What They Do Not Talk about When They Talk about Radicalism: (A Corpus-assisted Discourse Analysis on Islamic News Portals in Indonesia)

The lack of a clear definition of radicalism leads Islamic organisations to feel entitled to interpret it. It results in contention for the meaning of radicalism and forms a different reality for each reader. By taking a case study on two spectra of Islamic organisations, moderate Islamic organisation (NU Online) and Islamic extremist organisation (Portal Islam), this study aims to understand the construction of radicalism in two Islamic news portals. We utilised a dataset of news about radicalism from September 2018-2019 and analysed it using a combination of corpus linguistic (CL) and critical discourse analysis (CDA). While CL helped to reveal emerging discourses, CDA intended to observe the patterns and relate them into socio-political contexts critically. The results indicate that each site was blurring the information function of news portals into a propaganda function. They also generate fragmented knowledge, which leads to a misrepresentation of paradigm towards radicalism. This leads to discriminatory actions against other groups. The meaning of radicalism in the media may encourage group dichotomy, which is counterproductive to countering terrorism in Indonesia. This study contributes to a comprehension of the terrorism phenomenon by providing a closer view of how moderate and extreme Islamic organisations interpret radicalism.


INTRODUCTION
Terrorism is an objective reality that can be interpreted subjectively (Karaffa, 2012(Karaffa, , 2015. As an objective reality, terrorism is a crime against humanity that threatens civilisations and the sovereignty of countries in the world (Crenshaw, 2014;Hülsse & Spencer, 2008;Malkki & Sallamaa, 2018)it is extraordinary to see that what was once a marginal subject for social science has developed into a full-fledged program of "terrorism studies." In fact, recently a sociologist considered the subject of sufficient importance to write a doctoral dissertation and then a book on the "social construction" of the field (Stampnitzky 2013. In some ways, terrorism is considered an act rooted in radical Islam . It creates stereotypes in radical Islamic movements, which result in discrimination against groups with a strong Islamic identity who are suspected of being terrorists, extremists, or whole groups (Mubarok & Hamid, 2018;Umar, 2010). This legalises certain groups of people or even a state to commit acts of discrimination and violence against radical groups (Karaffa, 2012;Schmid, 2013). This situation is even more complicated due to the mass media bias in reporting on Islam. It is too focused on narrating news on incidents and acts of violence in the Muslim world (Chertoff, 2008;Maulana, 2018;J. Neumann et al., 2018;Powell, 2011)business processes, organisation structures as well as Enterprise Resource Planning infrastructure. The case study focussed on the summative evaluation of DOI: https://doi.org/10.29313/mediator.v13i2.5859 the programme post-implementation. This assessment involved 25 semi-structured interviews with employees across a range of managerial strata capturing more than 65 roles within the change programme at both local and global levels. Data relating to their perception of evaluation effectiveness and shortcomings were analysed by means of template analysis. The study identifies responsibilities for executing an evaluation alongside various methods and tools that are appropriate, thereby focussing on the "Who" (roles, responsibility for particular activities.
However, a term like radicalism has never been clearly defined in either the academic or the policy literature (Schlegel, 2018). In a practical sense, every organisation feels entitled to define what radicalism is. It is feared that this may lead to the narrowness of the meaning of radicalism, which will lead to misconstruction and misconception of the meaning itself. Furthermore, the term of radicalism is deliberately raised and exaggerated to seek publicity, which affects a larger audience than those who are direct victims so that it is treated more favoured than it should be (Crenshaw, 2014;Lauderdale & Oliverio, 2018)this chapter suggests that the social problem task is not to expose or define the terrorist of the week -be it the Unabomber or the Islamic State organization (ISIS. The problem is that every organisation tries to construct the meaning of radicalism and to persuade the audience to have the same definition as them. The contestation of radical narratives between moderate and extreme Islamic organisations on the internet is an interesting subject examined in this study. Mass media is an ideological instrument that shapes the social cognition of others. The media should not be seen as a neutral space that freely provides every element of society with a means of publication, but it is a means of equalising interests, ideas, and concepts from owners to the public. The media becomes a subject that constructs reality based on its interpretation and definition to be disseminated to the public (Dijk, 2003;Khotimah, 2003;Sudibyo, 2001).
There are many pieces of literature examining how propaganda is made by extremist groups to influence the individual behaviour to radicalise and join these organisations (Andersen & Sandberg, 2018;Mahood & Rane, 2017;Ömer Taşpınar, 2009;Schlegel, 2018;Schmid, 2013;Thomas, 2016). However, only a few studies discuss the struggle for radicalism meaning by Islamic organisations and how they try to construct its meaning. We argue that the lack of a clear definition of radicalism has been exploited by organisations to establish their own meaning of radicalism. Therefore, this study aims to determine how the radicalism discourse is constructed by Islamic organisations and the implications of this construction. Specifically, this study aims to provide a comprehensive understanding of the radicalism meaning by two spectra of Islamic organisations represented by NU Online and Portal Islam. NU Online represents moderate Islam, while Portal Islam is an extremist Islamic organisation (Maulana, 2018).

