Linguistic analysis of online conflicts: A case study of flaming in the Smokahontas comment thread on YouTube

Lotta Lehti
PhD, Senior Lecturer
School of Languages and Translation Studies
University of Turku

Johanna Isosävi
PhD, Senior Lecturer
School of Languages and Translation Studies
University of Turku

Veronika Laippala
Post-doctoral researcher
Turku Institute of Advanced Studies / School of Languages and Translation Studies
University of Turku

Matti Luotolahti
MSc (Tech.), Doctoral Candidate
Department of Information Technology
University of Turku

Tulostettava PDF-versio

We describe how online conflicts can be analysed with linguistic tools. The case under scrutiny is the ‘massive polylogue’ on YouTube (i.e. the multilingual and global comment thread following the video ‘What languages sound like to foreigners’ by Smokahontas). The comment thread displays themes and conflicts that re-emerge with surprising frequency. Our investigation shows that a linguistic study of online conflicts can reveal important aspects of communication patterns and group dynamics.


The advent of participative Internet has created a shared concern of the downsides of everyone being able to express themselves in public without moderators. One reason for this concern is the (seemingly) increasing number of verbal conflicts. For instance, school bullying has reached an unprecedented level – verbal or visual bullying takes place in digital, public environments; thus, the bullied pupils cannot escape the harm even when at home or by changing schools. Another current issue is the hate speech targeted either at social groups or at individual public figures. Overall, in our era of a ‘demotic turn’ (i.e. increased participation of ordinary people in the media and the production of culture), identities are more and more constructed online (Turner 2010, Dervin & Abbas 2009). In this process of identity construction, topics conceived as ‘meaningless’ and ‘trivial’ are now dealt with in public (Eronen 2015, 31; Turner 2004, 85). As Eronen (2015) shows in her study on online celebrity gossip, the expression of both sameness with and difference from others often creates conflicts that become emotional, aggressive and even hostile.

The purpose of this article is to present, to a readership in culture studies, some recent developments in the linguistic analysis of conflicts in digital language use. In addition, we hope to enhance the dialogue between researchers from different fields who are studying digital language use, communication and culture. In order to concretise linguistic notions in the study of (online) conflict, we analyse a specific case: user-generated comments posted on the viral YouTube video ‘What languages sound like to foreigners’ by Smokahontas. Previous research on different topics of YouTube videos and their comments has shown that the communication phenomena on this participatory website are complex and that interaction often becomes hostile and polarised (e.g. Walther & Yang 2012, 2; Bou-Franch & Garcés-Conejos Blitvich 2014, 20).

This paper explores what kind of verbal conflicts can be found in a YouTube comment thread dealing with a ‘trivial’ topic. We particularly explain how the notion of flaming (Danet 2013; Turnage 2007; Kayany 1998) is manifested in the material. We also briefly discuss what kind of group dynamics are reflected by the verbal conflicts in this type of online discussion.

The remainder of the article is organised as follows. We begin by presenting the Smokahontas video and YouTube as a communication medium. We then move on to the definition of (online) verbal conflict. Our analysis is organised according to three main elements of flaming: the structure of conflicts in the material, the impoliteness in these conflicts and playful language use in the conflicts. The article ends with some brief concluding remarks.

Smokahontas and YouTube

Video 1. “What Languages Sound Like to Foreigners” by Smokahontas.

As mentioned, the material under scrutiny in this paper is a comment thread from YouTube. The video that the writers are commenting on is the YouTube video presented above. Then 19-year-old Sara Maria Forsberg posted this video, in which she imitates the phonological and prosodic features of different languages but speaks gibberish instead of uttering real words of the languages in question. The video was posted on March 3, 2014, and by September 25, 2015, the video had over 15 million views and had received 28,922 comments. In this study, we concentrate on 1,402 comments. The comment thread is highly multilingual, but we only analyse comments that are mainly written in English.

The reason for examining this particular comment thread is its viral and global nature. The video shares features with most viral videos (see Guadagne et al. 2013); it is humorous and emotionally arousing. Many people viewing the video for the first time were mesmerised by the talent of the creator in imitating the phonology of different languages. It can be assumed that the video was shared via social media mainly in order to delight and entertain others. For some people, however, the main emotion evoked by the video was irritation. We assume that this is because one’s native language is an important dimension of identity (see Pascual y Cabo & Rothman 2012) and, as the conflicts analysed below indicate, many perceived Smokahontas’ imitation of their native language as offensive. Whatever the reasons, the video quickly became extremely popular; consequently, the comment thread is multilingual and the nationalities of the participants appear to be numerous. However, the purpose of the video is entertainment; therefore, our results will not be directly generalisable to, for example, discussions concerning ideological and political themes. Instead, the underlying topics generating discussion in this comment thread can be considered as ‘meaningless’ or ‘trivial’.

