised eyebrows jointly contribute to a sense of sarcasm.Impoliteness therefore arises from the 'mutual elaboration' of the modes (Stöckl, 2019: 51).Another relevant example is Creator6 emulating the critique of the chauvinistic portrayal of women in the Bible by transgender community (Fig. 4, bottom).The utterance carries a plausible ideational element given that feminist activists are often critical in this regard.However, as the creator adopts a voice that echoes somebody else's voice (if only a hypothetical one), aiming to condemn and mock it, this act gets an interpersonal interpretation as an instance of 'dissociative echoing' (Dynel, 2021;Dynel and Zappavigna, 2023).Contributing to this interpretation is the aforementioned synthetic modulation of the creator's voice, yielding a resemblance to a communist propaganda machine.This effect deviates from the voice quality of the imitated TikToker.Complete with the visual 'echoing' effect on the creator's face discussed earlier, along with the prolonged gaze directed into the camera and the accompanying frown, the snippet exemplifies an ingenious way of lampooning political adversaries, capitalizing on both the individual strength of each modality and their mutual elaboration in the process of meaning construction.It is worth noting that recontextualized elements lose their original illocutionary and perlocutionary effects (Gruber, 2019), and thus both the inclusion of circus music alongside the reviewed footage of a transgender activist testing makeup products (Fig. 5, top right) and the reimagining of the now-memetic Jessica Starr's reaction to Donald Trump's inauguration as a fire-breathing dragon (Fig. 5, bottom right), serve the purpose of disparaging the target, embodying the essence of mockery (Culpeper, 2011;Dynel, 2013;Dynel and Zappavigna, 2023).While recontextualizations may involve an intertextual orchestration of different voices aligned toward a common objective (Dynel and Zappavigna, 2023), such as Donald Trump's derogatory comment on woke culture or the aforementioned snippet from The King of the Hill where the conservative protagonist directs an offensive remark at the embedded TikToker (Fig. 5, top left), the intriguing aspect of the current dataset is what can be termed as 'derisive appropriation' of resources.In such cases, the incorporated element undergoes a recontextualization that not only evokes shifts in meaning and interpretation (as commonly observed in this process; Gruber, 2019), but also serves as a powerful means designed to undermine and belittle both the targeted individual(s) and their associated group.An illustrative example is the aforementioned footage of Justin Trudeau's mispronunciation of the LGBTQ+ acronym in a speech intended to demonstrate solidarity with the community.However, as the response video intends to mock the legitimacy of the community's struggle, the expression of solidarity is entirely overshadowed once the footage is utilized to diminish the sense of the very symbol of inclusivity and diversity.Similarly, a reviewed TikToker's choice to wear a glittery facemask, resulting in a more restrained speech style, is recontextualized to stifle the creator's voice (Fig. 5, bottom left), symbolizing the intention of silencing the transgender community.
These observations highlight the intricate collective contribution of different semiotic modes to the multimodal semantics of impoliteness in the dataset and their rhetorical impact.Previous research has demonstrated that multimodal recontextualizations, such as sarcastic images, are more persuasive and memorable than language alone, fostering deeper reflections on messages and improved memorization (Bülow and Johann, 2023).Likewise, the recurrent recontextualization of diverse audio-visual effects in the present dataset can be seen as enhancing the videos' impact and visibility, aiming to assert superiority over targeted groups while appealing to the creators' own communities.As illustrated in Fig. 5, to achieve this goal, YouTubers often leverage their internet-savvy and cultural awareness in a creative form discussed elsewhere as 'e-mpoliteness' (Andersson, 2023).The outcome frequently observed is the said 'derisive appropriation' of diverse semiotic modes that hold symbolic relevance to the target of impoliteness.Interestingly, this outcome embodies a nuanced participatory framework involving multiple entities, a framework that may be elucidated by reconceptualizing Goffman's (1981: 144) idea of 'speaker's roles' in spontaneous spoken interactions (see also Dynel, 2011).These roles involve the 'principal,' whose ideas are communicated, the 'author,' who formulates the words conveying those ideas, and the 'animator,' who utters the words.In digital communication, roles are redefined, allowing the author to take on responsibilities such as writing, creating videos, and designing images.In the investigated dataset, creators consistently embody the principal's viewpoint.Some recontextualizations, like Trump's slur on woke culture, align the author/animator's role with the principal's perspective.However, incorporating Trudeau's mistake in a video ridiculing his own political stance reduces him to a mere 'sounding box' (Goffman, 1981: 144), devoid of responsibility for the expressed sentiments (cf.Messerli, 2017 on fiction).As a result, the mispronounced LGBTQ+ acronym corroborates the unfavourable portrayal of the community intended in the video, resulting in Trudeau's transformation into an 'unwitting mouthpiece' for the creator's viewpoint.Such complex amalgamation of multifarious semiotic modes allows the creators to position themselves vis-à-vis both the audience and the target, effectively communicating an ideationally plausible message that resonates on interpersonal and compositional levels, all while contributing to a cohesive narrative of mockery and impoliteness.

