Introduction

The comment sections of online news sites have become a political arena for digital democracy thanks to remarkable technological advancements. Readers typically engage with contemporary new media by responding and commenting on news items. KhosraviNik and Esposito (2018) have pinpointed that comment sections serve as a communicative social media space that allows readers or non-elite text producers to support or oppose the discourse of established institutions. Online reader comments have thus evolved as influential fora of public discourses through which ordinary people can make their voices heard by interacting and engaging with a news item or other readers (Badarneh & Migdadi, 2018). Reader comment sections are primarily presented as a new system of interactivity that promotes public discussion and can be used to gauge reader involvement. However, such sections usually turn into platforms for the dissemination of offensive and disparaging language rather than serve the purpose of civil debate (Dordević, 2020). This trend is probably driven by the anonymity of computer-mediated interactions that technically occur in the absence of physical bodies and faces. Readers are hence encouraged to engage in more open communication than face-to-face interactions (Herring, 2001, 2004). Put differently, most of the contextual elements of face-to-face interactions, such as gestural, prosodic, and social, are missing from the comments and responses posted by readers on digital media. Jones, Chick, and Hafiner (in Trajkova & Neskhovska, 2018) assert that such computer-mediated communication is now the most widely used form of discursive interaction for the achievement of social or political goals, the performance of particular social identities, and the reproduction of specific social relationships.

The participatory nature of comment sections on online news websites enables readers to communicate their opinions openly. Thus, comment sections are often used to express support or resistance to a particular political issue, figure, or group. The communicative practices of comment sections of online news and other digital platforms such as Facebook, Twitter, and YouTube have prompted the modernization of political communication. As a result, the role of new media in political communication has become a topic of research in language and politics (e.g., Boyd, 2014; McChesney, 2000; Norris, 2001; Trajkova & Neskhovska, 2018; Towner & Dulio, 2012; Velazquez & Rojas, 2017; Vaccari, 2008). Numerous studies (e.g., Dordević, 2020; Erjavec, 2014; Erjavec & Kovačič, 2012; Zannettou et al., 2020) have also demonstrated that the significant growth of comment sections in online news sites has resulted in increasing rates of media criticism.

Readers express stances freely and openly when they post comments, claiming linkages to their comments, the entities mentioned in them, and their interlocutors or audiences. Thus, the stances taken by online readers in their comments significantly shape public opinion, particularly with respect to political issues and ideologies. Commenters may use particular linguistic expressions for specific pieces of information to affect public opinion or spread a particular ideology. Du Bois (2007) argues that stance-taking represents one of the most important functions of language because positions can accord value to objects, locate social actors to the objects, adjust the alignment between stance-takers, and invoke socio-cultural value systems. Stance can be conceptually apprehended from several discrete terms studied from an intersection of subareas of linguistics: corpus-based, systemic-functional, discourse-analysis, cognitive, sociological, and interactional.

Studies on online criticism related to news-based comments have generally analyzed English, predominantly American and British, and some other European languages (see Boyd, 2014; Bruce, 2018; Dordević, 2020; Erjavec & Kovačič, 2012; Fangen & Holter, 2020; Tenenboim & Cohen, 2015). Scant studies have attended to Indonesian criticism, particularly in the context of politics. The present investigation analyzed reader comments posted on political news items as a form of social interaction in a digital world. More specifically, it examined readers’ engagement with online news in the form of negative criticism they post in the news comment section. To this end, it employed a corpus of comments concerning the current Jakarta governor Anies Baswedan’s handling of the COVID-19 outbreak posted by readers on Detiknews, a pioneer of online journalism in Indonesia. Founded in 1998, Detiknews was the first online media in Indonesia. It operates autonomously without any print media as its parent. In terms of popularity, Detiknews is the most visited news and media publishing website in Indonesia (similarweb, 2022). Detiknews generally reports on events supported by interviews from particular sources and does not try to tell readers what to think on specific issues. Through this strategy, Detiknews hopes to maintain its neutrality, including toward the government. Nevertheless, in the case of the 2017 Jakarta gubernatorial election, Detiknews has tended to quote more sources from Anies Baswedan’s side than it has from his opponent Ahok’s side (see Rumata, 2017).

News reports on the Jakarta governor’s policies or actions often become the most commented items on the Detiknews website. The comments mainly comprise negative criticism posted by readers espousing different political views. Such comments probably occurred because of the historical background of the gubernatorial electoral system and Anies Baswedan’s political career. Juditha (2017) asserted that the 2017 Jakarta local gubernatorial election displayed an unprecedented enormous amount of hate speech in Indonesia’s local election history, resulting in the longest political polarization in Indonesia. However, some analyzed comments in the present research’s data can count as hate speech, and others cannot because each has a particular definition. Previously, the phase of political polarization in Indonesia began in 2014 (see Power & Warburton, 2020). In the 2014 and 2019 presidential elections, the contest between Jokowi and his opponent, Prabowo, is said to have rekindled the latent Islamic–pluralist division/split. Prabowo represented a conservative Islamic stream, whereas Jokowi represented the pluralist stream. Prabowo pursued a religiously themed campaign as he saw that Jokowi’s secular political orientation was vulnerable to this. This campaign was then used to defeat Jakarta’s popular governor, Basuki Tjahaja Purnama (known as Ahok), in the 2017 gubernatorial election. Islamist groups opposed Ahok, a Christian, ethnic Chinese Indonesian, Jokowi ally. Ahok’s opponent, Prabowo-aligned Anies Baswedan, joined forces with the hardline Islamist groups opposed to Ahok. Before the gubernatorial election, Anies was popularly known as a pluralist scholar, yet he came to support a conservative Islamic narrative.

The present research aimed to comprehend the linguistic characteristics of negative criticism on news websites and tackled the global issue of negative criticism, specifically in Indonesia. Therefore, the investigation attempted to accomplish three objectives by quantitatively and qualitatively analyzing the corpus: (1) to investigate the lexical features of negative criticism in the corpus; (2) to identify speech acts underlying the negative criticism in the corpus; (3) to reveal the stance-taking role discharged by the comment posters on the negative criticism in the corpus.