Dynel (2014, 37–52) characterises YouTube communication overall by ‘speakers’ and hearers’ spatial and temporal separation and infinite numbers of potential participants at the reception end, who are typically unfamiliar with one another’. Dynel (2014) distinguishes three levels of communication on YouTube: speakers and hearers in video interaction, senders and recipients of a video and writers and readers of comments. The active participants in the comment thread (i.e. those who write comments) are registered users (most often displaying a nickname), while passive, non-registered users can only read the comments. The present paper concentrates on the third level of communication, the comment thread, and only takes into account the comments written by active participants, thus leaving aside, for example, a reception study conducted via questionnaire.

Due to the infinite number of participants, Bou-Franch and Garcés-Conejos Blitvich (2014) and Bou-Franch et al. (2012) label YouTube comment threads as ‘massive polylogues’ (see also Marcoccia 2004, 144). The comment thread remains on the site for as long as the commented upon video clip remains, thus enabling asynchronous interaction that may continue for long periods of time. This allows, as Bou-Franch and Garcés-Conejos Blitvich (2014, 21) point out, ‘unprecedented access to researchers to the diachronic unfolding of conflict’. However, the massive polylogue also poses a challenge to methods of analysis. For instance, the participant structure of the exchange is extremely complex – commenters join in and drop out whenever they like and can either ‘post a new comment’ or ‘respond to a comment’. Sometimes, the first option is used even if the posted comment is a response to a previous one. Consequently, identifying which comments belong together is not a straightforward task. The analysis is further complicated by the fact that new comments appear at the top of the thread, while responses to comments appear underneath the posts. Bou-Franch and Garcés-Conejos Blitvich (2014, 21), however, show in their survey that this partly reverse chronological order of comments does not appear to trouble YouTube users.

Online verbal conflicts

Verbal conflict is, naturally, not a phenomenon native to the Internet. Fracchiolla (2013) studies verbal conflict from an ethological and anthropological perspective and highlights the perpetual nature of verbal conflicts between humans. She makes an analogy between the physical violence of animals in the competition for the same ecological niche and the verbal conflict of humans in the competition for symbolic territory (i.e. who has ‘the last word’ and who gains power). She also distinguishes verbal violence from verbal aggression. The former always includes the latter – verbal violence is, according to Fracchiolla (2013), aggressive language use that lacks meaning. In other words, verbal violence features domination of one person toward the other instead of an interaction (even if aggressive) between these two (ibid.). This distinction between aggressive language use and verbal violence is not established in the discourse analytical terminology of verbal conflicts. Moïse (2012, 2) defines verbal violence as a global process in communication surrounding an increasing tension in interaction caused by power relations, categorisations, identities and social actions.

Aggressive, violent and impolite language use all imply a conflict or disagreement, i.e. discrepancy of ideas, opinions or actions between interlocutors (Smith & Mackie 2000, 503, see also Langlotz & Locher 2012). In other words, disagreements are, according to Langlotz and Locher (2012, 1591), speech events in which ‘interlocutors judge the behavior of their communicative partner(s) in order to manage their social positions’. Disagreements range from supportive to highly oppositional – aggressive, violent and impolite language use pertain to the latter end of the disagreement scale (see Angouri & Locher 2012). In the study of this continuum, the concept of ‘face’, initially theorised by Goffman (1959), is crucial. Spencer-Oatey (2007, 644) associates face with affective sensitivity – as something vulnerable and resulting in emotional reactions. The violations of the face of the other are conceived as impoliteness.

In digital environments, the timeless phenomenon of verbal conflict takes new forms, some of which users and/or researchers have identified and labelled. One of these is flaming, a notion that Danet (2013, 639) describes as ‘a sequence of typed, synchronous or asynchronous, online exchanges involving sudden, intense conflict’. For Danet, flaming is thus a phenomenon native to digital environments. Moor et al. (2010), in contrast, do not consider flaming to be an exclusively digital phenomenon, although they claim it to be more common in computer-mediated communication than in face-to-face interaction. Indeed, their questionnaire study indicates that YouTube users perceive flaming as a common practice in YouTube comments. They consider it to be an irritating feature of YouTube communication to the extent that the fear of flaming is a reason for refraining from uploading videos. However, some users perceive conflicts as positive and amusing (Moor et al. 2010, Pagliai 2010, see also Bou-Franch & Garcés-Conejos Blitvich 2014, 20).

The main reason for the strong presence of flaming on YouTube is often considered to be the anonymity of and distance between the participants, resulting in a setting in which participants are less aware of others’ feelings and in which it is considered ‘safe’ to hurt others (see e.g. Bou-Franch & Garcés-Conejos Blitvich 2014, 20, Moor et al. 2010, Garcés-Conejos Blitvich 2009 & Pagliai 2010). However, Spears and Postmes (2015, 26) offer another explanation. According to them, the effect of anonymity – or deindividuation – is not one of increased individuality but of increased social regulation. In other words, in an anonymous online context, participants act according to the implicit norms of the interaction more carefully than in face-to-face settings (Spears & Postmes 2015, see also Moor et al. 2010, 1537, Kiesler et al. 1984, Diener 1977). These implicit norms of YouTube comment threads appear to allow and encourage aggressive language use and flaming. Anonymity and deindividuation also increase polarisation in online contexts: strong adherence to group norms reinforces the juxtaposition between different groups.