Concluding remarks
This study explored how different modes of non-verbal and digitally mediated communication convey entertaining impoliteness in YouTube reaction videos, focusing on both their individual characteristics and complementary contribution to this expression.As observed in the dataset, all creators engage in expressing impoliteness through a range of multimodal means, such as gaze, gestures, facial expressions, and audio-visual effects.Besides previously established behaviors, such as head shaking, angry stares, or frowning (Rancer et al., 2010), the study identified acts specific to the analyzed material.These include turning the head away from the footage while simultaneously covering the forehead, prolonged hard gaze, and sarcastic grinsamong others.Additionally, the creators were found to utilize the technological affordances of the platform, resulting in both creative exploitation of popular internet tropes (e.g., memes) and the use of innovative strategies such as visual frame alterations.However, the overarching observation from this part of the analysis was that while these diverse modes can be analyzed 'from below' (Stöckl, 2019: 52), a comprehensive understanding of the multimodal complexity of the videos is achieved through the study of their combination and mutual elaboration.A relevant example is the comprehensively discussed 'This Is Fine' memetic scenario in Fig. 2, where cues as subtle as the creator's grin and gaze, through the intricate recontextualization of relevant cultural references, collectively function as a conduit for his viewpoint, enriching and elaborating the intended expression of disparagement and sarcasm.While the presence of such diverse cues across the dataset implies at least a certain degree of conventionalization within communication and culture (Brown et al., 2023), the analysis of their interplay suggests that they are interpretable as indicators of mockery and/or impoliteness based on their relevance in the cultural and social context at hand (Sperber and Wilson, 1996).
This observation aligns with Stöckl's (2019) argument that genre knowledge plays a crucial role in interpreting multimodal relations and establishing connections between modes.Therefore, while one could argue that the observed utilization of different semiotic modes has been partially driven by the creators' adherence to YouTube's community standards, which may lead to the avoidance of explicit verbal impoliteness, multimodal means of expression can be viewed as an inherent component of the reaction video genre convention.In this context, such means function as a vehicle for ingenious content creation, thereby enhancing entertainment for the audience.Complete with the focus on the radical faction of woke culture, which may contribute to heightened hostility towards the identity of a larger, certainly heterogeneous group, the videos convey a coherent narrative aimed at eliciting feelings of superiority while appealing to the sense of 'vicarious pleasure' among the creators' cohorts (Dynel, 2016: 137).As a result, the examined dataset illustrates a specific convention: intrinsically critical and frequently hostile reactions to opposing political viewpoints, which generally adhere to official community standards while employing a variety of semiotic modes to increase appeal derived from denigrating outgroups.The audience's genre knowledge will then entail at least some level of awareness (and, presumably, expectation) of both the presence of impoliteness and the modes of its construal in the videos.
Importantly, these observations shed light on the nature of multimodal impoliteness, which can convey nuances not captured through words alone, attract attention, or even creatively amplify the message (Andersson, 2023;Knoblock, 2020).The paper therefore adds to our understanding of the multifaceted and nuanced nature of the phenomenon, which has previously been studied within the functional framework (Andersson, 2022;Khosravi and Babaii, 2017), effectively uncovering the role of specific contextual determinants in the construction of social meaning.Likewise, the present study not only provides insights into the practices of expressing impoliteness in YouTube reaction videos (or at least within their specific subgenre), but also offers a comprehensive picture of the cultural, social, and ideological dimensions of the discourse at play.This is particularly important in the context of social media, which are known for disseminating and amplifying cultural and social trends.Semiotic discourse analysis enables the discovery and evaluation of these trends, shedding light on how they gain traction in online communities.Therefore, future research should further investigate different modalities available on YouTube and how they interact to convey impoliteness.

Declaration of Generative AI and AI-assisted technologies in the writing process
During the preparation of this work the author(s) used QuillBot and Trinka in order to improve language correctness and readability.After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. The visual effect of a meme in Creator4's video.

Fig. 5
below showcases further examples of the interplay of various semiotic means focusing on their recontextualized meaning in context.

Table 1
Semiotic means and their aspects annotated in the analysis.

Table 2
Non-verbal behaviours in the dataset.

Table 3
Audio-visual expressions of impoliteness in the dataset.