MITIGATION IN ONLINE INTERACTIONS ON CONTROVERSIAL TOPICS: A METHODOLOGICAL

A BSTRACT This article puts forward a methodological proposal for the pragmatic analysis of mitigation in online written interactions, specifically in discursive genres such as the digital press forums and Twitter. Mitigation is a rhetorical and pragmatic strategy linked to the face needs and communicative goals of the participants in the communication (Albelda Marco & Briz Gómez 2020). For this reason, drawing on the existing literature concerning the selected discursive genres and on an analysis of some contributions about a controversial Spanish tradition, namely the Moors and Christians festivals, this article highlights both situational and linguistic features to be taken into account when analysing mitigation in these genres, and proposes adaptations to Albelda Marco et al .’s (2014) methodology for analysing mitigation in spoken interactions. In doing so, this study presents a methodological proposal directly suitable for analysing mitigation in online interactions in discursive genres such as digital press forums and Twitter.


INTRODUCTION 1
The development and omnipresence of the Web in contemporary society has led to the emergence of new practices in the digital sphere.These practices, characterised by specific situational contexts, have aroused interest in linguistics and pragmatics, with research on (im)politeness and facework, among others (e.g., Graham & Hardaker 2017, Xie 2018).There is very prolific literature on digital genres and facework.In this overview, we focus on the genres discussed in this paper.Facework and (im)politeness have been studied on social networking sites (for a general overview, see, for example, Mancera Rueda 2015) and forums (see Mancera Rueda 2009, Lorenzo-Dus et al. 2011, Vigara Tauste & Hernández Toribio 2011, Weizman & Dori-Hacohen 2017, Sanmartín Sáez 2019 as examples of empirical research).In this paper, facework and (im)politeness are analysed through the study of mitigation.
Mitigation has been studied from different perspectives.In politeness studies (see Brown & Levinson 1978, 1987, among others), mitigation, as a strategy for reducing or compensating for threatening acts that challenge the speaker's or interlocutor's face, is seen as a way of expressing politeness (Briz Gómez 2012).Nevertheless, mitigation strategies are not limited to the expression of politeness, as they can also pursue other goals, linked to the efficacy and argumentative activity of the communication (Briz Gómez 2012).While Fraser (1980) defines mitigation as the reduction of the unwelcome effects of a speaker's speech act on the hearer, Meyer-Hermann (1988) relates this phenomenon to the expression of a low speaker commitment to the propositional content.For Caffi (1999), mitigation is seen as a relational concept that includes both the interactional goals and the relational needs of the participants in an interaction, while Schneider (2013) combines in his definition the degree of commitment assumed by the participants as well as their face needs.In Spanish research, although the inclusion of "face" in the definition of mitigation has been debated (Albelda Marco 2016), mitigation is defined as a pragmatic strategy linked to the face needs and communicative goals of the participants (Albelda Marco 2016, Albelda Marco & Briz Gómez 2020), and as a preservative force aimed at maintaining the set of assumptions related to the speaker's face attributed to the hearer (Albelda Marco & Estellés Arguedas 2021).In this paper, mitigation is seen as a strategy linked to the face needs and communicative goals of the participants.For this reason, we consider it particularly relevant to propose a methodology focused on the linguistic analysis of mitigation in interactions on controversial topics that involve the participants' face, taking the specific situational characteristics of online interactions into account.
As a pragmatic strategy, mitigation is not inherent to specific linguistic forms, but stems from the context and is achieved in different ways (Albelda ANU.FILOL.ESTUD.LINGÜÍST., 13/2023, pp.1-26. ISSN: 20141-26. ISSN: -1408.DOI: 10.1344/AFEL2023.13.1 Marco & Briz Gómez 2020).According to Caffi's (1999) distinctions, mitigation can appear by depersonalising the utterance source, by weakening the illocutionary force of the speech act or by modifying the propositional content.Regarding the modification of the propositional content ("bushes"), Caffi (1999) clearly differentiates between mechanisms that reduce the quantity (e.g., diminutives) and those that make the propositional content less precise (e.g., approximators), as in example (1).The mechanisms that weaken the illocutionary force are "hedges" and are obtained, for example, through the use of conditional and modal adverbs, as in (2).Finally, Caffi proposes the term "shields" for mechanisms of depersonalization, such as the use of modals, the first-person plural pronoun and impersonal constructions, as in example ( 3).
( When one gets nervous one acts that way.(Caffi 1999: 898) Mitigation has been studied in numerous discursive genres (section 2) and methodological proposals have been formulated for its pragmatic analysis in various genres, mainly oral.Consequently, this article considers the linguistic and situational features of two digital discursive genres, interactions on Twitter and in digital press forums, as well as their effects on communication (section 3) in order to propose a methodological proposal for the pragmatic analysis of mitigation in these kinds of online interactions (section 4), based on an adaptation of Albelda Marco et al.'s (2014) proposal.Moreover, since these genres usually host online controversies that reflect societal debates, this article is based on a specific case study of an online controversy, namely interactions about the Moors and Christians Festival.This Spanish tradition commemorates the confrontations of the Reconquista and has been the subject of controversy.The controversy concerns mainly the representation of Muslims and the participation of women on equal terms with men in groups, parades and roles traditionally reserved for men, as highlighted in anthropological research (Gisbert Gracia 2010, Martínez Pozo 2015).Therefore, this controversy is testament to wider debates concerning minority relations and the social roles of women in contemporary society.