Flaming, as described above, is a sub-category of verbal conflict. In addition to being a ‘sudden, intense conflict’, it is characterised by the target of flaming. According to Kayany (1998, 1135), flaming is expressed at another person, whereas criticism is targeted at ideas and opinions. Another neighbouring notion is trolling, defined by Hardaker (2010, 237) as a user who ‘constructs the identity of sincerely wishing to be part of the group in question, including professing, or conveying pseudo-sincere intentions, but whose real intention(s) is/are to cause disruption and/or to trigger or exacerbate conflict for the purposes of their own amusement’. At the core of trolling is the active, malicious behaviour and insincerity of the troller. In cases of flaming, it is sometimes difficult for the reader to interpret whether a given participant is trolling (i.e. being insincere). We will present one such comment below in example (4).

In addition to the target, flaming is characterised by the possibility of having ‘expressive, playful, sporting aspects, not only aspects relating to “the facts”’ (Danet 2013, Vrooman 2002). In addition, Danet (2013) underlines that flaming is not entirely a negative phenomenon – it can also be used to express solidarity. Overall, Danet (ibid.) views flaming as an impolite conflict. In what follows, we will explore the manifestation of flaming in our material. We will specifically concentrate on three features: the structure of the verbal conflict, the nature of impoliteness and playful language use.

Flaming in the Smokahontas comment thread

Conflict structure

Content-wise, conflicts in the Smokahontas comment thread deal with the same theme with surprising frequency: commenters not understanding the core idea of the video, i.e. that Smokahontas is speaking gibberish instead of real languages. Examples (1) and (2) illustrate this.


[Thunayan Almajed]:

aha yo can’t speak Arabic like the other languages

sorry to say that but I had to because I am arabic and I did not know what you said !

try too work on it more ok

[Illyasviel von Einzbern]:



[Julianna Fonseca]:

I’m Brazilian, I speak portuguese and I didn’t understood a thing you said at the portuguese part

[DJ Lufin]:

@ [Julianna Fonseca] THAT’S THE FUCKING MEANING OF THIS GODDAMN VIDEO! She was talking gibberish, using the local “accent”. So, it sounds like she’s speaking the language, but she is really just imitating the accent.

In examples (1) and (2), the response to the first comment contains aggressive language use: capital letters and exclamation mark(s) signalling shouting and taboo words (‘fucking’, ‘goddamn’). This is not the case in all of the responses to the ‘I do not understand’ theme. Some of the responses feature a rather objective, neutral tone in order to inform the previous commenter of the core idea of the video. This type of response is illustrated in the comment response pair in example (3):


[omarmustafa khalil96]:

I’m arabi but i don’t understand what she said

[Lez Sky]:

You’re not supposed to. That’s what the language sounds like to foreigners, or people who don’t speak that language.

As stated above, our analysis concentrates on the interaction between participants in the comment thread; however, it is worth mentioning that some of the initial expressions of non-understanding signal a judging stance, even aggressiveness, toward the video and Smokahontas. In most cases, this initial comment is rather neutral, as in examples (1), (2) and (3).

Many conflicts on the misunderstanding theme consist only of two components, as the second turn receives no reply. Examples (1), (2) and (3) illustrate this – the second comment in all three examples did not receive a reply. In these cases, the only connection between the participants is the second one reading and commenting upon the first one. The first commenter might never access the comment thread again and is thus possibly not even aware of someone expressing disagreement toward his or her comment.

According to Danet (2013, 646), conflict consists of three components – ‘an initial statement or other speech act by speaker A; a challenge by Speaker B expressing disagreement with Speaker A; and a response by Speaker A expressing disagreement with Speaker B’ (see also Avgerinakou 2003, 279; Schiffrin 1984, 316). A conflict is thus triggered by a ‘challenge’ (Danet 2013) or an ‘offensive event’ (Jay 1992, see also Bou-Franch & Garcés-Conejos Blitvich 2014, 23) (i.e. an impolite expression of discrepancy, which in turn is likely to receive a conflictual response). An offensive event is illustrated in the comment by [Illyasviel von Einzbern] in example (1) and in the comment by [Dj Lufin] in example (2); this is in contrast to the comment by [Lez Sky] in example (3). The most important difference between the first two and the latter example lies in the degree of politeness of the response. Indeed, the response of [Lez Sky] in example (3) is an explanation instead of an offensive event. The analysis of impoliteness in language use thus appears to be critical in the analysis of conflicts. We will account for this inherent element of flaming in the next section, which deals with impolite language use. As mentioned, many of the conflicts in our material end at the offensive event. We label the discrepancies in our material consisting of only two comments as fractional conflicts.