Corpus and Methods

The research employs a mixed-method research design, which was designed on the assumption that readers freely express criticism and insults in their reactions to online news articles posted in the comment sections (see Dordević, 2020; Tenenboim & Cohen, 2015). It aimed to determine how readers subscribing to different political views engage in negative criticism in the comment section of a news website. A corpus of reader comments was constructed on news articles on the Detiknews site about the Jakarta governor’s tackling the COVID-19 outbreak. The corpus was compiled from all readers’ comments in response to eight news articles on Detiknews published between June 5 and 10, 2020, when the governor of Jakarta, Anies Baswedan, decided to end the Pembatasan Sosial Berskala Besar (PSBB) or large-scale social restrictions in the capital city of Indonesia and enter a transition phase. The corpus consists of 882 comments but does not contain information about the commenters’ backgrounds because we lack access to the metadata. This is a limitation of the present study. The highest number of comments for a news article is 227, while the lowest is 44. The other news articles received 191, 120, 105, 82, 67, and 46 comments. The comments were generally brief because of the 1000-character limit established by the news provider. The shortest comment comprised one word, i.e., mbelgedesss (a Javanese swear word), and the longest contained 168 words. Most comments included insulting terms and used pejorative language, swear words, or sarcasm targeting specific individuals or groups of people. Comments were posted primarily by readers who supported the Jakarta governor and those who opposed him.

The corpus of comments posted by Detiknews readers totaled 18,190 words. The corpus is relatively small, but there is a primary advantage to using a small, domain-specific corpus in pragmatic research. According to Koester (2010) and Vaughan and Clancy (2013), a context-specific small corpus allows researchers to access actual, naturally occurring language and retain a close link between language and context. The present study performed a combination of lexical and pragmatic analyses to examine three focal concerns: determining word frequencies, identifying speech acts underlying negative criticism, and revealing the negative criticism-related stance taken by the comment posters on the news site. Quantitative analyses were executed to ascertain the frequencies of words, speech acts, and stance-taking. The quantitative results were subsequently interpreted qualitatively to evaluate the linguistic features used by the comment posters to generate negative criticism and determine how they responded to political news articles based on the stances reflected in their comments.

The analysis began with the examination of word frequencies to specifically examine some prominent lexical items of the texts in the corpus that served to build negative criticism. Studies of corpus research often begin with word frequency analyses because these analyses can reveal the silent phenomena in a corpus and indicate the prominent concerns of language users. Consequently, researchers can determine themes worthy of discussion (Hyland, 2015). The present investigation employed a corpus tool named Sketch Engine to generate a word frequency list. The use of nouns, deictic persona, and negation lexis salient to the corpus was determined from the 100 most frequent words.

The second step utilized the pragmatics perspective to study how criticism was produced through the reader comments corpus. This study intensively references research conducted by Trajkova and Neshkovska (2018) and Atifi and Marcoccia (2017) to investigate the role of social media readers and viewers apropos political issues. It follows their pragmatic analysis to examine reader comments based on speech acts, to determine their illocutionary value in identifying speech acts underlying criticism, and to reveal the communicative functions of reader comments. The speech act assessments were conducted referencing categories proposed by Atifi and Marcoccia (2017): evaluative, a sub-category of the expressive, indicating the attitudinal expressions of speakers or writers toward the discussed entities; directive, or messages produced by speakers or writers designed to make the audience act according to their intentions; and analytical, a sub-category of the assertive, signifying descriptive discourse that speakers or writers employ to interpret the discussed entities (Trajkova & Neshkovska, 2018). The analysis of the corpus data disclosed that some readers posted comments comprising more than one message. Thus, these comments do not contain an individual category of speech acts but a combination of several categories. In this case, a different order of speech acts underlying a comment counted as a different category. Thus, for example, a comment consisting of evaluative and directive speech acts was considered a different category from a comment consisting of directive and evaluative speech acts.

The final step of the analysis uncovered the role of readers by investigating their stances and positions in engaging in criticism in the comment sections of Detiknews. This scrutiny was grounded in concepts mooted by Atifi and Marcoccia (2017) and subsequently developed by Trajkova and Neshkovska (2018), categorizing three social roles discharged by commenters on social media: judge, activist, and analyst. Readers are considered judges when they express their evaluation of a particular situation, in this case, the Jakarta governor’s policies in tackling the COVID-19 outbreak. Readers functioning as activists focus on persuading the audience to confirm or act in a manner conforming to the stated belief on the discussed topic, for example, by asking questions or issuing commands. Readers taking on the mantle of analysts investigate and clarify specific situations to ensure comprehensive understanding by the audience.

Results and Discussion

Primarily, this research utilized the methods elucidated in the previous section to conduct a study including word frequency analysis as well as speech act and stance analysis. The sections that follow discuss the results derived from both analyses.

Word Frequency Analysis

Table 1 presents the Sketch Engine-generated list of the 100 most frequent words found in the Detiknews comments corpus. The table evinces that some nouns were used to refer to the Jakarta governor, each with relatively high occurrences: Anies (8576 word/million), pak ‘Mr’ (4178 word/million); gubernur ‘governor’ (4013 word/million); Anis (3408 word/million); gabener abbreviated from gubernur gak bener, meaning an incompetent governor (2254 word/million); wan, a term of address for Arabic descent, (2199 word/million); and si, an article used before proper nouns to show intimate relations or to insult (1704 word/million).

