Between Saying and Not Saying

This contribution focusses on forms and functions of explicit silence in user comments on the Facebook page of the New Right political party Alternative für Deutschland and on the way in which they allow users to participate in the communication while withholding content. These explicit silences create vagueness and ambiguity which users exploit to at the same time insinuate, adhere to, and scandalise taboos.


Introduction
The present article was prompted by previous work in two ways. First, and more immediately, I collected user comments from migration-related Facebook posts by the Alternative für Deutschland (AfD) from the central AfD Facebook page to write about the self-victimising discourse that occurs in these user comments in the context of online hate speech (Schröter 2022). Second, I had already published a few contributions looking into the metadiscourse about silence and silencing in public discourse perpetuated by the political right in Germany (Schröter 2015(Schröter , 2019a(Schröter , b, c, 2021. Because of the latter, I could not help noticing the various instantiations of this metadiscourse in the Facebook user comments while working on the former contribution. In this article, I am going to analyse the same user comments, but with a view on users' performances of explicit silence. I am using the term for the phenomenon of users commenting on a Facebook post while explicitly stating that they cannot or will not say what they would otherwise have to say. I.e., rather than just keeping silent, they perform keeping silent and rather than just remaining absent from the interaction on the Facebook post, they seek to make their intention of refraining from commenting known. Beyond just noting that explicit silence can be one of the forms that silence may take, it will firstly be interesting to consider the communicative situations in which it occurs and its affordances and secondly, to look at its occurrence in discourse contexts. Doing so shows that explicit silence allows commenters to ambiguate their contribution to the communication -stating that they refrain from commenting or indicating that they deliberately withhold (parts of) statements allows them to participate in the communication, but to remain vague about content in the light of a perceived, or enacted, inability to speak their mind.
I am first going to discuss explicit silence in the context of research on silence and in the context of social media and specifically the AfD Facebook page. I am then going to describe my dataset and afterwards, I am going to analyse and discuss various ways in which silence and silencing is performed in the user comments, followed by a conclusion in which I discuss my findings in the light of the metadiscourse about silence and silencing within the German political right.

Between saying and not saying: explicit silence
Thus far, research on silence has mainly noted explicit silences, such as stating ›No comment‹, or ›I am not going to answer this question‹ in the context of political interviews (e.g., Ekström 2009) as well as police interrogation (e.g., Garbutt 2018) or court settings (e.g., Kurzon 1998). The former have been discussed with a view on political leadership, accountability and journalists' interviewing practices, and the latter two with regard to legal provisions and possible consequences for individuals refraining from making statements.
Explicit silence in the context of Facebook post comments is rather different. Unlike when being asked a question in an interview or interrogation, which entirely rely on question-answer sequences, no individual needs to feel called upon to engage, unless they are perhaps tagged in a post. This is even more the case on sites with a high number of followers. If individuals choose not to engage, there will just be fewer responses to the post in question. It is less common that anyone will specifically miss a comment from a particular individual Facebook follower or friend. Non-commenting by Facebook friends and followers is ambiguous for those who posted something and expect reactions. We cannot be sure whether a user has just refrained from Facebook, overlooked the post, not cared enough about the post, intended to react or comment but forgot about it, or intended to react or comment but decided against it, intended to react or comment but struggled to express themselves, intended to comment but only reacted with one of the prefabricated options instead. However, at an individual level, abstention from commenting on a site like the AfD's with more than half a million followers at the time of data collection is unlikely to become a notable and meaningful absence. It will only show as the collective result of individuals' decisions in the form of some posts triggering a higher or lower level of response. While decisions by individuals to react to a post might be based on considerations of increasing the response rate of a post, it is more likely that they react with a comment if they feel prompted by the post's content and feel they have something to say about it. Why, then, would they bother commenting just to declare that they have nothing to say, or cannot or do not want to say all they have to say? Explicit silence is less ambiguous than non-commenting in that the act of commenting, even if only to state ›not saying‹ excludes the previously mentioned options. The user has clearly noted the post and decided to comment, they show communicative engagement and action with the aim of telling us something, rather than nothing at all. What they tell us, though, is that they are saying nothing or not saying everything. What explicit silence remains ambiguous about is its content and possibly its intention. Explicit silence leaves us wondering what it is that could or would have been said, and why the user chooses, as they claim, to withhold it.