PREVIOUS EMPIRICAL RESEARCH ON MITIGATION
Mitigation as a rhetorical and pragmatic strategy aimed at protecting the participants' faces in an interaction has been extensively studied in face-to-face oral interactions.In Spanish, Albelda Marco (2010) highlights the importance of considering the speech act and its illocutionary force, the situational characteristics of the act of communication (i.e., the contribution), and the implication/participation of participants' faces in the interaction, whereas Villalba Ibáñez (2020) proposes practical tests to identify mitigation phenomena.Others are interested in the relationship between mitigation, face and politeness.Briz Gómez (2003Gómez ( , 2005)), for instance, points out that mitigation can be seen as a linguistic strategy linked to the minimization of the propositional content in argumentative activities, as well as a communicative strategy linked to the protection of the participants' faces in an interaction, with or without politeness.This author draws attention to the possible dissociation of facework and politeness.This dissociation relates to contexts in which mitigation strategies are used to reduce the speaker's benefit or contribution, or to soften any disagreement with an interlocutor in strictly argumentative activities, which leads to facework (mainly related to the speaker's face) without necessarily leading to politeness activities (Briz Gómez 2005: 58).In (4), a child uses the diminutive -ito in rotito "small tear" to minimize the problem he is reporting to his mother.According to Briz Gómez, this use protects the speaker's face by minimising the importance of the tear in order to avoid his mother's recrimination (Briz Gómez 2005: 58).Consequently, the mitigation strategy is not used for politeness purposes to protect the interlocutor's face.However, example (5) illustrates a mitigation strategy linked to politeness, given that the use of poquito "a bit" softens a face-threatening act (you are distracted) for the interlocutor (Briz Gómez 2005: 57).Furthermore, mitigation has also been studied in specific discursive genres (e.g., Villalba Ibáñez 2017 on trials), and from the perspective of variation across discursive genres.For example, De Cock et al. (2018) conduct research on the presence of mitigation in different spoken discourse settings, while Albelda Marco (2018) carries out a study on the variation in both oral and written genres.This existing research on spoken and written genres shows that each genre has specific characteristics both in terms of the need to protect the participants' face according to their relationship and status, and in terms of linguistic mitigation strategies.
In recent years, research focuses on mitigation in digital discursive genres in light of the growing importance of this type of discourse (see, for example, Albelda Marco & Mihatsch 2017, Helfrich & Pano Alamán 2018, Moya Muñoz & Carrió-Pastor 2018).Some of these studies are based on the methodology developed by Albelda Marco et al. (2014) for the pragmatic analysis of mitigation in discursive corpora.This methodology has been developed within the framework of the Pan-Hispanic project ES.POR.ATENUACION, whose purpose was to establish a theoretical and methodological basis for the pragmatic analysis of mitigation in different Spanish and Portuguese corpora (Albelda Marco et al. 2014).However, this methodology is more suited to certain genres.For this reason, it has been mainly employed to analyse mitigation in oral corpora and, to a lesser extent, in written corpora (see, for example, Pano Alamán 2018Alamán , 2020)), in which case some adjustments were required when looking at genres with peculiar situational features, as highlighted by Pano Alamán (2018) in her study on mitigation in comments on Spanish digital articles.However, there has been no methodological proposal for the analysis of mitigation in this kind of discursive genre.Even though the literature on mitigation also offers other proposals based, for example, on a sociolinguistic perspective (Alfano Rodríguez 2015), this article presents a proposal based on Albelda Marco et al.'s (2014) methodology because of its pragmatic perspective and its focus on Spanish corpora.Therefore, the following section presents Albelda Marco et al.'s (2014) methodology, before considering the characteristics of the digital discursive genres examined in this article and proposing a specific methodology for analyzing mitigation in these genres.et al.'s (2014) methodology is divided into feature clusters related to linguistic, structural, expository and situational aspects, as summarised in Appendix 1.

Albelda
The linguistic analysis includes the function of mitigation, namely the reason for its presence, and the mitigation strategy, which can be morphological, lexical, syntactic, or prosodic, paralinguistic and gestural.The structural aspects are composed of the textual typology and the discursive position of the mitigation strategy in relation to the element causing the mitigation and the attenuated linguistic element.The expository aspects analyse the relationship between the propositional content and the participant's face, and the illocutionary force of the speech act.As for the situational aspects, the variables comprise information on the social and communicative circumstances of the interaction (the topic, the purpose, the physical setting, the register and the type of communicative activity) and the characteristics and relationship between interlocutors (experiential and social relationship between interlocutors, age, gender, educational attainment, origin and language).
As can be deduced from the description of the situational aspects, the analysis of mitigation in a digital press forum and on Twitter requires the adaptation of these variables, in line with Villalba Ibáñez's (2015) conclusions on the applicability of Albelda Marco et al.'s (2014) methodology in oral trials.In this article, we propose reflecting on the applicability of this methodology in digital interactions, specifically in digital press forums and on Twitter, digital spaces that share the feature of allowing their users to express opinions on news, as we discuss in the next section.