Table 1 Top 100 frequent words in the corpus of Detiknews comments
Full size table

These listed nouns denote that readers who posted on the news comment sections addressed primarily the Jakarta governor by his first name, Anies. Many commenters misspelled the governor’s first name, writing Anis instead of Anies. The deliberate misspelling may have been employed to convey a disparaging impact, as seen by the fact that those who favored Anies mainly spelled his name correctly. The higher frequency of the noun Anies than the colloquial term of address pak, abbreviated from Bapak, which is used by speakers to address an older or respected man, also suggests that readers displayed an egalitarian relationship with the Jakarta governor when commenting on news articles. Some readers explicitly used insulting terms for the Jakarta governor: gabener, wan, and si. The word gabener is a pun, abbreviated from gubernur gak bener, meaning an incompetent governor. The racial term wan to address the governor has negative connotations. It was popularly used in a famous Indonesian soap opera to address a cunning antagonist of Arabic descent. The word si is an article (the) used before a surname to address a less respected person or someone intimate with the speaker. Apparently, anti-Anies readers used colloquial insulting terms in negative comments about the Jakarta governor. Such language is generally considered rude, as shown in the following examples.

  1. [1]

    hahahaha wanabut planga plongo gakbecus kerja, wan aibon munafik pinokio bohong mulu

    ‘hahahaha, Wanabut is gawking because he doesn’t know how to do his job, Wan Aibon is a hypocrite who loves to lie like Pinocchio’

  2. [2]

    dari segala urusan, banjir, virus, dll Kenyataann Gabener terbodoh ini pura2 lupa membandingkan

    ‘from all the issues, flood, virus, etc., the fact is that the most foolish Gabener pretends to forget to compare’

  3. [3]

    ….semakin orang tahu betapa gobloknya si anies…. semakin banyak cuap-cuap dia…

    ‘….more and more people know how imbecile Anies is…he talked more and more…’

Such terms of address represent the linguistic features of negative criticism posted by readers subscribing to a political view different from the Jakarta governor and his supporters.

Pragmatically, some of the lexical items (Anies, Anis, gabener, and wan) selected by readers to address the governor of Jakarta, Anies Baswedan, in their posts on the news comment sections indicate a communicative strategy designed to threaten the governor’s public image and cause disharmony (Culpeper, 1996). Polite language is principal in Indonesian culture because of the cultural influence of the two largest ethnic groups of Indonesia, Javanese and Sundanese. Both indigenous languages are classified based on contexts and speech levels rooted in the relationships between speakers and hearers. Lexical honorifics represent a type of politeness marker used for older people or individuals considered worthy of respect, such as teachers, lecturers, and state officials (Rahayu, 2017). Some honorifics utilized in colloquial Indonesian include Bapak (Mr.), pak (Mr., but less formal than Bapak), Ibu (mother or Mrs.), Bu (less formal than Ibu), Mbak (a Javanese loanword used to address a sister or young woman), Mas (a Javanese loanword employed to address a brother or young man), and Kak (abbreviated from the word Kakak for an older brother/sister in the family or an older man/woman). The use of lexical honorifics is pervasive in Indonesian; it reflects the relationships between speakers based on age, gender, status, social distance, and formality. The honorific system is encoded in the Indonesian linguistic patterns and constitutes normative politeness. It is thus deemed impolite for speakers to address listeners without due consideration of the honorific system. Given this context, the higher frequencies of the words Anies, Anis, gabener, wan, and si for the Jakarta governor in the corpus of the Detiknews comments than the word pak, as well as the fact that the word Bapak is not found in the top 100 frequent words, suggests that readers posted rude comments on news articles reporting on the Jakarta governor and confirms that readers contributed to the construction and dissemination of negative criticism.

Pronouns represented another linguistic feature used in the Detiknews comment sections to construct negative criticism. Three pronouns were found in the 100 most frequent words in the Detiknews corpus: dia (he) was the most recurrent, and kita (we) was the least regularly used, as shown in Table 2.

Table 2 Pronouns appearing in the 100 most frequent words used in the corpus of Detiknews comments
Full size table

The pronoun dia (he) was used by the readers in their comments solely to refer to the Jakarta governor, Anies Baswedan. Meanwhile, the pronoun saya (I/my/me) alluded to the readers who posted comments on the news comment sections, and the pronoun kita (we) was used by readers referring to themselves along with other readers considered as being in the same position. The inclusive kita (we) served to facilitate the engagement of other readers with the commentator’s opinions and sought to persuade them to act as exemplified in [4] and [5]:

  1. [4]

    Kita harus maklumi karena anis ahli tata kata bukan ahli tata kota…

    ‘We need to understand that Anis is only an orator, not an expert in urban planning.’

  2. [5]

    Mari kita diskusi dengan tema: IMPEACHMENT GUBERNUR DKIFootnote 1 DI TENGAH PANDEMI COVID19 DAN BANJIR ROB.

    ‘Let’s have a discussion with a theme: THE DISMISSAL OF JAKARTA GOVERNOR IN THE MIDST OF THE COVID-19 PANDEMIC AND ROB FLOOD.’

The frequent use of personal references indicates interactions among readers who posted comments and demonstrates that the language used in the news comments has the characteristics of spoken language (Biber, 1988; Chafe, 1982; Halliday, 1985, 1987). In this case, the context-dependent reference deictic is used to express the opinions of readers vis-à-vis news on the Jakarta governor. Such terms are specifically employed to defend commenters’ political views and generally influence other readers. However, no reference deictic was utilized to directly address the Jakarta governor or other political figures, demonstrating that commenters preferred nouns to pronouns to address the Jakarta governor. This finding differs from Trajkova and Neshkova’s (2018) results that a relatively high number of Twitter and Facebook authors employed the second-person pronoun you to express anger and hatred and directly address political figures. This disagreement in results may be attributed to the difference in the topic of the comments analyzed or the culture between Indonesian and Macedonian speech communities.