A quantitative study on the number of comments left by each user on a sample[1] of comments from the Smokahontas comment thread suggests that if there are conflicts in the material, the large majority of them are fractional. In the case of a definite majority – 1,309 of the 1,402 comments examined for this count – the users left only one single comment. A total of 62 users posted two comments and seven posted three. Three individuals posted more comments: four, six and eleven, respectively.

However, not all of the conflicts are fractional, even if the initial speakers do not return to the comment thread after posting their comment. We argue that for a conflict to unfold it is not necessary for the initial speaker to respond to the offensive event – it can be others who pursue the conflictual exchange. Indeed, according to Dobs and Garcés-Conejos Blitvich (2013), in polylogal communication, it is often a witness that responds to an impolite act instead of the initial speaker to whom the impolite utterance was directed.

Example (4) represents a sequence consisting of an initial comment and 13 responses. There are eight participants in the sequence. Because of its length, we present example (4) in five parts (A–E). In this particular sequence, the initial comment is already impolite because it contains a taboo word (‘fucking’) and explicitly associates the other with a negative aspect (‘prude’). However, we consider the following comment by [yålå swag] to be the actual offensive event. The third comment by [Eric Hossner] is an example of a reaction of a witness to the offensive event: [Eric Hossner] disagrees with [yålå swag] in a very offensive, ironic way.



gosh you Europeans try to call Americans stupid but you don’t understand that she is talking in gibberish? Fucking prudes

[yålå swag]:

We don’t “try” to call u stupid, we ARE calling u stupid.

[Eric Hossner]:

@ [yålå swag] So, because I happen to have been born in a particular country, I am automatically stupid? That makes absolutely no sense. I’m sorry I was born in this country. I’m sorry I’m white. I’m sorry I’m a man. Please mock and ridicule me more. Call me stupid, ignorant, fat, racist, and uncultured. I deserve it because it just so happens that I was born in the United States of America.

Bou-Franch and Garcés-Conejos Blitvich (2014: 29) distinguish ‘attacks directed at a social group’ from ‘attacks against a specific participant’. These two types of attacks are mixed in the sequence in example (4). For instance, in (4A), [Eric Hossner] responds to the attack by [yålå swag] directed at Americans in general as if the attack was directed at him personally. The importance of nationality to an individual’s identity, triggered by the imitation of languages in the video, is thus the core of the conflict in the sequence. In (4B), [TheXtraMan99] presents an opinion that is not straightforward to interpret because of the quotation marks and the ellipses at the end. It could, however, be interpreted as a racist and offensive opinion – claiming those of one nationality to be less intelligent than others. One alternative is to interpret the comment as trolling (i.e. as insincere), but the comment in (4D) by the same commenter suggests that the writer is not presenting fake opinions.



ummm Idk what were talking about here but, americans really are “less intelligent” than most other people…

The following comments, presented in example (4C), are reactions to [TheXtraMan99]. The first one to react, [Sneakysneaky88], labels the comment as hypocritical. The second, [OurSolemnHour13], aligns to some extent with the categorisation of [TheXtraMan99] by stressing the importance of nationalities.



@ [TheXtraMan99] You do realise how hypocritical this comment is, right? :P


@ [Sneakysneaky88] Actually it’s funny I’m an American but I like to identify as Polish cause that’s where my paternal line comes from. I want to move to Europe in the future. I still get pissed when Europeans mock me even though I’m probably more European than most Europeans. I do grudgingly admit I do kinda love America though. You won’t understand unless you’re born here.

The third reaction to the comment of [TheXtraMan99] is presented in (4D). [Eric Hossner] expresses his disagreement by aggravating the categorisation and requesting confirmation.


[Eric Hossner]:

@ [TheXtraMan99] So every single American is less intelligent than every single person of every single other country?


@[Sneakysneaky88]…especially because he misplaced the comma and forgot an apostrophe. He must be some ignorant 11-year-old who was indoctrinated by stuck-up parents.


@ [akpokemon] No, I’m not a 11 year old and my parents aren’t stuck up and why do you expect all people to know English perfectly?! It’s not my native language and yes it is useful but you can’t expect everyone to speak and write/type English perfectly… kinda dumb of you, huh? I rest my case…

In (4D), the conflict continues as [akpokemon] indirectly criticises [Sneakysneaky88] for posting a patronising comment (see [4C]). In an ironic manner, [akpokemon] accuses [Sneakysneaky88] of judging others. [TheXtraMan99] does not understand the irony but responds to the comment by [akpokemon] as if he had truly been labelled as lacking intelligence because of spelling errors. [TheXtraMan99] posts no more comments. What follows is a sequence featuring a submission/assent (see Bou-Franch & Garcés-Conejos Blitvich 2014: 29) in which [akpokemon] and [Eric Hossner] acknowledge each other’s claims and admit their own faults in communication.



@ [TheXtraMan99] Okay, I apologize for making baseless assumptions, but you seem to be doing the exact same thing. So looks like we’re both in the wrong here. […]

[Eric Hossner]:

@ [akpokemon] I think all the points you made are right. But why do you have to sound so “uppity” and shit? No one’s impressed by how you type. Just as long as you use proper grammar.