I am referring to such communicative acts as explicit silence here, observing that the act of being silent or withholding communication can be explicitly performed. The functions and meanings of doing so can be elucidated by comparing it to silence in the form of withholding communication. Contributions in the volume edited by Jung (2019) look at practices of abstention from (political or societal) participation and consider the possibility of such abstention being an intentional, meaningful act rather than mere oblivious passivity. Scott (2019) calls such practices acts of commission, »when we deliberately choose not to do or be something with conscious intentionality. This typically involves negative motives of avoidance, disengagement, disavowal, refusal or rejection.« (p. 13) As observed just above in the context of social media, abstention might only become notable and meaningful at a collective, not at an individual level, such as non-voting or withholding applause within a large audience: If one person abstains, even if it is a deliberate and meaningful act for them, there will be little noticing of this, unless many other individuals do the same.
It would therefore seem as though Facebook users perform silence explicitly because they want to make sure that their individual withholding of communication gets noticed. In further characterising acts of commission, Scott also asserts that such »social acts are performative displays, designed with an audience in mind: the actor wants to be seen not to be doing something and to have their reasons acknowledged.« (Scott 2019, p. 13.) Therefore, the explicit silences that I am going to analyse might be understood as acts of commission; Facebook users may want others to think about why they state that they refrain from commenting, or they even provide a reason themselves, and they may want others to wonder about what they would have said. There is therefore an ambiguity inherent in the very act of stating these explicit silences: The users want to abstain, but they want their abstention noticed. They want to refrain from ›saying something‹, but make their act of commission known to others. There is an ambiguity here between, at the same time, saying and not saying.
By stating their silence explicitly, we would assume that the Facebook users' silence is intentional, rather than symptomatic. Kurzon distinguishes between intentional silence »in the form of a modal expression ›I will not speak‹ or ›I may/ must not speak‹« (2007, p. 1676), whereby the latter »seems to imply an external source that forces the person not to speak« (Kurzon 2007(Kurzon , p. 1677, and unintentional silence which »may be glossed as ›I cannot speak‹ and relates to psychological inhibitions that may prevent the person from opening his or her mouth.« (Kurzon 2007(Kurzon , p. 1677) I will show below that all of these types of silences are put on display by Facebook users commenting on AfD posts. In one of the seminal early articles on silence, Johannesen (1974) names several possible meanings of silence, a few of which appear to apply to the explicit silences stated by Facebook commenters, such as »avoiding discussion of a controversial or sensitive issue out of fear«, being »emotionally overcome« or feeling »inarticulate despite a desire to communicate« (p. 29). In another seminal early article, Jensen (1973) points out the inherent ambiguity of silence by distinguishing a number of functions that it can assume, all of which consist of two opposites: With regard to the linkage function, silence can bring people together, but also isolate; the affecting function means that silence might be either wounding or have healing effects; with regard to the revelation function, silence can be revealing in that it could point to a psychological disposition, but it can also be used to cover up. Within the activating function, silence can indicate either activity or inactivity and, finally, silence can mean either affirmation or negation in what Jensen calls its judgmental function. Silence as a means of communication is therefore already inherently ambiguous. I will also show below how Facebook users oscillate particularly around the revelation function and add to this already inherent ambiguity of silence by utilising both opposite aspects of the revelation function, i.e., to reveal and to conceal at the same time.

Explicit silence in the context of the AfD and its Facebook environment
The context of the AfD Facebook page might further help us to understand such occurrences. Firstly, the polarising and galvanising effect of political communication on social media has often been noted (e.g., Sunstein 2017). Medina Serrano et al.
(2019) find that content posted on the AfD Facebook page tends to be provocative or even sensationalist, and that the »topics discussed are controversial, which encourages users to engage with the posts and express personal opinions.« (218). They also show that the AfD amplifies certain issues on its social media platform, compared to its manifesto: 21% of manifesto content deals with economic politics and 19.2% with immigration; whereas only 4.5% of Facebook content pertains to economic issues, but 16.1% to immigration (ibid., p. 222; based on data from January 2015-May 2018). Secondly, the phenomenon of decreased inhibition in online communication has also been noted, such as users venting extreme views or using hate speech (Udupa/ Gagliardone/Hervik 2021; Assimakopoulos/Baider/Millar 2017). This indicates that explicit silences might be due to an awareness of taboos. Users might avoid saying anything that is taboo (for example, offensive, discriminating, outrageous, or disgusting) and which might therefore even be sanctionable, such as having their posts deleted or rendered invisible to others, or being barred from commenting for a while. 1 Of course, they might avoid taboo by refraining from communicating altogether. If instead they choose explicit silence, they imply, or partly also explicitly state something like: I would like to say certain things, but if I do, this might have undesirable consequences (see sections 5.2 and 5.3 below). At a meta-level, they can thereby point out the fact that what they have to say is subject to tabooing. Explicit silence creates ambiguity about their participation in the communication, by at the same time offering and hiding content. It keeps the proposition of their contribution vague, which helps navigating such taboos, working around them by insinuating, but not spelling out offensive content. For AfD followers on Facebook, the ambiguity and vagueness that can be created with explicit silence is therefore an opportunity (see Felder, in this issue) for making themselves noticeable as participants in the communication while avoiding (parts of) what they have to say in the face of (purported) tabooing of discriminatory or otherwise offensive content.