The other notable linguistic feature used to post negative comments in the corpus is evaluative lexis, the first of which is negation lexis comprising ga, gak, tidak, bukan, and jangan, all of which mean no (see Table 1). The other negative terminology found in the corpus of Detiknews comments, not listed in the top-100 most frequent words as yet, consists of g (796 words/million), nggak (796 words/million), kagak (609 words/million), ngga (468 words/million), and enggak (49 words/million). These negations are commonly used in spoken language and social media, particularly the word g. The negation lexis often co-occurs with particular words, such as bagus, becus, bener, beres, jelas, kerja, malu, and prestasi, to convey a negative evaluation to the governor, for example, ga ada prestasi/tidak berprestasi (no achievement), ga ada terobosan (no breakthrough), ga becus/gak becus (incompetent), ga beres/gak beres (not right), ga jelas/gak jelas/tidak jelas (illogical), ga bisa kerja/tidak bisa bekerja (can’t work well), ga cocok jadi pemimpin (not qualified to be a leader), ga malu/tidak punya malu (shameless), gak bener (not good), gak bagus (not good), etc. Commenters also use evaluative lexis in negative comments, such as kata (words), pemerintah (government), pusat (center), presiden (president), and orang (person). The word kata frequently co-occurs with tata to form tata kata (speech). It is used when evaluating Anies Baswedan as a governor who is good at speaking but not at working. The other interesting negative evaluation given by commenters is that they think Anies always wants to compete with the central government under President Jokowi by adopting policies different from those of the central government. In this context, they frequently use the words pemerintah and pusat, which represent Jokowi. Many also commented that one reason why Anies often adopts policies at variance with those of the central government is that he is campaigning to be the next president.

Speech Acts of the Comments

The analysis of speech acts demonstrates that the comments published by the readers in the corpus of Detiknews contain evaluative, analytical, and directive speech acts. In addition, we found comments that consist of the combination of several speech acts manifesting in different orders: two acts (directive and evaluative; evaluative and directive; analytical and evaluative; evaluative and analytical; analytical and directive; directive and analytical), three acts (evaluative, directive, and evaluative; evaluative, analytical, and evaluative; evaluative, directive, and analytical; evaluative, analytical, and directive; analytical, directive, and evaluative), four acts (evaluative, directive, evaluative, and directive; directive, analytical, directive, and evaluative; directive, evaluative, directive, and evaluative; evaluative, directive, analytical and directive), and five acts (directive, evaluative, directive, evaluative, & directive).

The most frequently used speech act is evaluative, indicating that the comments posted by the readers are dominantly in the form of a single speech act (see Fig. 1). Commenters use the evaluative speech act to show their attitudinal expressions toward the discussed entities in the news articles, which, in this case, are predominantly about the Jakarta governor and his policies in tackling the Covid-19 pandemic.

Fig. 1
figure 1

Speech acts of the comments

Full size image

In contrast, the least frequently used speech acts are those that occur only once in the corpus, i.e. directive and analytical; evaluative, analytical, and directive; analytical, directive, and evaluative; directive, analytical and evaluative; directive, evaluative, and analytical; directive, analytical, directive, and evaluative; directive, evaluative, directive and evaluative; evaluative, directive, analytical & directive; and directive, evaluative, directive, evaluative, and directive. The comments are more commonly used to convey a negative evaluation than a positive one. The negative evaluations are addressed mainly to the Jakarta governor and the readers who support the governor.

Positioning of the Commenters

Based on the analysis of commenters’ stances and positions, it is found that there is a one-to-one relationship between the three types of speech acts, i.e., analytic, directive, and evaluative, and the three types of social roles, i.e., analyst, activist, and judge. The speech acts underlying commenters’ utterances taking the role of analyst, activist, and judge are analytical, directive, and evaluative, respectively. Hence, the results of the stance analyses demonstrated similarly to the speech act analyses’ results, i.e., readers’ comments on each news article could contain more than one position. In some cases, commenters assumed multiple roles in their generated comments. The majority of the commenters acted as judges (665), followed by the other single-role commenters: activists (33) and analysts (24). Interestingly, several commenters also assumed a combination of two or three roles in their posted comments. In such instances, 6% of the posts evinced the activist and judge combination, followed by judge and activist (5%), analyst and judge (2%), judge and analyst (2 %), and the amalgamation of judge, activist, and judge (2%). The other role combinations occur less than 2% in the corpus.

Commenters as Judges

As previously indicated, most news commenters take the role of judges in comments posted in response to news articles on the Detiknews website. Comments assigned to this category tended to encompass evaluative speech acts designed to evaluate, assess, assert, and draw an inevitable conclusion concerning the issues raised or reported in the articles. This study demonstrates that most such comments negatively evaluated Anies Baswedan’s policies, particularly with respect to the handling of the COVID-19 outbreak in Jakarta.

  1. [6]

    masyarakat sudah tau anis gak becus membuat kebjiakan. jakarta jd terbesar penyebaran covid salah satunya dikarenakan kebijakan efek kejut anis yg membuat antrian panjang dmana mana. pembagian sembako tidak merata, gabener amatiran.

    ‘Everybody knows Anies is not good at making policies. Jakarta is becoming the city with the most positive cases; one of the causes is Anies’ shock effect policy that makes for long queues everywhere. The groceries are not evenly distributed, such an amateur gabener (an incompetent governor)’

The data in [6] reveal a negative assessment of policies adopted by Anies Baswedan as Jakarta’s governor. The commenter is dissatisfied with the measures the governor imposed to contain the spread of COVID-19, and the comment intentionally uses specific words to disparage or belittle him. For example, Anies Baswedan is mentioned by his first name, which is even misspelled as Anis. The reader also uses the corruption gabener instead of gubernur to refer to him as an incompetent amateur governor. Let us now consider another comment.

  1. [7]

    Gubernur DKI yg paling rendah hati sepanjang sejarah...... Habiskan dana APBD trilyunan tanpa sekali pun memamerkan keberhasilannya..... Salut pak Anies.....sungguh sangat bersahaja.....

    ‘The humblest governor of Jakarta in history...... Spent trillions of APBD funds without ever showing off his success..... Salute Pak Anies.....how very modest..’