A general description of Twitter and online newspaper comments
Twitter is a microblogging platform characterized by conciseness and multimodality since it enables users to post messages of up to 280 characters (140 characters until November 2017) and to insert pictures, videos, links and so on (García Aguiar 2019).Moreover, this platform makes it possible for users to subscribe to the profiles of other users they want to "follow," and to like and republish tweets, and has developed a set of typographical conventions in order to tag someone (@) and to identify the topic of a tweet (#) (Zappavigna 2011).These features have led, according to Zappavigna, to a type of "public conversation" characterised as a "multiparty, temporarily fluid and highly intertextual" conversation (Zappavigna 2011: 790).
Concerning the topics, even though initially Twitter asked users ("what are you doing?") for their status updates, a variety of themes can be observed, including personal thoughts and experiences, professional and information sharing, as well as exchanges with well-known people and discussions on news items (García Aguiar 2019).However, Twitter differs from other online platforms where users seek contact and sociability with friends, because this type of microblogging is mostly used to inform, keep informed and share opinions with other users on news topics of common interest (Mancera Rueda & Pano Alamán 2013, Oz et al. 2018, Sifianou & Bella 2019).This supports Zappavigna's conclusion that Twitter is the place you go when you want to find out what people are saying about a topic right now and in order to involve yourself in communities of shared value that interest you in this given moment.(Zappavigna 2011: 803-804) This latter characteristic is shared with the genre of comments on digital press forums.
Although there is no consensus on how to categorize comments on digital press articles, they can be considered as a sort of forum included in the newspaper (Moya Muñoz 2015).This position is adopted in various studies on comments to the digital press.In this article, we therefore use the term "digital press forum" to refer to this space, which is not a forum in the traditional sense of the term, but a comments section at the end of an article for readers to meet and share opinions (Vigara Tauste & Hernández Toribio 2011).These comments can be seen as part of a hybrid genre composed of the traditional press article on the one hand, and interactional discourse -the comments-on the other, since the latter stem from the topic and content of the article (Bruce 2010).Moreover, comments convey and collect opinions and knowledge based on individual stances and experiences from the reader's private sphere (Johansson 2017).In addition, depending on the newspaper, the commentators sometimes have the option or are required to create a profile to be able to participate in the forum, or to subscribe to another account (their Facebook account, for example), but there ANU.FILOL.ESTUD.LINGÜÍST., 13/2023, pp.1-26.ISSN: 2014-1408.DOI: 10.1344/AFEL2023.13.1 is generally very limited information about the users and it is not possible to follow them.
As shown in this section, the microblogging platform Twitter and the digital press forums share the feature of allowing their users to express opinions on news.We now address the specific features of these two digital genres using a case study of an online controversy in these genres.

Case study of an online controversy in digital press forums and on Twitter
The methodological proposal for the study of mitigation in online interactions presented in this paper is based on previous research on mitigation in spoken genres (section 2) and on the selected discursive genres (section 3.1).In doing so, it fills a gap in that it focuses on mitigation in online genres.The methodology is illustrated by means of an empirical analysis of data about the Moors and Christians Festival, chosen because they contain the main characteristic criteria of the digital genres analysed.Starting with a search using hashtags and keywords about the festival with the Woodpecker tool (reference removed for anonymization purposes) between October and November 2019, the corpus has been further narrowed down manually according to three criteria: the date of publication (from 2010 to 2019), the topic for discussion (the Moors and Christians Festival), and the presence of comments and tweets in response to the initial article or tweet.Table 1 shows the number of tweets and comments collected and the size of each sub-corpus.In the following sections, we have reproduced the contributions as they appear without correcting language errors.However, for privacy reasons, we chose to anonymise the data for the citation, except for the public figures.Therefore, in order to be able to identify the relationship between interlocutors, we decided to indicate whether the users chose to show their first name or full name, or to employ a pseudonym.This is why we have substituted the usernames for "first name", "full name" or "pseudonym", followed by a number that distinguishes between interlocutors.Moreover, at the beginning of the contributions, we have included [DPF] (for digital press forum) or [TWI] (for Twitter) to make it easier to identify the type of contribution.The translations from Spanish are proposed by the author of this article.