Thirdly, the AfD is met with little acceptance on part of the other political parties and most of the widely circulating media in Germany. It is a relatively new party, founded in 2013 at first with a mainly neoliberal, EU-sceptic agenda in reaction to the European debt crisis (cf. Häusler/Roeser 2016). Since 2014, the party surpassed the threshold of 5% of the votes to constitute parliamentary fractions in the sixteen federal states. In the years 2014 and 2015, right-wing and anti-immigrant positions in the party became more prominent. Some of the founding members left the party and the ethno-nationalist wing gained dominance, hovering between right-wing populist and extreme right stances (cf. Kim 2017) and tensions keep on arising within the party around where to draw a line to the extreme right. In this formation, the party scandalised the immigration of refugees, which brought them electoral success in the 2017 federal election, where the AfD won 12.6% of the vote -the third largest share of votes among all parties. Their share of votes decreased slightly in the 2021 election to 10.3%. The AfD therefore needs to be considered a powerful political player in Germany, even though it positions itself as an anti-establishment party.
The lack of acceptance of the AfD and the degree of its problematisation despite its status as a political party and relative electoral success also means that followers who agree with the AfD and see it as their main political platform might feel that their views are not widely accepted. Therefore, explicit silence might also be used to indicate powerlessness: There is no point in me saying what I have to say, because I do not expect it to have the desired resonance (see section 5.4 below).
Finally, the AfD is an opposition party and as such more likely to attempt devaluing and scandalising current politics, and it is surrounded by a lot of controversy. This means that the AfD is more likely than a party involved in government to post scandalising content that might stir up its users, which might also be intended for doing just that. Confronted with such input, users may find themselves in a position where what they have read renders them unable to appropriately verbalise a coherent response -they are speechless: I want to say something about what I have just read, but I find it so outrageous and incomprehensible that I have difficulties with formulating the kind of response I would like to provide (see section 5.1 below)

Data and methodology
The AfD Facebook page can be found under www.facebook.com/alternativefuerde. It was initiated on 3 March 2013 and as of 12 April 2021, it attracted 512,193 likes and 538,916 followers. According to Müller/Schwarz (2021), the party's reach on Facebook exceeds that of other German political parties. The AfD page does not state rules of conduct for posting comments and it is »consistently more focused on refugees than that of traditional news reports and frequently contains loaded terms that civil rights groups have identified as ›hate speech‹« (Müller/Schwarz 2021, p. 3, cf. also p. 33). Compared to all other German political parties, Medina Serrano et al.
(2019) find that the AfD Facebook page has the most followers, that it posts most actively, and that its posts attract the most comments, as well as likes and shares.
In a first step, to gain an overview over the activity on the AfD Facebook page, and over the kind and amount of migration-related posts, a sample time span of three months was selected during which all posts were surveyed for content and amount of interaction. The selected time span covers the months of October, November and December 2020. The time span was not selected with the intention to capture any specific events or likely issues occurring during the time span. It was selected so as to be practically manageable -first in terms of the volume of analysis and second in accessibility of posts. During these three months, typically three or four posts appear on the page every day. Posts were considered to relate to migration if they referred to a) movement by foreign nationals, including refugees, into Germany, or b) the presence of and provision for refugees, or c) conflicts arising from current or past immigration, including posts highlighting criminal activities by immigrants and d) posts critical of Islam. In October, there were 33 migration-related posts, 27 in November and 8 in December, which amounts to a total of 68 migration-related posts during the selected time span. Thematically, these 68 migration-related posts pertain to 1) purported criminal activity by Muslims in Germany or elsewhere in Europe (14 posts); 2) advocating sanctions for such purported criminal activity (15 posts); 3) criticism of Islam and the presence of Muslims in Germany (16 posts); 4) provisions for refugees such as accommodation or rights granted (9 posts); 5) (potential) movements by foreign nationals, including refugees, into Germany and the EU (12 posts); 6) individual migrants who are presented as in agreement with the AfD (2 posts).