The data in [7] offer another example of commenters as judges. Unlike [6], however, [7] does not directly comment negatively or adversely assess Anies Baswedan in his capacity as governor. The mere reading of the first line of the comment Gubernur DKI yg paling rendah hati sepanjang sejarah does not yield any critical connotation. Instead, the comment seems more like praise or a compliment. However, the next line, Habiskan dana APBD trilyunan tanpa sekali pun memamerkan keberhasilannya, reveals the satirical intent of the commenter apropos the governor’s use of Jakarta’s budget. In this instance, the commenter asserts that Anies spent trillions of Rupiahs, but the immense expenditure did not result in any constructive outcomes. The satire ends with Salut pak Anies….sungguh sangat bersahaja. The last line is addressed to Anies Baswedan using pak Anies. This honorific is intentionally used to intensify the negative evaluation of Anies’ actions. Thus, readers of online news who oppose Anies Baswedan would always deliberately attempt to belittle, censure, and even insult him using all available techniques in their posts. The two comments [6] and [7] typify their methods of conveying their judgments on Anies Baswedan. Now, let us examine a slightly different comment from a commenter taking the role of judge.

  1. [8]

    rumah susun: rumah Lapis. Banjir: air Genangan Normalisasi: Naturalisasi New Normal: Transisi. ada yg bs menambahkan perbedaan yg di Buat anis?

    ‘Flats: layered houses. Flood: Water Inundation. Normalization: Naturalization. New Normal: Transition. Can anyone add other new terms to all the terms Anis created?’

Similarly, the data in [8] also enable the categorization of this comment as a judge remark. Again, a critical question is directed toward Anies Baswedan about the terms he used in his policies. For example, the commenter asks if anyone can offer other examples apart from the differences between normalization and naturalization, new normal, and transition. Pertinently, the governor prefers using the term ‘normalization’ over ‘naturalization,’ and ‘transition’ over ‘new normal.’ The overlooked terms are used by the central government, the group that opposes Anies Baswedan. In this instance, readers posting such comments desire to cynically demonstrate Anies Baswedan’s penchant for creating terms that are essentially similar to previously used terminology.

Commenters as Activists

The proportion of commenters taking a single role of activists is 3.7%, positioning it as the fourth most prevalent role in the corpus. The comments under this classification demonstrate a directive speech act through the use of questions or commands with the intention to prompt readers to support or comply with the articulated belief regarding the discussed topic. Such comments were directed, for instance, against Governor Anies Baswedan, who was believed to be comparing the success of his policies to those implemented by the central government to fight the COVID-19 outbreak. Indeed, there was an implicit request from a reader for the Corruption Eradication Commission (KPK) to initiate an investigation into Anies' implementation of policies, as seen in [9].

  1. [9]

    Kpk mana nih koq diem aja?

    ‘Where is the KPK (Indonesia’s Corruption Eradication Commission), and why is it so quiet?’

This comment appears to have been prompted by cases of corruption that have transpired within the context of regional governance in Indonesia. In 2016, a significant majority of corruption cases, precisely 97%, were reported to have taken place within regional governments. Furthermore, it was found that 18 governors, accounting for 50 percent of all provincial governments in Indonesia, were implicated in these corruption cases (see Suhardjanto, 2018). In the above comment, the reader appears to have an implicit suspicion regarding potential corruption perpetrated by the governor. Moreover, the reader seeks to influence other readers, urging them to support initiating an investigation by the KPK into the policies implemented by the governor.

Commenters as Analysts

According to the findings, analysts emerged as the fifth most frequent role in the corpus. The comments falling under this category display the characteristics of an analytical speech act. This observation reveals that approximately 2.7% of the readers actively posted comments to analyze and elucidate the situation, enhancing understanding for other readers, as shown in the example below.

  1. [10]

    Dulu pemotor dibebaskan di Thamrin karena udah ada putusan Mahkamah Agung, Anies menjalankan putusan Mahkamah Agung tersebut.

    ‘Previously, motorcycles were permitted to cross the Thamrin roads due to a Supreme Court order; therefore, Anies was simply carrying out the Supreme Court's decision.’

Data [10] showcases a commenter’s analysis on a topic discussed in a news article, highlighting the governor's changing policy on motorcyclists. Anies initially permitted motorbikes to cross Thamrin roadways, which his predecessor, Ahok, had previously banned. Anies then decided to make it illegal for motorcyclists to use the streets during the large-scale social restrictions period. Comment [10] was posted by the commenter who appears to support Anies, the governor. The commenter provided justifications for the governor's implementation of the prior policy and criticized the news article and other commenters who argued in favor of Ahok’s policy.

Commenters as Activists and Judges

Commenters also play the role of activists in around 6% of the posts, constituting the second highest profile in the study’s corpus. The comments classified into this type contain directive and evaluative speech acts that primarily intend to persuade people to do something and to assess specific issues. The speech acts can appear as questioning, ordering, imploring, challenging, and summoning (Trajkova & Neshkovska, 2018). Most such posts in the present study’s corpus tended to criticize Anies Baswedan’s handling of the COVID-19 pandemic and assail his other policies as governor. In general, however, the comments did not include disparaging words because they aimed principally to question or censure Anies Baswedan, as can be noted in [11].

  1. [11]

    Pak kalo tidak tahu ya di edukasi dong. Kesannya masyarakat indonesia bodoh semua apa…apa kata new normal lebih sulit di pahami di banding kata *social distancing? *Lockdown? *Work from Home? *Masker? Dan kata2 lain baru selama adanya covid Ini?

    ‘Sir, if people don't know, please educate them. It seems that Indonesian people are all stupid... Is the term new normal more difficult to understand than the terms *social distancing? *Lockdown? *Work from Home? *Face Mask? And other new terms during the Covid-19 pandemic?’

The data in [11] represent a critical request and assessments or questions expressed by commenters concerning some of the COVID-19 terms they thought were too complicated for the ordinary citizens of Jakarta. The reader asks whether Anies Baswedan cannot use easy-to-understand terms. Alternatively, the reader asks the governor to educate the people of Jakarta about the meaning of the COVID-19-related terms. However, the reader’s queries are far from objective because terms such as social distancing, lockdown, and work from home were not devised by Anies Baswedan. In fact, they were used worldwide in relation to the COVID-19 pandemic and activities undertaken to contain the spread of the virus. Thus, such activist comments are posted to bluntly and biasedly criticize Anies Baswedan even though they used the honorific pak to address the governor.