@ [Eric Hossner] Yeah that’s a very valid point. I was pretty worried about coming off that way while writing it, but after I finished it, I didn’t want to go through it all and dumb it down just to look more humble. I mean, how stupid would THAT be?! But yes, the way I worded it definitely makes me sound stuck-up myself. But it’s the way my ideas flowed…and it gets the point across the way I want it to, and that’s my main concern.

[Eric Hossner]:

@ [akpokemon ] Oh, I’m not mad about it. It’s not that big of a deal. Your points are what matter. I’m just trying to make sure you know you did it. People write that way too much and it pisses me off because most of those people are assholes.


Ameritard detected…

The comment sequence (4) ends with a comment posted by a new participant [antiregayton1], the content of which is basically the same as in the comment by [TheXtraMan99] in example (4B). None of the other participants respond; thus, we can claim that the others have withdrawn from the conflict. In Bou-Franch and Garcés-Conejos Blitvich’s study (2014: 31), withdrawal was the most common type of terminal exchange.

On the basis of the analysis presented, we claim that Danet’s (2013) conception of conflict as consisting of three parts and most often of only two participants does not correspond to the reality of massive polylogues in fora such as YouTube. Also, Bou-Franch and Garcés-Conejos Blitvich (2014: 33) problematise the traditional three stages of conflict (beginning, middle and end) because of the ‘multi-sequentiality of conflict in polylogal, mediated interaction’, and perceive the YouTube polylogues as ‘undergoing a continuous middle’: conflicts are neither dyadic nor linear. They are not always discursively resolved, like a face-to-face conflict. In the sequence presented in example (4), the conflict is partly resolved: [Eric Hossner] and [akpokemon] appear to find some degree of consensus; however, the final comment by [antiregayton1] leaves the conflict open.

The main theme – national stereotypes and racism – of the sequence in example (4) is a frequent theme in this material. In addition to the participants commenting on each other’s national characteristics and the accuracy of those stereotypes, many commenters claim that Smokahontas is racist because of the way she represents the speakers of different languages. Many commenters stress especially Smokahontas’ naming the Italian language ‘pizza’ as racist. Further, many consider Smokahontas’ imitation of Portuguese to be unsuccessful, and her perception of Portuguese elicits numerous comments dealing with racism.


As mentioned, the degree of politeness of the expression of disagreement is crucial in the emergence of conflict and flaming. Danet (2013) considers impoliteness to be a component of flaming. Culpeper (2013, 5) argues that “situated behaviours are viewed negatively – considered ‘impolite’ – when they conflict with how one expects them to be, how one wants them to be and/or how one thinks they ought to be”.

According to Brown and Levinson (1987, 61), positive face is the desire for others to respect an individual’s wants and needs, whereas negative face is the desire for ‘freedom of action’ (see Culpeper 1996). Culpeper, Bousfield and Wichmann (2003) categorise impoliteness into five categories: (1) bald on record impoliteness, where there is much face at stake and where there is an intention on the part of the speaker to attack the face of the hearer; (2) positive impoliteness, which uses strategies designed to damage the addressee’s positive face wants (‘ignore, snub the other’, ‘exclude the other from the activity’, ‘disassociate from the other’, ‘be disinterested, unconcerned, unsympathetic’, ‘use inappropriate identity markers’, ‘use obscure or secretive language’, ‘seek disagreement’, ‘make the other feel uncomfortable, e.g. do not avoid silence, joke, or use small talk’ ‘use taboo words’, ‘call the other names’); (3) negative impoliteness, which uses strategies designed to damage the addressee’s negative face wants (‘frighten’, ‘condescend, scorn, or ridicule’, ‘invade the other’s space’, ‘explicitly associate the other with a negative aspect’[2], ‘put the other’s indebtedness on record’, ‘hinder or block the other – physically or linguistically’, etc.); (4) sarcasm or mock impoliteness, which uses politeness strategies that are obviously insincere and thus remain surface realisations (these are clearly the opposite of banter, in other words, mock impoliteness for social harmony); and (5) withhold politeness, that is, to keep silent or fail to act where politeness work is expected.

We randomly selected 400 successive comments in the middle of the polylogue and analysed all of the comments written in English (309) according to the above-mentioned classification. Impoliteness was found in 28% (n = 88) of the comments analysed. Positive impoliteness represented 66% of the impoliteness in the data analysed, negative impoliteness 27% and sarcasm 7%. Interestingly, in a study by Lorenzo-Dus et al. (2011, 2587), positive impoliteness was also the most-used strategy. In addition, we consider the feature ‘explicitly associate the other with a negative aspect’ to be part of positive impoliteness (Lorenzo-Dus et al. 2011, 2587) and not negative impoliteness (see Culpeper et al. 2003), for instance: ‘You sucked at Arabic not even close […]’.