The above overview of posts demonstrates that they relate to issues perceived as the most problematic aspects of migration. There is likely to be a relation between the gist of the post and the reactions to it voiced in the comments. Wahlström/ Törnberg/Ekbrand (2020) observe that the »framing in the top post« can determine »the occurrence of violent and dehumanizing rhetoric in the discussion threads.« (p. 14) This reinforces concerns about the potential of Facebook posts to provoke or intensify the expression of hate. Hafeneger/Jestädt/Klose/Lewek (2018, p. 12) characterise the AfD's position on immigration, according to which immigrants are purportedly exploiting the social benefit system instead of participating in the labour market. The situation is portrayed as pending social unrest and a creeping extinction of European culture. The AfD advocate a severe limitation of immigration and the processing of asylum applications outside German borders. They also advocate deporting immigrants whose first applications were declined without the right of appeal. Müller/Schwarz (2021) analyse the narrative in AfD Facebook posts that contain the word Flüchtling (refugee) 2 . It »centers around the idea that the ›elites‹-politicians and mainstream media outlets-have betrayed ›the people‹ by allowing ›streams‹ of illegitimate ›economic refugees‹ to enter the country, who are described as being criminals and rapists for ›cultural reasons‹«. (Müller/Schwarz 2021, p. 9). The authors also observe that responses to these posts show that »the overwhelming majority appear to agree with the positions of the AfD.« (Müller/ Schwarz 2021, p. 9). In short, the topics described above serve to highlight, amplify, and exaggerate only the most problematic aspects of immigration, and to portray immigrants as a problem and a burden. 3 In a second step, nine posts were selected for a more detailed analysis of user comments. The selection includes posts from the three months for which data was surveyed as well as different topics from the above list, but it excludes posts about migration-related events or policy responses in other countries. The table in the Appendix provides an overview together with an indication of the post content, with reference to the topic category above. The topic caption in the left column of the table indicates the gist of each post.
In a third step, comments on the posts were retrieved with the aid of an online tool (https://exportcomments.com, last access 20/06/2021). The tool extracts all publicly visible comments and replies to comments to an excel sheet, including links to external content and emojis. The table in the Appendix indicates the level of interaction on posts and comments both as per Facebook's own count as well as per number of comments and replies that were retrieved by the tool. The number of comments and replies that were retrieved by the tool is lower than the number indicated on Facebook. Kalsnes/Ihlbaek (2021) investigated practices of moderating debates on political parties' Facebook pages, since »pressures and public expecta-tions have increased for political actors to pay attention to and delete comments perceived as hateful or derogatory.« (p. 329) It can be assumed that some degree of comment moderation takes place on the AfD Facebook page. The discrepancy in the apparent and retrieved number of comments might be an indicator that some comments were hidden by moderators 4 , so that they are still included in the Facebook comment count and visible to the respective commenter, but not retrievable due to a lack of visibility of the comment to anyone else. There are also indications in some user comments that users had previously had their comments removed or that they were temporarily barred from commenting: Users themselves explicitly mention the fact that this has happened to them, or state that they refrain from writing what they would wish to write in order to not be banned or to not have their comments removed (see section 5.2 below).
Both figures, however, indicate that there is a sizeable amount of interaction with the post content, although numbers vary between different kinds of interaction: Generally, the number of simple reactions (like symbol, heart symbol, or emoji) surpasses the number of post shares, the number of post shares surpasses the number of comments, and the number of comments surpasses the number of replies to comments. The average number of user comments retrieved by the downloading tool across the nine selected posts is 1,626.
In a fourth step, rather than examining all comments per post in detail, the analysis will draw on the retrieved comments as a large pool of examples to, firstly, identify instances of explicit silence. This necessarily has to be a qualitative analysis, since, as we shall see, explicit statements of silence can be worded in various ways. However, while retrieving and categorising these instances across a range of comments to nine different posts, it is important to, secondly, ensure that each noted phenomenon recurs across different posts and different users' comments, and is therefore not a singular, isolated occurrence. This serves to demonstrate that explicit silence is a recurrent pattern in user comments, and that the different forms of stating silence discussed below are also recurring across different comments by different users on different posts.