Commenters as Judges and Activists

The study’s corpus also contains posts whose commenters simultaneously discharge the roles of judges and activists. This category comprises 5% of the data, making it the third most prominent in the corpus. Comments of this type exhibit an amalgamation of two speech acts, as displayed by [12] and [13].

  1. [12]

    memang enak menyalahkan….ini masalah rumit, pengamat sak karepe dewe ngobrol. kesehatan dan perut kenyang adalah dua pertimbangan yg berat. disuruh taat 50% saja sulit. aturan gubernur sdh pas,krn transisi. maju pak gubernur. yg penting sampean kerja, hasilnya untuk rakyat jakarta. pengamat juga kerja dg caranya sendiri dan hasilnya untuk dirinya sendiri.

    ‘It indeed feels good to blame someone….this is a complicated problem; observers just comment casually. Health and people’s financial conditions are two tough considerations. Being told to obey 50% capacity is difficult. The governor's rules are right because of the transition. Go Governor. The important thing is that you work; the results are for the people of Jakarta. Observers also work in their own way, and the results are for themselves.’

The data in [12] exemplify commenters simultaneously adopting the roles of judges and activists. In this case, the commenter displays an evaluative act by stating that it is easy to blame the regional government of Jakarta. The commenter adds an evaluation of situations in which political and social observers have criticized the governor’s policies without considering how difficult it is for him to make the right decision, which also implies negative evaluation for the political and social observers. The commenter continues using directive acts that motivate Anies Baswedan to persist with his policies of enforcing a so-called transition time in which Jakarta workers were required to work in shifts. This comment is obviously written by a commenter supporting Anies Baswedan, although this position is not specifically mentioned in the post. It is thus reasonable that no insulting or disparaging words or phrases are found in this example. Let us examine another similar example from a reader opposed to Anies.

  1. [13]

    ramaah gabanernya…maniss gabenernyaa.maniss kebijakannya.nikmatin deeh 2 tahun lagii.

    Such a friendly gabener.. such a sweet gabener. Such a sweet policy.. enjoy two more years of his government…

Such an assumption of the conjoined judge and activist profile can also be seen in [13]. Just like [12], the reader begins the post with an evaluative act in stating that Anies Baswedan is a friendly and sweet governor and that his policies are also agreeable. The post ends with a directive act calling on the people of Jakarta to enjoy another 2-year term of Anies Baswedan’s leadership. There seems to be nothing untoward in this comment at first glance, especially if only the evaluative aspect (the commenter as a judge) is considered. A different meaning emerges when the line containing the directive act (the commenter as an activist) is read. This part of the statement discloses the intent of issuing a satirical critique of Anies Baswedan’s policies. The comment's content reveals that it is posted by someone who opposes Anies Baswedan. Significantly, the commenter does not use insulting or disparaging words or phrases even though its response to the governor is critical. Let us consider another example below.

  1. [14]

    Saya kira kritik dan masukan yang ada sudah bagus sehingga sebaiknya diterima dan ditindaklanjuti, tinggal bagaimana Pemprov menyikapi hal itu untuk perbaikan, gak usah mengelak cari cari alasan pembenar. Kritik memang pedas, tapi kalau kritiknya bagus sebaiknya ya dijadikan masukan…

    ‘I think the existing criticism and input are good in that they must be accepted and followed up on; it’s just a matter of how the Provincial Government reacts to it for improvement; there is no need to dodge it by looking for justifications. Criticism is scathing, but if the criticism is good, it should be used as input...’

Comment [14] is similar. Here, the commenter begins with evaluative act, assessing that the existing criticism against the regional government of Jakarta is already good. It also uses an evaluative-expressive act to pronounce a hypothetical judgment: a situation exists in which the government evades criticism by looking for unnecessary justifications. Then, it is followed by a directive act to convince the regional government of Jakarta to accept any reasonable criticism and act accordingly. The comment ends with an assertive and directive act to explain that critiques are very often insensitive and call on the government to accept unsympathetic criticism as long as it is constructive or makes sense. Commenters simultaneously assume two pragmatic roles in [13] and [14] but no unkind words or disparaging phrases are used in [14], unlike in instance [13]. The post encompasses cruel words or belittling expressions generally aimed at Anies Baswedan as governor. Such comment cited in this paper appear objective, and the commenter attempts to place circumstances into perspective. Notably, such commenter also tends to avoid personal attack or censure and do not mention personal names such as Anies Baswedan. Rather, they often use institutional nomenclature, for example, regional government. This feature could indicate the efforts of such commenter to remain as objective and unbiased as possible.

Commenters as Judges and Analysts

In approximately 2% of the corpus’s posts, commentators serve as judges and analysts. This type of comment contains evaluative and directive speech acts primarily intended to evaluate specific issues and persuade people to act, as seen in the example below.

  1. [15]

    tak pantas naik kelas manusia seperti anis ini. kerjanya bukan biasa biasa saja. cenderung kerjanya gak becus dan menghambur hamburkan uang rakyat. contoh paling nyata tgupt mengabiskan uang rakyat tak ada sumbangsihnya untuk rakyat dan warga dki.

    ‘Humans like this Anies don’t deserve to go to class. His work is not ordinary. His work tends to be incompetent and wastes people’s money. The most obvious example is that TGUPT spends public money without making any contribution for the citizens of Jakarta.’

The data in [15] exemplifies posts by commenters who assume the position of both judge and analyst, which comment takes the opposite direction in its assessment of Anies Baswedan’s performance. It indicates the polarization of pro- and anti-Anies Baswedan supporters. Specifically, this commenter evaluates Anies Baswedan’s performance in an extremely negative manner and asserts the opinion that Baswedan is an incompetent governor who likes wasting the people’s money. The reader then delivers the analytical conclusion that Anies Baswedan does not contribute to the people of Jakarta. In this instance, the reader deliberately uses very harsh terms and demeaning phrases to criticize and insult the governor. The post also addresses Anies Baswedan with his first name in an intentional attempt to denigrate him. The comment indicates that the reader is anti-Anies Baswedan.