The positive impoliteness strategy ‘call the other names’ was frequent both in our study and in Lorenzo-Dus et al. (2011, 2587), for example, ‘it’s called EXAGGERATION, dumbass. In our study, the positive impoliteness strategy ‘taboo words’ was common too, for example, ‘Not to sound mean or anything but your […] FUCKIN RACIST’. As this example shows, these strategies are also overlapping: in addition to the use of the taboo word ‘fucking’, a commentator ‘explicitly associates the other with a negative aspect’, namely, ‘racist’. The most frequent negative impoliteness strategy both in our study and in Lorenzo-Dus et al.’s (2011, 2587) is ‘condescend, scorn, ridicule’, for example, ‘It was just wat it sounded like to her…cant u read?

The study of impoliteness always involves interpreting the verbal act in its context. In other words, no linguistic forms or speech acts are impolite per se, but speakers/writers can express themselves verbally in a manner that can be perceived as polite or impolite. For instance, taboo words, usually conceived of as impolite language, function as solidarity markers in certain communities. In example (5), we interpret the comment by [Illyasviel von Einzbern] as impolite.



I did not get a dingle word of portugusese xD And thats my language!! I heard it 5 times now

[Illyasviel von Einzbern]:



@ [ChoiPipo] Cause she’s not talking portuguese, she’s just trying to express how it sounds, if you dont actually speak the language.


Thank you very much for the explanation @[The42ndUniverse]! And points to you for maturity ;)

In example (5), the comment by [Illyasviel von Einzbern] is an imperative speech act directed at [ChoiPipo]. We characterise the comment as impolite because of the following indices. The imperative contains a taboo word and an emoticon representing an angry face; further, it is written in capital letters and it contains repetitive exclamation marks, altogether representing a shouted imperative. Example (5), however, illustrates the context dependency of the social meaning of linguistic markers: In his or her first comment, [ChoiPipo] uses repetitive punctuation. The comment as a whole can nevertheless hardly be interpreted as impolite – rather, the repeated exclamation marks highlight the wonder and disappointment experienced by the commenter.

It must be mentioned that [Illyasviel von Einzbern] repeats the same comment as in examples (1) and (5) several times in the comment thread we analysed. This comment can be labelled as an offensive event, but as example (5) shows, the comment does not always trigger a conflictual response. In example (5), the initial commenter [ChoiPipo] responds to [Illyasviel von Einzbern]’s impoliteness indirectly in an ironic way: he or she complements [The42ndUniverse]’s maturity, which can be interpreted as an evaluation that [Illyasviel von Einzbern] is immature in his or her behaviour.

Playful language use

According to Danet (2013), another important component of flaming is playful language use. It appears that as the Smokahontas comment thread evolves over time, the language becomes more and more playful. This applies especially to comments pertaining to the above-mentioned ‘non-understanding’ theme. This conflict reoccurs with such frequency in the comment thread that it characterises the material as a whole. The reoccurrence of this same conflict indicates that for many participants, YouTube video commenting does not include reading comments that have already been posted. Thousands of commenters express their frustration at not understanding their native language as presented by Smokahontas, despite the fact that numerous comments with similar content already exist in the comment thread. These motivate other users to post playful, ironic comments imitating them. Separate examples (6), (7) and (8) from the comment thread illustrate these metacomments on the ‘non-understanding’ theme:




I’m Gibberish and I did not understand anything at all.


Step 1: Read the title: “What Languages Sound Like to Foreigners” Step 2: Observe key word: “SOUND LIKE” Step 3: Repeat it loud after me: “SOUND LIKE” Step 4: Write your comment anyway that you do not understand what she says in your language. Step 5: Feel smart

In example (6), the user plays with characteristics of computer-mediated communication: the ‘insert’ command is a frequent formula in online text functions. The author of example (7) plays with the term ‘gibberish’ and uses it as a name of a nationality. Finally, the format of example (8) resembles directions for using a device – as YouTube comment directions, the irony lies in the designated audience, who are obviously lacking intelligence.

The ‘non-understanding’ conflict and the varied responses to it indicate the existence of an ‘in-group’ among the participants, an ‘us’ distinguished from ‘them’ (see Bou-Franch & Garcés-Conejos Blitvich 2014, 28). ‘We’ refers to those who understand the main idea of the video and ‘they’ refers to those who do not. Participation in the in-group seems to allow flaming: impoliteness and playful language use.

Results and discussion

The video discussed in the comment thread under scrutiny in this paper is one of entertainment, in contrast to the comment thread studied by Bou-Franch and Garcés-Conejos Blitvich (2014), which pertains to ideological questions concerning sexual minorities. We therefore assume that the motivations of the participants in the Smokahontas comment thread resemble those in online celebrity gossip, as described by Eronen (2015). She views the comments as appeals to communicate and as self-expressions; the participants express proximity or even an intimacy to certain groups as well as distancing from others, the latter allowing mockery and moralisation. In our study on the Smokahontas YouTube comment thread, the focus is not on commenting on the celebrity in the video but on the interaction between the commenters. However, the motivation of posting these comments is most certainly similar to the self-expression in online celebrity gossip. The mockery – or flaming – toward other participants appears to be most common in our material among those who experience and want to express intellectual superiority with regard to others.