In a fifth step, after identifying and illustrating the uses of the different kinds of stating silence, the possible functions of each of them will be considered in their context. The analysis thereby broadly follows the model of the Discourse Historical Approach to Critical Discourse analysis (Reisigl/Wodak 2009). In line with this, discourse is understood to pertain to specific fields of social action, and to exhibit argumentativity. In order to situate the discourse at hand accordingly, it was contextualised in the previous sections with regard to user-generated content on social media (as a social practice) and within the political party and its Facebook platform (i.e., the field of social action) on which the user comments occur. The relation to the macro-topic migration was characterised above. Apart from noting the above level of interaction, the more micro-dynamics of the interaction, i.e., structures of comments and replies are not in the focus of this analysis. Having said this, the immediate co-text and other intertextual and interdiscursive relations will be taken into account and build on the above characterisation of the discourse of the AfD. I will draw the observations based on the occurrences of explicit silence together in the conclusion, where I will discuss them as ambiguation and point out its functions, as well as situate them as part of a metadiscourse about ›being silenced‹ and purported censorship 5 on part of the German political right. By commenting a ›non-comment‹ rather than refraining from commenting, what users seem to want to demonstrate here is a strong and overwhelming reaction to what they read in the post. They appear to be struggling to rationally grasp what is going on (›incredible‹, ›unbelievable‹, ›just makes me angry‹) either because of their own incapacity or because of the irrationality of the situation (›madness‹, ›unheard of‹). They use explicit silence to display that their mind is going blank, thoughts do not form clearly; this and other strong reactions (›anger‹, ›bile‹) defy verbalisation. The explicit silence performed in the comments above is therefore portrayed as an unintentional silence (I would like to say something about this, but...) that comes upon the commentator (›I cannot speak‹), relating »to psychological inhibitions that may prevent the person from opening his or her mouth.« (Kurzon 2007(Kurzon , p. 1677. When it comes to considering the possible meanings of these explicit silences in Johannesen's words, it would be either being »emotionally overcome« or feeling »inarticulate despite a desire to communicate« (Johannesen 1974, p. 29). Bruneau (1994) also notes this type of silence in »(m)oments of high sensation« which »seem almost to demand silence and absence of cognitive control.« (Bruneau 1994, p. 21). However, in stating their speechlessness, their silence is not a silence anymore, and it is also not the kind of symptomatic silence anymore that they indicate -they have already got their act together to write something, so rather than being just symptomatic, what they do must have meaning and function.
Stating their speechlessness could therefore at one level just be telling other Facebook users that they have experienced a strong emotional response to what they have read in the post. However, other Facebook users who may engage with their comment are likely to be strangers to them who may not be particularly interested in how they feel. It makes more sense to do this when users can assume that they are participating in a shared experience of similar feelings on other users' parts. Therefore, at another level, their explicit silence could invoke a shared frame of interpretation of the post content which considers the events described in the posts as so unreasonable that it defies verbalisation. The status of the users as participants in the communication therefore remains ambiguous; they contribute while saying nothing, and in doing so, rather than refraining from commenting altogether, they can show themselves as a member of a ›community of experience‹. They can communicate and hope to be understood without offering propositional content which therefore remains vague. Including comments that just state »Without words«, explicit silences as demonstrating speechlessness like in the examples above occur 120 times across the nine posts, i.e., on average 13.3 of these explicit silences occur per post in the comments.