Commenters as Analysts and Activists

Commenters as analysts and activists constitute the eleventh-highest comments, making up 2% of the posts in the study’s corpus. The comments belonging to this category encompassed analytical and directive speech acts and aimed to analyze specific situations or issue clarifications so that readers and other commenters could understand the issues reported in the news (Atifi & Marcoccia, 2017). Specifically, the commenters tried to explain, contextualize, enlighten, clarify, and analyze certain discussed situations or issues. This category is exemplified in [16].

  1. [16]

    Faktanya status DKI masih zona merah, Penambahan pasien positif covid-19 masih banyak, kenapa dilakukan pelonggaran PSBB dengan bahasa PSBB Transisi, mohon ditinjau ulang dan dievaluasi kebijakan PSBB transisi agar tidak jadi gelombang kedua covid-19.

    ‘The fact is that DKI still has the status of a red zone; there are still numerous additional patients testing positive for Covid-19. Why is the PSBB relaxed with the term Transitional PSBB? Please review and evaluate the transitional PSBB policy so that a second wave of Covid-19 does not occur.

In data [16], the reader’s analysis of the Jakarta situation appears to be objective with respect to Anies Baswedan’s actions to tackle the pandemic. This comment consists of a combination of analytic and directive speech acts. An analytic act is produced in the first part of the comment by stating that Jakarta is a COVID-19 red zone because of the steady increase of new cases of infection. In the second part, the commenter politely queries whether the policymakers could review and reevaluate the relaxation of the so-called large-scale social restrictions to avoid the second wave of COVID-19 cases. Notably, commenters playing the analyst role do not use any insulting words in response to the governor’s policies during the pandemic in Jakarta. They try to report facts and offer sound arguments. In addition, the comment does not mention Anies or his government, although [16] is evidently critical of his policies. Let us see another comment of this type directed toward the other readers.

  1. [17]

    Faktor utama nya ya rakyat. Kalau anda semua mau tetap hidup dan tetap sehat tolong ya disiplin. Kita tak bisa menyalahkan semua pada pemprov atau pemerintah pusat.

    ‘The main factor is the people. If you all want to stay alive and stay healthy, please be disciplined. We cannot blame everything on the provincial government or the central government.’

The data in [17] typify commenters assuming the roles of both analysts and activists. In this example, the reader asserts that the people should function prominently in the success of implementing large-scale social restrictions and contribute to efforts to contain the transmission of COVID-19 in Jakarta. The commenter further urges people to discipline themselves if they want to remain healthy and to not blame regional or central governments if efforts to contain cases of COVID-19 fail. Such readers try to be objective in their evaluations of the situation in Jakarta during the governor’s handling of the COVID-19 outbreak. They do not appear to side with either Anies Baswedan as the governor of Jakarta or Joko Widodo (Jokowi) as the president of Indonesia. It must be remembered that the residents of Jakarta tended to polarize into two extreme groups in terms of their political affiliations: one group stood with Anies Baswedan, and the other with Jokowi. Moreover, both the regional and central governments appeared to stake claim to each other’s territories or authorities in attempting to handle the COVID-19 pandemic.

Commenters as Activists and Analysts

The comments assigned to the category of the combined activist-analyst comments constitute only 0.1% of the corpus and contain directive and analytical speech acts as in [18] .

  1. [18]

    tenang pak anies, ketegasan pak anies dalam covid19 membuat murka pengusaha dan akhirnya mereka mengutus buzzer bayaran membuat survey, makanya kepala daerah yg permisif langsung naik dalam survey, pemimpin ga boleh menyenangkan untuk elektabilitas, pemimpin adalah melindungi nyawa rakyatnya, bukan duit penguasa.

    ‘Calm down Pak Anies, Pak Anies’ firmness in handling Covid-19 has made businesspersons angry, and finally, they sent paid buzzers to do a survey. That is why permissive regional heads immediately took surveys; leaders should not play around with electability, leaders protect the lives of their people, not the ruler’s money.’

The commenter in [18] begins with directive act to support Anies Baswedan, followed by a positive analysis of Anies Baswedan’s tough policies in dealing with the pandemic, which made entrepreneurs react negatively. In this example, the reader supports Anies Baswedan for being assertive and thinks his assertiveness causes certain business people to turn against him. Perhaps, they could pay polling buzzers to downgrade his electability ratings. This commenter asserts and assesses the situations by offering the outcomes of personal observations or by analyzing why permissive regional leaders tend to demonstrate higher electability ratings. In this instance, the analysis yields a positive judgment of the governor’s policies for handling the Covid-19 outbreak. The comment's content suggests that the commenter is a supporter of Anies Baswedan’s supporters or at least a group of people who are sympathetic with or objective to his policies.

Commenters as Judges, Analysts, and Activists

The study’s corpus also clarifies the simultaneous judge-analyst-activist role discharged by commenters. Such comments comprise only 0.1% of the data, performing evaluative, analytical, and directive acts as in [19]:

  1. [19]

    FKM UI trus solusinya apa? PSBB terus? Yang mau kasih makan pekerja harian siapa? Negara? Kuat sampe brapa lama? Yang bener itu naik karena penungkatan test masal tinggal sosialisasi terus psbb transisi. Anggota FKM UI begitu karena mereka dapat salary bulanan coba dipikirkan ekonomi pekerja harian jadi mikirnya gak pembatasan dirumah terus. Kalo gak ada solusi lebih baik diam.

    ‘FKM UI, then what is the solution? PSBB continues? Who will feed the daily workers? Country? How long can it last? The truth is that the cases went up because of the increase in mass testing, and they just continued to socialize the transition PSBB. FKM UI members are like that because they get a monthly salary. Try to think about the financial condition of daily workers, who won’t only think about the staying-at-home policy. If there is no solution, it is better to be silent.’