We analysed the interaction between participants in the Smokahontas comment thread on YouTube with a specific focus on flaming (i.e. sudden, intense conflicts involving impoliteness and playful language use). The conflicts in the material correspond to the definition of flaming, especially in that they represent discrepancies concerning other participants instead of criticising their ideas or ideologies. We have shown that conflicts in a massive polylogue do not necessarily continue after an offensive event due to the volatile nature of participation. The impoliteness detected in the comment thread comprised mainly threats to the positive face of the other, for instance, using taboo words and disrespectful naming. The study also showed that impoliteness could not be interpreted without the study of words and utterances in their context.

The volatility of participation also engenders repetition – a feature characterising the Smokahontas comment thread interaction: the same themes and the same type of conflicts emerge frequently. Indeed, the infinite and constantly changing number of participants both synchronically and diachronically, as well as the complex nature of the comment thread structure, create a ‘massive polylogue’. These massive polylogues are settings in which different virtual groups are formed. The interaction within and between these groups creates implicit conventions of verbal presentation of self, especially in cases of disagreement.



Primary sources

Comments from “What languages sound like to foreigners” video by Smokahontas (Sara Maria Forsberg), YouTube, 3.3.2014, Link verified 18.2.2016.


Angouri, Jo and Miriam A. Locher. 2012. “Theorising disagreement”. Journal of Pragmatics, 44 (12): 1549–1553. DOI: 10.1016/j.pragma.2012.06.011

Avgerinakou, Anthi. 2003. “Flaming in computer-mediated interaction”. In Rethinking Communicative Interaction: New Interdisciplinary Horizons, edited by Colin B. Grant, 273-293. Amsterdam: John Benjamins.

Brown, Penelope, and Stephen C. Levinson. 1987. Politeness: Some universals in language use. Cambridge: Cambridge University Press.

Bou-Franch, Patricia, Nuria Lorenzo-Dus, and Pilar Garcés-Conejos Blitvich. 2012. “Social interaction in YouTube text-based polylogues: a study of coherence”. Journal of Computer-Mediated Communication 17 (4): 501–521. DOI: 10.1111/j.1083-6101.2012.01579.x

Bou-Franch, Patricia, and Pilar Garcés-Conejos Blitvich. 2014. “Conflict management in massive polylogues: A case study from YouTube.” Journal of Pragmatics, 73. 19–36. DOI: 10.1016/j.pragma.2014.05.001

Culpeper, Jonathan. 1996. “Towards an anatomy of impoliteness”. Journal of Pragmatics 25, 349–365. DOI: 10.1016/0378-2166(95)00014-3

Culpeper, Jonathan. 2013. “Part I. Impoliteness: Questions and Answers”. In Aspects of Linguistic Politeness, edited by Demis Janet and Manuel Jobert, 2-15. Cambridge: Cambridge Scholars Publishing.

Culpeper, Jonathan, Derek Bousfield, and Anne Wichmann. 2003. “Impoliteness revisited: with special reference to dynamic and prosodic aspects”. Journal of Pragmatics, 35. 1545–1579. DOI: 10.1016/S0378-2166(02)00118-2

Danet, Brenda. 2013. “Flaming and linguistic impoliteness on a listserv”. In Pragmatics of Computer-Mediated Communication, edited by Susan C. Herring, Dieter Stein and Tuija Virtanen, 639–664. Berlin/New York: De Gruyter.

Dervin, Fred, and Yasmine Abbas. 2009. “Introduction”. In Digital Technologies of the Self, edited by Yasmine Abbas and Fred Dervin, 1-14. Newcastle upon Tyne: Cambridge Scholars Publishing.

Diener, Edward. 1977. “Deindividuation: Causes and consequences”. Social Behavior and Personality 5 (1): 143–155.

Dobs, Abby Mueller, and Pilar Garcés-Conejos Blitvitch. 2013. “Impoliteness in polylogal interaction: Accounting for face-threat witnesses’ responses”. Journal of Pragmatics 53, 112–130. DOI: 10.1016/j.pragma.2013.05.002

Dynel, Marta. 2014. “Participation framework undelying YouTube interaction”. Journal of Pragmatics 73: 37–52. doi:10.1016/j.pragma.2014.04.001

Eronen, Maria. 2015. Rhetoric of Self-Expressions in Online Celebrity Gossip. Vaasa: University of Vaasa.

Fracchiolla, Béatrice. 2013. “De l’agression à la violence verbale, de l’éthologie à l’anthropologie de la communication”. In Violences verbales : analyses, enjeux et perspectives, edited by Béatrice Fracchiolla, Christina Romain, Claudine Moïse and Nathalie Auger, 19–36. Rennes: Presses universitaires de Rennes.