Characteristics of the communicative situation
As Albelda Marco (2010) notes, situational characteristics of communication affect people's behaviour and the presence of mitigation.This section therefore looks at the situational features of the digital press forums and Twitter.
Digital press forums, such as discussion forums in general, present some characteristics already observed in other digitally mediated genres (see Herring 1999, for example).The communication is public, not in-person and allows for some anonymity because, even though users usually need to sign up to participate and post comments, publicly available personal information is generally limited to a pseudonym and, in some cases, a picture that rarely reveals the user's identity.Moreover, this form of digital communication is characterised by thematic fragmentation and sub-topics derived from the initial topic, and personal attacks on an interlocutor, as in ( 6) below.In this excerpt, Pseudonym1 begins the discussion thread by responding to the content of an article about the participation of women in the Moors and Christians Festival.His/her comment receives a reply from Pseudonym2 (about the prohibition of festivals or groups in society in general), who in turn gets a response from Pseudonym3 (who attacks him/her personally by using the word cortito "simple-minded").These characteristics can also be observed in the interactions on Twitter, as shown by Honeycutt and Herring (2009) and illustrated in excerpt (7).Albert Rivera, a Spanish politician, starts a topic tweeting about his participation in the Moors and Christians Festival in Elda.This first tweet receives some feedback about the use of the word "Moors" and about Spanish society in general from The digital genres studied in this paper allow for multi-user responses regarding the content of an article (in digital press forums) or the content of a post (on Twitter) and are liable to disruption and off-topic digression, particularly when there are a large number of participants (Honeycutt & Herring 2009).
As a result of this multi-user participation, the analysis of asynchronous group communication stresses that "multiple responses are often directed at a single initiating message, and single messages may respond to more than one initiating message" (Herring 1999: n.p.).Consequently, this kind of online communication can offer various participatory frameworks with different kinds of addressees (Marcoccia 2004b, Moya Muñoz 2015), which leads Graham and Hardaker (2017) to note that Goffman's (1981) observations are a good starting point to consider the specific frameworks of online discussions.
According to Goffman (1981), the participatory framework of everyday interactions can involve ratified and non-ratified participants.Ratified participants are interlocutors and comprise the speaker and ratified hearers (addressees or not), whereas non-ratified participants correspond to the overhearers and are composed of bystanders and eavesdroppers.Drawing on Goffman's (1981)  framework of specific online genres.Marcoccia, for example, concludes that "the opposition between ratified participants and bystanders does not seem very well suited to newsgroups" (Marcoccia 2004a: 140), because the eavesdropper, whose identity and presence are unknown to the speaker, is simultaneously seen as a ratified participant.
Later, in her analysis of the participatory framework in YouTube interactions, Dynel discusses Marcoccia's (2004a) observations and proposes considering that "in the case of any public/media discourse, there is hardly any possibility of unratified hearers being present" (Dynel 2014: 46).She explains that publicising a turn, a speaker must be mindful of the fact that it will be widely available and may potentially reach even those to whom the speaker does not wish to communicate a given message.(Dynel 2014: 46) Therefore, according to Dynel, any individual who reads some contributions becomes a participant (or at least a passive participant) and is considered a ratified non-addressed hearer (a third party) (Dynel 2014: 48).In the category of ratified hearers, Dynel consequently distinguishes between the addressee ("a hearer to whom the speaker directs a given utterance") and the third party ("a ratified listener to whom an utterance is not addressed but who is fully entitled to listen to it and make inferences") (Dynel 2014: 40).
This observation is related to Scannell's (1991) notion of "double articulation," which Bou-Franch et al. (2012) also apply to YouTube interactions.According to the latter authors, the open and public nature of YouTube interactions generates a double level of reception characterised by, on the one hand, one-to-one interactions or intergroup discussions and, on the other hand, a large audience of users who participate passively by reading the contributions (Bou-Franch et al. 2012).As a consequence, various researchers regard this kind of communication as hybrid, blending characteristics of mass communication because of its public nature as well as interpersonal communication, given that it makes individual exchanges between internet users possible (Marcoccia 2004a(Marcoccia , 2004b;;Moya Muñoz 2015).
To summarise, the digital genres studied in this article are characterised by public communication that allows for interpersonal interactions between anonymous and not acquainted interlocutors, as well as with public and famous figures and groups.Moreover, thematic fragmentation and secondary discussions are typical of these genres, due to the constant participation of multiusers.These features can affect people's behaviour and, consequently, the presence of mitigation in these genres, this is why section 4 discusses how to take these observations into account for the analysis of mitigation.