Explicit silence for fear of sanctions
Explicit silences also occur as demonstrative self-silencing. Users claim that they remain silent because they would not be allowed to say what they had to say, as the following comments indicate: The comments above illustrate that users also use this kind of explicit silence to participate in the communication and to signal their readiness to contribute. However, the type of explicit silence indicated by them is more deliberate and intentional; users decide to draw a line, but this line is externally imposed on them, i.e., »some person or body or even social norm seems to impose the silence on the silent person(s)« (Kurzon 2007(Kurzon , p. 1676 or, in Johannesen's words, they are »avoiding discussion of a controversial or sensitive issue out of fear.« (Johannesen 1974, p. 29 Where this occurs in interaction among users, i.e., in replies to other users' comments, it also becomes apparent that users here advise others to draw a line so that they will not become subject to such sanctions. Stating explicit silence with reference to sanctions if they spoke freely instead of remaining silent seems to serve mainly two purposes here. First, users can scandalise the fact that voicing their opinions might be sanctionable as in examples 21 and 22 which contain critical remarks about this state of affairs that point to an inhibition of political discourse and a lack of democracy. Second, ambiguously stating and not stating at the same time allows them to avoid any sanctions. Overall, such posts are quite frequent; I counted 151 comments that refer to imposed silence, an average of 16.7 such comments on every post. One user observes the other users' refraining from posting what is on their minds, leading to a reflective comment about the limitation to the ›freedom of opinion in Germany‹ through certain topics becoming ›taboo‹: (26) When I read the comments -we are still allowed to read AfD-comments, or not? -I notice that freedom of opinion in Germany has become something that you can only afford in your own home with good friends. Frequently, comments end with »I would have a lot to say, but I better keep it to myself«. This means that everything that relates to unregulated migration and to the costs it incurs is already a taboo topic in Germany. Second, users can use their stated silence with reference to sanctions to invite the other users to imagine what they would have said if they had felt able to do so as in examples 24 and 25 which evoke consensus or invite readers to imagine something extreme. This allows users to ›say and not say‹ at the same time; by stating their refusal to spell out their thoughts but indicating what these thoughts might be (same as other users', or ›cruel and brutal‹), they do not just leave a blank, but invite readers to fill the blank. The in-group is invited to resolve the vagueness arising from withheld content by inserting their own ideas. The latter aspect seems to also play into exaggerations of possible sanctions if the users wrote what was on their minds, which appear to invite, or to fire up readers to fill the blank with extreme content, which can be seen in the following examples: Similar to the examples above, users ambiguate their status as participants in the communication by saying and not saying at the same time. What they would have said remains vague, but their metadiscursive statements about why they don't say have identifiable content and function. They frame the withheld content as subject to censorship by Facebook or by the state (referring to imprisonment, special task force, Stasi -the latter also framing a state institution as illegitimate by comparing it to a dictatorship institution) and they scandalise this purported censorship and position themselves as victims of silencing. The in-group is invited to resolve the vagueness by inserting their own thoughts about what could or should have been stated, and the out-group cannot proceed with negotiating content in the manner envisaged, e.g., by Felder (in this issue). The following exchange further illustrates that users may quite consciously draw and walk the line between saying and not saying. The comment that triggers the exchange starts with apparently simply making a comment about winter temperatures, but the wording of the hashtag that is added to it, with reference to furnace 6 that are still there points to the ovens that were used to burn corpses in concentration camps. In the immediate co-text of the comment, it therefore suggests fighting the cold temperatures by reactivating these furnaces. In the context of the Facebook post, it therefore suggests the reactivation of concentration and extermination camps as a way to deal with migrants who are making demandsthe AfD post was about an ›Islamist demonstration‹ in Hamburg. Two users seem to understand the post in this way and signal that while they are not opposed to the idea, they find the comment risky and consider whether it would be better to refrain from posting such content or from indicating agreement. The original commentator replies with confidence in their ability to walk the line between saying and not saying by ›circumscribing it nicely‹. The use of explicit silence with reference to fear of possible sanctions therefore appears to be a way to manage constraints in posting extreme content and hate speech. Explicit silence allows them to ›not say‹ and therefore avoid sanctions. Rather than just refraining from commenting, posting explicit silence however allows them to ›say‹ at the same time by indicating a blank that other users are invited to fill. Rather than abstaining from commenting, this way they can become visible as a member of the AfD Facebook community of experience; they suppose, enact and reinforce mutual agreement without posting a potentially precarious proposition. At the same time as walking the line between saying and not saying might be an entertaining game for users, as comments 27-31 suggest, by providing a reason for ›not saying‹, they can scandalise and mutually reinforce their dismay at the constraints that keep them from posting anything they like in any words they like. Because of the ›not saying‹ that the post performs at the same time, we cannot be sure whether what they would have said would have been in fact perfectly acceptable, as examples 11, 17 and 26 seem to suggest, or indeed more extreme, as examples 27-30 suggest. Especially in the former case, sanctions can be scandalised without allowing any discussion or negotiation as to where to draw a line, and why, because a proposition is withheld.