Garcés-Conejos Blitvich, Pilar. 2009. “Impoliteness and identity in the American news media: the “Culture Wars””. Journal of Politeness Research 5 (2): 273–303. DOI: 10.1515/JPLR.2009.014

Goffman, Erving. 1959. The Presentation of Self in Everyday Life. New York: Anchor Books.

Guadagno, Rosanna E., Daniel M. Rempala, Shannon Murphy, and Bradley M. Okdie. 2013. “What makes a video go viral? An analysis of emotional contagion and Internet memes”. Computers in Human Behavior, 29 (6): 2312–2319. DOI: 10.1016/j.chb.2013.04.016

Hardaker, Claire. 2010. “Trolling in asynchronous computer-mediated communication: From user discussions to academic definitions”. Journal of Politeness Research, 6: 215–242. DOI: 10.1515/JPLR.2010.011

Jay, Timothy. 1992. Cursing in America: A Psycholinguistic Study of Dirty Language in the Courts, in the Movies, in the Schoolyards and on the Streets. Philadelphia and Amsterdam: John Benjamins.

Kayany, Joseph M. 1998. “Contexts of uninhibited online behavior: Flaming in social newsgoups on Usenet”. Journal of the American Society for Information Science 49 (12): 1135–1141. DOI: 10.1002/(SICI)1097–4571(1998)49:12<1135::AID-ASI8>3.0.CO;2-W

Kiesler, Sara, Jane Siegel, and Timothy W. McGuire. 1984. “Social psychological aspects of computer-mediated communication”. American Psychologist, 39: 1123–1134.

Langlotz, Andreas, and Miriam A. Locher. 2012. “Ways of communicating emotional stance in online disagreements”. Journal of Pragmatics, 44: 1591–1606. DOI: 10.1016/j.pragma.2012.04.002

Lorenzo-Dus, Nuria, Pilar Garcés-Conejos Blitvich, and Patricia Bou-Franch. 2011. “On-line polylogues and impoliteness: The Case of postings sent in response to Obama Reggaeton YouTube video”. Journal of Pragmatics 43. 2578–2593. DOI: 10.1016/j.pragma.2011.03.005

Marcoccia, Michel. 2004. “On-line polylogues: conversation structure and participation framework in internet newsgroups”. Journal of Pragmatics 36 (1): 115–145. DOI: 10.1016/S0378-2166(03)00038-9

Moïse, Claudine. 2012. “Argumentation, confrontation et violence verbale fulgurante”. Argumentation et Analyse du Discours, 8. URL:

Moor, Peter J., Ard Heuvelman, and Ria Verleur. 2010. “Flaming on YouTube”. Computers in Human Behavior, 26: 1536–1546. DOI: 10.1016/j.chb.2010.05.023

Pagliai, Valentina. 2010. “Introduction: Performing Disputes”. Journal of Linguistic Anthropology 20 (1): 63–71. DOI: 10.1111/j.1548-1395.2010.01048.x

Pascual y Cabo, Diego and Jason Rothman. 2012. “Multilingualism and Identity”. In The Encyclopedia of Applied Linguistics. Wiley Online Library. DOI: 10.1002/9781405198431

Schiffrin, Deborah. 1984. “Jewish argument as sociability”. Language in Society 13 (3), 311–335. DOI:

Smith, Eliot R., and Diane M. Mackie. 2000. Social Psychology. Hove: Psychology Press.

Spears, Russell, and Tom Postmes. 2015. “Group identity, social influence, and collective action online: Extensions and applications of the SIDE model”. In The Handbook of the Psychology of Communication Technology, edited by S. Shyam Sundar, 23–46. Chichester: Wiley Blackwell.

Spencer-Oatey, Helen. 2007. “Theories of identity and the analysis of face”. Journal of Pragmatics, 39 (4): 639–656. DOI: 10.1016/j.pragma.2006.12.004

Turnage, Anna K. 2007. “Email flaming behaviors and organizational conflict”. Journal of Computer-Mediated Communication, 13 (1), article 3. DOI: 10.1111/j.1083-6101.2007.00385.x

Turner, Graeme. 2004. Understanding Celebrity. London: SAGE.

Turner, Graeme. 2010. Ordinary People and the Media. The Demotic Turn. London, California, New Delhi and Singapore: SAGE.

Vrooman, Steven S. 2002. “The art of invective: Performing identity in cyberspace”. New Media & Society 4(1): 51–70. DOI: 10.1177/14614440222226262

Walther, Joseph B. and Jeong-woo Jang. 2012. “Communication processes in participatory websites”. Journal of Computer-Mediated Communication 8 (1): 2–15. DOI: 10.1111/j.1083-6101.2012.01592.x


[1] The YouTube API did not allow us to download all of the comments; therefore, we counted the comment frequencies from a sample.

[2] Lorenzo-Dus et al. (2011, 2587) consider this feature, ‘explicitly associate the other with a negative aspect’, to be part of positive impoliteness and not negative impoliteness; this is because it displays a functional similarity to the positive impoliteness strategy ‘call the other names’. We agree with this view.