Characteristics of digital language
Internet-mediated communication is influenced not only by the kind of communicative activity but also by the characteristics of the platforms on which it takes place.Coesemans and De Cock (2017) show that politicians use linguistic strategies on Twitter that enable them to comply with the character constraints (section 3.1) while pursuing their professional and communicative goals.In interactive genres such as digital press forums and social networking sites, written communication usually shares features with the spoken language (Mancera Rueda & Pano Alamán 2013, Ridao Rodrigo & Rodríguez Muñoz 2013).In this respect, Ehret and Taboada (2020) show that the comments in digital press forums share some features of spontaneous conversation (such as informal and emphatic language).However, they also emphasise differences in relation to spontaneous conversation due to the written and asynchronous nature of the comments in digital press forums.For this reason, all these authors regard this kind of communication as a new modality, alongside written and spoken communication.
As Álvarez (2011) mentions in her analysis of interactive blogs from Spanish newspapers and TV channels, the language used in these kinds of genres is usually characterised by lexical, grammatical and orthographic features, as well as typographic features, which are particularly interesting for a rich pragmatic interpretation of online messages.Typographic features are textual deformations such as capitalisations, repetitions of letters and punctuation marks, and emojis (Yus Ramos 2014).These features guide the interlocutor towards the correct interpretation of the propositional content and also of the speaker's attitudes and emotions in relation to his/her message (Yus Ramos 2005Ramos , 2014)).
Emojis have some functions similar to those of non-verbal behaviour (Derks et al. 2008) and communicate emotions and non-emotional attitudes traditionally expressed through facial expressions and body language (Dresner & Herring 2010, Yus Ramos 2014).Moreover, these typographic elements can also signal the illocutionary force, strengthening or mitigating a complaint, utterance or request that could threaten the interlocutor's face (Dresner & Herring 2010, Yus Ramos 2014).In example (8), the winking smiley can be interpreted as a strategy to mitigate the illocutionary force of the suggestion made to the politician through the depersonalising structure hay que, also used as a mitigation strategy in Spanish.In punctuation studies, González García's presentation at the Congreso Internacional de Atenuación Lingüística in 2016 highlighted the fact that ellipsis can be used in posts on Facebook and WhatsApp for numerous purposes, including the reduction of the illocutionary force of a speech act.This could also be the case in example ( 8), where the use of the smiley and ellipsis both contribute to mitigating the suggestion.Moreover, Figueras Bates (2014) observes that the repetition of punctuation makes it possible to communicate propositional attitudes and, for example, the repetition of exclamation marks can reinforce the strength of the message and solidarity with the interlocutor (Figueras Bates 2016).This also seems to be the case in example ( 8) with the duplication of the final exclamation mark.
Thus, the communication produced in digital press forums and on Twitter is characterised not only by lexical, grammatical and orthographic features, but also by typographic particularities such as textual deformations and emojis.Given the importance of these typographic features for the pragmatic interpretation of online communication, it seems essential to include them in the analysis of the mitigation strategies in online interactions.

Specific features of digital press forums and Twitter for the analysis of mitigation
The case study proposed in the previous section points out situational and linguistic features of the digital interactions discussed in this article.On the one hand, the analysis of the communicative settings shows that some situational features are determined by the platform itself and the type of communicative activity in which the speakers are involved.The public nature of the communication and the purposes of these digital places allow for multi-users opinions in response to the content of an article or post, as well as different interactional moves.Consequently, different types of participatory framework can be observed, as the speaker can implicitly or explicitly address a specific interlocutor or the large audience of ratified hearers.On the other hand, some information is unknown about the interlocutors due to the frequent anonymity of the users and the lack of knowledge among them.As mitigation is affected by the situational characteristics and the implication of the participants' faces, the use of situational variables that capture these specific features and their possible influence on mitigation is important.Moreover, the analysed interactions also present linguistic features, such as typographical elements that communicate emotions and attitudes traditionally expressed through prosodic and paralinguistic elements.Given that mitigation is also affected by the illocutionary force of the speech act, the influence of these elements also has to be considered when analysing this pragmatic phenomenon in digital interactions.Therefore, the next section examines how to take these features into account for the study of ANU.FILOL.ESTUD.LINGÜÍST., 13/2023, pp.1-26.ISSN: 2014-1408.DOI: 10.1344/AFEL2023.13.1 mitigation in online interactions on the basis of the methodology developed by Albelda Marco et al. (2014).

METHODOLOGICAL PROPOSAL FOR A PRAGMATIC ANALYSIS OF MITIGATION IN ONLINE INTERACTIONS
The methodology for a pragmatic analysis of mitigation developed by Albelda Marco et al. (2014) suggests analysing mitigation on the basis of some feature clusters related to linguistic, structural, expository and situational aspects.Moreover, we have seen in section 2 that mitigation is affected by the speech act and illocutionary force, by the situational characteristics and by the implication/participation of the participants' faces, whose attributes are specific to each discursive genre (Albelda Marco 2010), as illustrated in section 3 through a case study of an online controversy in digital press forums and on Twitter.This section presents the variables we propose in order to take into account the situational characteristics (4.1), the speech act and illocutionary force (4.2) and the implication of the participants' faces (4.3) in the analysis of mitigation in online interactions.