Explicit silence as indicating and avoiding taboo
Facebook users also make elliptic statements to at the same time indicate and avoid spelling out taboo content. Ellipses have been discussed as a form of silence (e.g., Schmitz 1990) and I would argue that they are forms of explicit silence in that they leave a given, anticipated structure incomplete and in doing so draw attention to what is missing. Again, ellipses allow users to walk a line between saying and not saying by pointing attention to the gap that they are leaving and thereby inviting others to fill it. In some instances, users simply avoid spelling out taboo words such as ›arse‹, ›shit‹ and ›fuck‹, as in the following examples. Users may be wary of algorithms picking up on certain discriminatory language, so they can avoid sanctions by not spelling these words out and ›not saying‹ completely. What they indicate is however enough to ›say‹ and for others to fill in what is left out. They also make their silence explicit by visualising it graphically with dots, asterisks or a combination of other signs. They thereby make the gap they are leaving visible and invite others to fill it. As Thurlow/Moshin (2018) write with regard to strategies of representing taboo language in journalistic writing, they »are busy playing with witty ways of speaking it nonetheless and leaving little or no doubt in the reader's mind about what the word is« and thereby they perform »acts of silencing which are patently far from silent« (Thurlow/Moshin 2018, p. 315). Apart from indicating but avoiding taboo words and swear words, there are also instances where users leave out part of a word or leave out a word entirely that would be needed to syntactically complete a sentence which indicate calls for violence. Again, this becomes a notable absence, often but not always indicated with dots or an emoji, and users are invited to fill in the blank or replace the emoji with a word. These examples show once more how users utilise explicit silence in the form of notable gaps to walk a line between saying and not saying which allows them to at the same time indicate and avoid taboo content. Ellipses helps them to state what they want to say clearly enough without having to spell it out, so that they can avoid possible sanctions. Given that the examples in the previous section above and in this section also show that users are conscious of possible limitations to the acceptability of what they write, they might take a certain pleasure in finding ways to at the same time avoid and indicate precarious content, or in seeing others do so successfully; it could be a game of checking what they can get away with. For ingroup communicative purposes, the elements of vagueness that arise from omitting (parts of) the obviously intended content do not create much ambiguity as readers will probably be able to complete the insinuated content. The ambiguation here lies not so much in the post writers' status as participant in the communication but is directed towards page moderators or algorithms trained for spotting hate speech to either evade or deny the blurred content directed at the in-group. I counted in total 108 of such uses of ellipses as explicit silence, which makes an average of 12 occurrences in the comments per post.

References to silencing and lack of voice
Beyond the use of explicit silence to ›say and not say‹ at the same time analysed in the previous subsections, users also refer to silences that they feel subjected to or surrounded by outside of the Facebook context that they are engaged in. While this is a somewhat different phenomenon to what I discussed above, it still links to the scandalisation of the silencing which these Facebook users see themselves subjected to. It therefore links to their use of explicit silence, and to their discourse about silence and, together with the following section, helps completing a wider picture that links to the metadiscourse about silencing and taboo that is perpetuated by the political right in Germany. For this reason, a few such examples shall be discussed in this section. Users state their lack of voice or purport that they are not being consulted -they lack political influence and feel powerless: According to comments like these, users or people who think like them are being silenced; nobody will ask them for their opinion or pay attention to it if they do voice it -they will not be acknowledged or heard. Moreover, the issues they care about are being ignored and silenced as well by powerful instances such as the government and the media: Another way to keep them silent or to shut them up, according to the comments on Facebook, is to declare them to be ›Nazis‹ or right-wing, as comments 55-57 illustrate: (55) We pay multiple times for these false border crossers, but you are not allowed to say this out loud, otherwise you're a »Nazi«-but the money has to come from somewhere [Wir zahlen doppelt und dreifach für diese falschen Grenzgänger, das darf man bloß nicht Laut sagen, sonst ist man ein »Nazi« -aber irgendwo muß das Geld ja herkommen ] (01/10) (56) Are refugees....get everything and if not then it's arson or murder and then we are the Nazis when we say anything about it [Sind Flüchtlinge... Bekommen doch alles und wenn nicht dann wird eben in Brand gesetzt oder gemordet und dann sind wir Nazis wenn man den Mund aufmacht] (14/10) (57) And if we did this, to take to the streets to demand respect for us, then we will be the brown swamp again [Und wenn wir das machen Für Respekt uns gegenüber auf die strasse gehen heißt es wieder der braune Sumpf ] (09/11) These examples show that beyond utilising silence themselves, users also perpetuate a metadiscourse about silence and silencing which they see themselves exposed to. In addition to using explicit silence themselves to walk the line between saying and not saying, they also talk about silence. Their comments about silence at a metalevel exhibit similar concerns as those noted in subsection 5.2 above. They are ›not allowed‹ to voice their concerns in that if they try, they are made political out-casts (right-wing), their point of view is not represented anywhere in more widely circulating discourse (examples 52-54) and they scandalise the fact that they are supposed to silently condone a disagreeable situation (examples 48-51). Overall, I counted 114 instances of such metalinguistic references to silence, with an average of 12.6 comments per post.