The data in [19] exemplify commenters playing the roles of judges, analysts, and activists. This comment begins with a question for the Faculty of Community Health at the University of Indonesia: what solution can it offer other than the imposition of large-scale social restrictions in Jakarta? The commenter does not agree with the prolonged implementation of social restrictions because the worker economy would be most affected. The post then analyzes the situation and asserts that increases in testing caused the amplified reports of the number of COVID-19 cases in Jakarta. The commenter concludes that the faculty took such a position because of its guaranteed monthly salary by the government. Hence, money would not be a problem for such academics even if the social restrictions were prolonged. The comment ends by asking the faculty to remain silent if it does not have any solutions to the problem.

The data in [17] and [19] share the characteristics of not appearing to side with a certain group and of attempting to place phenomena in perspective. However, [18] also demonstrates a different standpoint: the commenter does not explicitly assert a stance toward the regional government’s policies but the reader’s disagreement with prolonged large-scale social restrictions and preference for transitional large-scale social restrictions become clear on careful reading. This reader does not directly oppose the government; instead, the Faculty of Health Community is censured for not offering a better solution to the problem other than advocating the extension of large-scale social restrictions. Significantly, both [17] and [19] are devoid of abrasive words or offensive phrases, suggesting such commenters endeavor to take objective positions without aligning with any particular group or opinion. However, posts by commenters assuming triple roles are rare.

Commenters as Judges, Activists, and Analysts

Commenters who combine the roles of judges, activists, and analysts represent only 0.3% of the corpus and their posts usually incorporate several speech acts as can be seen in [20], which encompass evaluative, directive, and analytical acts.

  1. [20]

    biasa aja, cuma dr parte itu aja yg biasa protes, bukannya beri bantuan u meringankan masyarakat yg kesusahan krn pandemi, emang mo diawasin satu2 mobilitasnya, coba liat banyak masyarakat yg sdh tdk bekerja, g ada uang apalgi tabungan, buat beli paket data aja u bljr anak2nya cari pinjaman, itupun kalo dpt, kendaraan ditarik, listrik nunggak, makan cuma nasi aja itupun pembagian sembako, padahal keadaan inilah yg membuat banyak orang tdk berpikir jernih, mudah tersulut emosinya, mudah2 tdk smp trjadi….

    ‘No wonder; It’s only that political party that likes to protest, they don’t give help to relieve people who are in trouble because of the pandemic. Do you really want to monitor one’s mobility? Look at the many people who don’t work and don’t have money or savings just to buy data packages for their children to learn; they look for loans; their vehicle is withdrawn; the electricity is in arrears; and they only eat rice from the government. This is the situation that makes many people not think clearly and easily ignited by emotions; hopefully worse things don’t happen...’

The data in [20] demonstrate the simultaneous assumption of the three roles of judge, directive, and analyst. The post begins with an evaluative-assertive speech act to assess a party that likes protesting but never helps those in need because of the pandemic. In this case, a negative judgment is delivered toward the party’s attitude in response to restrictions imposed because of the pandemic. The commenter then produces a directive speech act by asking the protesting party to view the fact that many people were unemployed and thus lacked the funds to pay bills, buy Internet quotas for their children, or satisfy their daily requirements. Finally, the post examines the situation using an assertive act and concludes that people could act irrationally and become easily provoked to commit acts transgressing law and order because of the increased economic difficulties faced by many citizens.

Conclusions

This study aimed to investigate the linguistic expression of negative criticism in the Detiknews corpus, identify speech acts underlying negative/positive evaluation, and reveal the roles assumed by commenters in terms of their stance-taking apropos the negative criticism.

Readers who post negative criticism use three linguistic features, i.e., nouns, pronouns, and negation lexis. The findings about these linguistic features are based on the list of the 100 most frequent words in the corpus. The nouns used to refer to Anies Baswedan had a relatively high frequency, for instance, Anies, pak, gubernur, Anis, gabener, wan, and si; the last four were utilized primarily to disparage and denigrate the governor. Three pronouns appeared frequently in the corpus: dia, saya, and kita. Dia referred to Anies Baswedan, saya to the commenters themselves, and kita to the commenters and other people in their positions. Five different negation lexis word forms also often occurred in the study’s corpus, i.e., ga, gak, tidak, bukan, and jangan. The frequent use of the informal lexical items ga and gak indicates that the comment posters preferred colloquial Indonesian. The three linguistic features above were employed to maximize the pragmatic effects of insulting comments on Anies Baswedan.

The speech acts underlying the negative criticism were evaluative, directive, analytical, and a combination of two or more speech acts. They appeared in the corpus on their own or in combination and demonstrated how commenters positioned themselves with respect to their comments. Twenty-one types of social roles were determined for the commenters: one act (activists; analysts; judges); two acts (directive and evaluative; evaluative and directive; analytical and evaluative; evaluative and analytical; analytical and directive; directive and analytical), three acts (evaluative, directive, and evaluative; evaluative, analytical, and evaluative; evaluative, directive, and analytical; evaluative, analytical, and directive; analytical, directive, and evaluative), four acts (evaluative, directive, evaluative, and directive; directive, analytical, directive, and evaluative; directive, evaluative, directive, and evaluative; evaluative, directive, analytical and directive), and five acts (directive, evaluative, directive, evaluative, & directive). Most commenters play the role of judges according to the most frequent speech act (evaluative act). Indeed, such judgmental readers predominantly produced insulting comments on Anies Baswedan. It is also found that there is a one-to-one relationship between the three types of speech acts and the three types of social roles. For example, the speech acts underlying readers’ comments taking the role of analyst, activist, and judge are analytical, directive, and evaluative, respectively. Therefore, the number of the role types and speech act types are the same.

In general, the news commenters could be classified into three groups: those supporting Anies Baswedan, those opposing him, and those attempting to remain objective or unbiased toward him or others. However, it is difficult to identify the socioeconomic and educational backgrounds of the commenters because the forum of news comments is a computer-mediated interaction where participants are anonymous. Future studies may need to consider the comments posted by readers by, for example, examining the grammatical structures of sentences they produce or investigating the correlations between the simplicity of the comments and the number of insults or impolite words.