Situational characteristics
The main changes we propose in relation to Albelda Marco et al.'s (2014) methodology, which was particularly employed for the analysis of oral corpora, affect the situational variables because, in online interactions, this information remains sometimes unknown to the users (e.g., the age and gender of the interlocutor) and is sometimes defined by the communicative setting (e.g., the type of communicative activity).
As observed in Section 3.2.1, the interactions in digital press forums and on Twitter are of a public nature and enable multi-user participation, though concrete threads do not necessarily contain many active participants.As a result of the possibility of multi-user participation, various participatory frameworks including different kinds of addressees can be observed.For this reason, taking into account Scannell's (1991) notion of "double articulation," applied by Bou-Franch et al. (2012) to another kind of online interaction, we propose two variables in order to identify the participatory frameworks of the interactions: the type of contribution and the relationship between the interlocutors.
Firstly, this methodological proposal suggests identifying whether a contribution starts a discussion thread, or responds to the content of an article, to a previous comment from another user or to the entire discussion, as detailed in The contribution of example ( 9) is analysed as 2 (responding move to another user's contribution) because the speaker reacts to what the politician said, as Fullname4's response in example ( 11) because the speaker replies to the last sentence of the previous comment reproduced in example ( 10).The latter is analysed as 4 (responding move to the entire discussion) because the speaker replies to the discussion as a whole. (9) [TWI] Fullname3 12:40AM Jun 4, 2016 Replying to @Albert_Rivera Perdona Albert, en Alicante gusta nombrarnos cómo alicantinos siempre que se menciona "valenciano".Consejo de simpatizante.'Excuse me, Albert, in Alicante we like to be called "Alicantinos" whenever "Valencian" is mentioned.Advice from a supporter.'(10) [DPF] Pseudonym5 06.abr.2012 23:15 #43 Pongamos un fallero mayor de Valencia, un rey de carnaval de Tenerife, a una Jesucristo en la recreación de la Semana Santa…..por què no? Para que respetar las tradiciones?En mi municipio hay comparasas formadas exclusivamente por chicas, y bien contentas y horgullosas que estan ellas, no les hacen falta hombres….Lo que no entiendo es por qué algunas chicas insisten tanto en formar parte de algo que no las quiere…no se, en fin.'Let's put a man as fallera mayor [head fallera, lady of honour] in Valencia, a king of carnival in Tenerife, a woman as Jesus Christ in the re-enactment of the Holy Week….whynot?Why respect traditions?In my town, there are troupes formed exclusively of girls, and they are very happy and proud, they don't need men….What I don't understand is why some girls are so insistent on being part of something that doesn't want them… I don't know, anyway.' (11) [DPF] Fullname4 07.abr.201201:34 #44 Creo que no se trata de insistir en algo que te rechaza, si no de tener la opción de poder hacerlo.Pienso que también esta bien que alguien quiera empezar algo, en este caso las mujeres formar parte de un colectivo hasta ahora machista, por que no?A caso no pagan ellas la fiestas?Y si la cuestión va a ser siempre sobre porque ir en contra de lo que te rechaza… Entonces adiós a los derechos y a la evolución, a la constitución y a todo lo que conocemos por democracia.Por otro lado me parece absurdo que le den tanta importancia a algo que me parece cateto en el sentido literal de la palabra.Que interés tiene algo que tiene como base el rechazo?Es una fiesta pagana que termina con alcohol y que seguramente tenga como fundamento hacer un club de hombres.Pues que se vallan a cazar si quieren ir de machos… NEXT! ANU.FILOL.ESTUD.LINGÜÍST., 13/2023, pp.1-26.ISSN: 2014-1408.DOI: 10.1344/AFEL2023.13.1 'I believe that it isn't a question of insisting on being part of something that rejects you, but of having the option to be able to do it.I also think that it's great that someone wants to start something, in this case that women want to join a hitherto macho group, why not?Don't they also pay for the festival?And if the question is always going to be about why go against what rejects you… Then, goodbye to rights and evolution, the constitution and everything we know thanks to democracy.Furthermore, I think it's absurd to attach so much importance to something bumpkin in the literal meaning of the word.What interest is there in something that's based on rejection?It's a pagan festival that ends with alcohol and that surely is based on the creation of a men's club.Well, they can go hunting if they want to behave like machos… NEXT!' Secondly, we propose characterising the relationship between the interlocutors on the basis of the recipient of the contribution and his/her (possible) identification.We have seen in section 3.2.1 that many contributors post anonymously or using a pseudonym, with the result that participants often do not know the person to whom they are reacting in this kind of interaction.An exception is the case where the users communicate directly with public figures or groups (on Twitter) or when they react to the digital newspaper they are reading (in digital press forums).Moreover, like Dynel (2014), we consider that, given the public nature of these interactions, any reader of the contributions becomes a passive participant and, then, is one of the ratified hearers.This leads us to distinguish between the addressees and the large audience of ratified, nonaddressed hearers.Consequently, instead of identifying previously shared experiences and shared knowledge between interlocutors on the one hand, and social and functional relationships on the other, as in Albelda Marco et al.'s (2014) methodology, this proposal suggests the options listed in Table 3.In light of the notion of "double articulation" (Scannell 1991), a stranger or a public figure who speaks to a specific addressee (cases 3 and 4) is in this proposal considered to be aware of the public nature of his/her contribution and, consequently, to know that he/she is being read by a wider audience.For example, Fullname3's response to Albert Rivera's tweet in example ( 9) is analysed as 3.2 (a stranger addresses a public interlocutor) because it reacts to something published by a public figure, and Pseudonym5's comment in example (10) is analysed as 2 (a stranger speaks to a large audience of ratified hearers) because the speaker responds to the entire discussion and then to an indefinite audience, whereas Fullname4's comment in example ( 11) is analysed as 3.1 (a stranger addresses another stranger), as the speaker replies to the previous comment posted by an unknown user.
Examination of the excerpts also reveals thematic fragmentation and the presence of sub-topics derived from the initial topic.For this reason, we propose detailing the variable "topics", depending on the topics in the corpora studied, considering the fact that contributions can relate to the initial topic, can be broadened to society in general and can even consist of personal attacks.We argue that attacking an interlocutor or expressing a personal stance that affects societal groups involves the participants' faces in different ways (the speaker's and interlocutor's faces in the former case, and the speaker's and outsider's or group's faces in the latter case).This is why this variable makes it possible to identify which topics are favourable for mitigated speech acts.Table 4 is an example of a thematic classification based on an empirical analysis of the corpus about the Moors and Christians Festival.Finally, section 3.2.1 also reveals that online interactions are marked by different moves between the contributions, as is the case with offline interactions.However, this observation enables us to suggest another distinction for the variable "purpose of the contribution," instead of identifying interpersonal or transactional purposes, as in Albelda Marco et al.'s (2014) methodology.Since the purpose of the speech act affects the management of the participants' face and consequently the presence of mitigation in these conflictual interactions, this proposal suggests the options listed in Table 5.
1. To answer a question or to react to a neutral statement; to express a personal stance 2. To contradict a previous statement 3. To support a previous statement 4. To attack the interlocutor, newspaper, etc. (without expressing an opinion) TABLE 5. Purposes of the speech act in online interactions.
Examples ( 9) and ( 11) are analysed as 2 (to contradict a previous statement or opinion) because the speaker contradicts something expressed by his/her interlocutor, whereas example (10) is 1 (to express a personal stance), as the speaker expresses his/her personal stance on the topic presented in an article without explicitly relating it to another interlocutor's stance.
On the whole, consideration of these situational characteristics makes it possible to analyse the contexts in which mitigation appears in the discursive genres discussed in this article.These variables enable us to observe whether ANU.FILOL.ESTUD.LINGÜÍST., 13/2023, pp.1-26.ISSN: 2014-1408.DOI: 10.1344/AFEL2023.13.1 mitigation functions more frequently in some situations, according to the specific features of this kind of interaction.