Conclusion
The analyses above demonstrate that Facebook users may state their silence explicitly in different ways and with different functions, which cut across the sections above: Firstly, to participate in a shared experience; secondly, to highlight imposed restrictions; and, thirdly, to not say, but to say at the same time. Finally, users also refer to ways in which issues they care about are being silenced more widely. Apart from the seemingly symptomatic explicit silence discussed in 5.1, all the other forms of explicit silence are related to silences that the users regard as imposed on them, i.e. forms of silencing. Such user comments reflect a wider metadiscourse on part of the German New Right in which they claim to be silenced and excluded from discourse that I have traced through a number of debates (Schröter 2015), individuals (Schröter 2019c), political parties (Schröter 2019a) and key periodicals (Schröter 2021). Individual Facebook users appear to have adopted the habitus of the silenced and marginalised political opponent that more prominent exponents of the German political right have been exhibiting for years. Thiesmeyer (2003) discusses silencing in discourse contexts, noting on the one hand that discourse analysis is concerned with structures of power and control and, therefore, with opportunities for expression, but on the other hand, »that silencing will be practised where the potential for counter-silencing exists«, which means that »[s]ilencing co-exists with the potential for the expression that it seeks to abolish.« (Thiesmeyer 2003, p. 13) The way in which users make use of explicit silence to walk the line between saying and not saying poignantly highlights this co-existence and points to metadiscourse and ellipses as a way of navigating it, and of ambiguating one's participation in a potentially precarious exchange. She also notes that those who are subject to silencing may keep on inventing forms of expression that could »rupture their silencing« and »appeal to potential listeners« (p. 9) and that we should not assume a »static boundary between chosen and imposed silences« (p. 8). Facebook users appear to make a conscious choice of avoiding certain content, but at the same time indicate that this choice is a result of an imposed silence. Those who present themselves as silenced can use explicit silence, rather than refraining from communicating altogether, to ›rupture‹ the silencing that they experience so that in fact, they do not remain entirely silenced. Instead, users create ambiguity and vagueness deliberately and strategically, with both an in-group and an out-group in mind. The in-group is invited to resolve vagueness of content by completing ellipses and imagining and possibly exaggerating explicitly withheld content. The out-group is prevented from negotiating or indeed sanctioning the omitted content. The forms of explicit silence discussed above ambiguate along potential addressees in that they serve to withhold a proposition from the out-group, but signal shared experience to the in-group -including the shared experience of ›being silenced‹ -through participating in the communication, even without users stating what they might have otherwise had to say, and/or by inviting in-group members to use the absent content as a launchpad for their own thoughts. Anthonissen (2003) shows how news media navigate the challenge to publish content that is subject to censorship by highlighting the imposed restrictions and notes that »[r]eporting on the stipulations and effects of censorship constitutes an act of protest.« (p. 99) The various Facebook user comments on the danger of having their comments removed or of being barred from commenting fulfil the same function of highlighting and scandalising such practices. This scandalisation rests on the appreciation of an unrestrained public discourse and of the right to freedom of opinion which can be observed in the wider metadiscourse about silencing on part of more public New Right discourses (cf. Schröter 2015, 2019b, Anthonissen 2008. Anthonissen (2003) also observes that indicating but not spelling out ›forbidden‹ content shifts the »responsibility for fixing an interpretation from the author and the test itself to the reader«, so that »the author cannot be held accountable for all possible interpretations« (p. 102). The ambiguity created by indicating but not spelling out taboo content, and the indication of users' suppressed intentions to post something extreme both work in a similar way. Thus, explicit silence creates ambiguity and vagueness which enables users to carry out various communicative purposes while at the same time possibly suppressing certain content. This removes those who revert to explicit silence from the need to state and to negotiate content. Explicit silence operates at a meta-level; the omission of possible content is explicitly indicated. Just like in the wider discourses about silencing on part of the New Right, it becomes impossible to assess or to deliberate whether, how or why the omitted content should, or should not be stated (cf. Schröter 2019b). It thereby ambiguously invokes, but at the same time prevents a debate about norms applicable to public discourse. Instead, users can indulge in the scandalisations of norms they see themselves subjected to. Users can also indulge in a game of walking a line of ambiguity between saying and not saying and indicating, but not performing transgression of such norms.

Data access statement:
Queries about accessing the data analysed for the current study can be directed to the author.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4. 0/.