Speech act and illocutionary force
The speech act and illocutionary force affect the presence of mitigation.As highlighted in Caffi's (1999) distinctions, the reduction of the illocutionary force is a form of mitigation.In the methodology developed by Albelda Marco et al. (2014), the linguistic mechanisms are mentioned in the context of the variable "mitigation strategies" and the identification of the illocutionary force is another variable situated in the expository aspects (cf.Appendix 1).In this methodological proposal, we propose an analysis that starts with the identification of the type of mitigation and then names the strategies that contribute to this purpose.
Moreover, as observed in section 3.2.2concerning this kind of digital communication, typographic features such as strategic use of punctuation and emoji can reduce the illocutionary force and mitigate speech acts that could threaten an interlocutor's face during the interaction.For this reason, following the example of Albelda Marco et al.'s (2014) proposal, which includes the analysis of non-linguistic elements (gestural, for example), we propose a variable that encodes these typographic features as written code elements, in parallel with the prosodic, paralinguistic and gestural elements, specific to the oral code, from Albelda Marco et al.'s (2014) methodology.This could enable us to identify that there is a reduction of the illocutionary force of a directive act by means of written elements (emoji and ellipsis) in the speech act from the above-mentioned comment (8).
This methodology then proposes introducing the typographic elements typical of digital communication into the analysis and including the identification of the speech act directly with the analysis of the strategies of reduction of the illocutionary force, which makes it possible to identify directly which types of speech act are more usually mitigated in this kind of digital interaction.

Participants' face
Finally, face is the last factor that affects the presence of mitigation in each discursive genre (section 2).As mentioned in section 3.2.1, the situational characteristics of digital press forums and Twitter offer various participatory frameworks and Goffman's (1981) observations are a good starting point to consider them in digital communication (Graham & Hardaker 2017), as demonstrated by research that adapts Goffman's proposal to specific online polylogues (Marcoccia 2004a, Dynel 2014 Accordingly, since mitigation can affect different kinds of participants in this digital context, we suggest introducing Goffman's (1981) observations about the participatory framework, more specifically the concept of ratified participant, and linking them to the analysis of the management of the participants' face during the interaction.Thus, this methodology proposes the changes presented in Table 6 in order to correspond as closely as possible to the situational characteristics of these genres.
Distinctions made by Albelda Marco et al. (2014) Our proposal Relationship between the propositional content and facework: − Mitigation of a negative content for a ratified participant in the interaction (speaker or listener).− Mitigation of a negative content for a nonratified participant in the interaction (or an institution).− Mitigation of a positive, flattering content for a ratified participant in the interaction (speaker or listener).
Relationship between the propositional content and facework: − Mitigation of a content in which the speaker's face is involved.− Mitigation of a content in which an explicitly addressed interlocutor's face is involved.− Mitigation of a content in which an implicitly addressed interlocutor's face is involved.− Mitigation of a content in which an outsider's or group's face is involved.TABLE 6. Changes from Albelda Marco et al.'s (2014) methodology concerning the relationship between the propositional content and facework.
The distinction between explicitly and implicitly addressed interlocutors is based on the cues given by the speaker to identify the interlocutors whom he/she is addressing.Cues such as the use of vocatives, forms of address and the position of the contribution in the discussion thread (when the platform explicitly shows that a user is replying to another user) lead to the analysis of an explicitly addressed interlocutor, whereas an implicitly addressed interlocutor is observed when the content of the contribution is presented as the only cue of a response to another user.Furthermore, as the relationship between the propositional content and facework is closely linked to the functions of mitigation, we suggest combining the analysis of the interlocutors' face with the function of mitigation, instead of analysing them separately, as detailed in For example, following this methodology, the mitigated speech act in ( 12) is analysed as 2.1 (prevention because an explicitly addressed interlocutor's face is involved) because the speaker contradicts something expressed by a previous participant, who is explicitly addressed using the number of his/her comment (#11).The speech act in ( 13) is analysed as 2.2 (prevention because an implicitly addressed interlocutor's face is involved), given that the speaker mitigates his/her criticism of previous interlocutors' stances by introducing the verb I think, which lets him/her present the criticism as a personal belief, but does not explicitly mention to whom he/she is responding.In this example, in spite of the mitigating effect, the criticism remains present, but the mitigation strategy allows the speaker to express the fact that his/her belief is personal and not necessarily shared by others.This reduces the possible threat to the interlocutor's face and, for this reason, it is analysed as a prevention strategy, despite the fact that other elements of the utterance may still maintain a threatening effect.Moreover, the proposed methodology considers that, when the interlocutor's face is threatened, the speaker's face is also affected, whereas when the speaker's face is threatened, the interlocutor's face is not necessarily affected.Finally, the use of I think in ( 14) is also analysed as 1.1 (self-protection because the speaker's face is involved), because, as in the previous example, the criticism is not cancelled out, but the use of this verb enables the speaker to reduce his/her responsibility by claiming uncertainty concerning factual information that is debatable.(12) [DPF] #11 hombre, decir que se cuida el detallen la vestimenta es verdad, pero decir se cuida el rigor histórico en la vestimenta en Alcoy es no haber ido nunca a ver las fiestas.
'#11 Well, to say that they take care in the detail of the clothing is true, but to say that they take care in the historical exactitude of the clothing in Alcoy entails never having attended the festival.'(13) [DPF] Creo que estáis exagerando lo que son esas fiestas y esas comparsas.
'I think you're exaggerating what this festival and these troupes are.' (14) [DPF] Creo que estan haciendo un rodaje del desembarco de normandia de la segunda guerra mundial y las feministas han hecho una protesta de que en el rodaje tiene que haber 50 % de mujeres en la escena del desembargo o sino pararan el rodaje y tiene visos de que lo puedan conseguir.'I think they're making a film of the Normandy landings in the Second World War and feminists have protested that the film must have 50 % women in the landing scene, otherwise they'll stop the filming and it seems that they can pull it off.' Since mitigation is a strategy aimed at protecting the participants' face by reducing the possible negative effects of the communication or the speaker's commitment (Albelda Marco & Briz Gómez 2020), this methodology argues that analysing the function and relationship between the propositional content and facework together makes it possible to observe directly whether this pragmatic phenomenon appears more frequently in some contexts than in others.

CONCLUSION
This article aims to present a methodological proposal for the pragmatic analysis of mitigation in specific online discursive genres that allow their users to express opinions on news topics: interactions in digital press forums and on Twitter.The existing literature on mitigation reveals important studies of its characteristics, its presence in specific discursive genres (written, oral and online), and a methodology for its pragmatic analysis (Albelda Marco et al. 2014).However, because of the particularities of the digital genres studied in this article, it seems necessary to reflect on the appropriateness of Albelda Marco et al.'s (2014) methodology for the pragmatic analysis of mitigation in these digital genres.
The analysis of authentic contributions has enabled us to stress situational features (such as the participatory framework, characterised by the "double articulation" of the communication, the relative anonymity and the thematic fragmentation), as well as linguistic features (the typographic elements) of these digital genres.As mitigation depends on the speech act and illocutionary force, the situational characteristics and the implication of the participants' face (Albelda Marco 2010), these observations have led to the formulation of a methodological proposal for its study in these genres, based on Albelda Marco et al.'s (2014) methodology.We propose variables that consider the thematic fragmentation, the relationships between contributions and interlocutors, and the diversity of the participatory frameworks in the analysis of the situational characteristics of the mitigated speech act.We also argue for considering the diversity of the participatory framework for the analysis of the function and the implication of the participants' face.Finally, we include the typographic aspects in the analysis of the mitigation strategies.

1 .
A public figure or group speaks to a large audience of ratified hearers 2. A stranger speaks to a large audience of ratified hearers 3. A stranger speaks to (a) specific addressee(s) 3.1.(An)other stranger(s) 3.2.(A) public interlocutor(s) 4. A public figure or group speaks to (a) specific addressee(s) 4.1.(A) stranger(s) 4.2.(A) public interlocutor(s)TABLE 3. Types of relationships between interlocutors.

1 .
The festival itself 2. Spanish society 3.An attack on an interlocutor 4. Others TABLE 4. Topics of the speech act on the basis of the excerpts.

TABLE 1 .
Corpus of the case study on the Moors and Christians festivals.
who may show their real names or choose pseudonyms that hide their identities to a greater or lesser extent.

TABLE 2 .
Types of contributions in online interactions.