Radicalism: Search for a Single Definition?
Although various disciplines have discussed it, the term radicalism is not well defined and still has an ambiguous meaning (Porta & LaFree, 2012;Sedgwick, 2010). This is due to increasing complexity and the added effort of each individual (or organisation) to define the problematic concept (Schmid, 2013). This lack of consensus definitions leads each researcher to make their definitions or justify using one another. However, the functional approach argues that radicalism is an increased preparation with a higher commitment to conflict between groups. Theoretical approaches to radicalism can be placed at two ends of the continuum, which is the Anglo-Saxon approach as a passive approach and Europe as an active approach (Neumann & Kleinmann, 2013). The passive approach emphasises that any belief, no matter how extreme, does not pose a threat to society as long as it is expressed peacefully. On the contrary, the active approach emphasises the cognitive aspects of the process and assumes that radicalism may lead to violence. Academics also debate the endpoint of radicalism. Some of them think it is a purely cognitive phenomenon (Neumann, 2013). Meanwhile, others think this phenomenon may lead to acts of violence (van de Weert & Eijkman, 2019) European states have become highly concerned with being alert to early signs of radicalisation processes. As a result, youth workers as well as other local professionals have been confronted with the task of detecting these early signs. However, despite training and increased knowledge, the question remains whether youth workers are sufficiently equipped to assess potential risks in youth who show no concrete plans for criminal action. In these cases, prevention targets ideas rather than violent behaviour. This article details qualitative results of a case study among Dutch youth workers and suggests that no clear framework exists for detection of radicalisation processes into (violent. The absence of a single interpretation for radicalism has been used as a gap by the media with various ideologies. Therefore, the media aims to construct the meaning of radicalism according to them. The radicalism discourse then transformed, re-contextualised by adding purpose to legitimate certain actions (Leeuwen, 2008).

METHOD
This study combined Corpus Linguistic (CL) and Critical Discourse Analysis (CDA) methods. This combination was performed, considering that one method's usage was not sufficient to interpret the linguistic patterns occurred and answer research questions. In this study, CDA helped to complete the sociopolitical context, which is missing from CL analysis, while CL helped with data triangulation problems in CDA. This combination is a means to complement the shortcomings of these two methods (Baker et al., 2008;Cheng & Cheng, 2015;Partington, 2015).
In the CL method, the corpus is defined as a text collection with much linguistic information to be studied (Hizbullah et al., 2017). The authors would reveal the emerging discourse by critically observing the patterns seen in the existing linguistic corpus. Therefore, the consequences of re-contextualising specific themes and how a word and sentence form certain groups and exclude other groups could be seen (Leeuwen, 2008). This study talked about how the chosen corpus interpreting radicalism and what the corpus do not talk about when they talk about radicalism.
To fully understand the discourse, we use the critical discourse approach from (van Dijk, 1993Dijk, , 1994Dijk, , 2013). Van Dijk's approach, which concentrates on media discourse and social cognition (Amerian & Esmaili, 2014), is relevant to this study's objectives. In examining discourse, van Dijk investigated the triangular relationship between discourse, cognition, and society. Consequently, this study did not attempt to reveal discourse as a productive act of social meaning (Fairclough, 2001) or a social-historical search (Reisigl & Wodak, 2001). However, it concentrated on mental representations of radicalism formed by the two media portals and the pragmatic consequences of the discourse.
This study started by building two corpora derived from news articles with the theme of radicalism on the NU website and Portal Islam, respectively, for a year from September 2018 to September 2019. Primary data in this study are articles that have headlines and contents containing the word radicalism found on the websites. The collection of articles was searched using the Google advanced search engine by employing the keyword of "radicalism site:nu.or.id" for NU Online and "radicalism site:portal-islam. id" for Portal Islam. Subsequently, we manually checked to match results from searches on Google and individual sites. Finally, we found 98 articles on radicalism on the NU website and 38 news on the Portal Islam website. The same method was also employed by several contentanalysis studies such as Abd Mutalib et al. (2017) and Opeibi (2019). The usage of the AntConc application (developed by (Anthony, 2019)) is intended to search for keywords, concordances, and collocations.
After having a list of words, we scrutinised the data and contextualised it by considering and examining its social, political, or historical context (Jaworska, 2016;Leeuwen, 2008). After categorising words and sentences based on particular themes, we carefully translated them from Indonesian to English and ensured that context and nuance were not reduced. Moreover, we explored the relationship between the words in the discourse to determine how this contributes to the readers' social cognition, which in turn forms mental representations and "us versus them" attitudes in society. This was obtained by exploring the discourse presented and uncovering the missing discourse on radicalism in both sites.

RESULTS AND DISCUSSION
This section contains findings obtained using CL techniques and assisted by the AntConc application. The results of our keywords, collocations, and concordances clustering are as follows:

Keywords
In the first analysis, we focused on the keywords from each corpus by thematically categorising the word list using a quick concordance analysis to determine which group they fit into. In this study, keywords are intended to see the meaning of a particular word and primarily why it was used.
We found six categories (TABLE  1), including social groups, verbs, figures, ideologies, and methods. The social groups in this study are a collection of institutions mentioned by two spectrums of Islamic organisations on their websites. The striking difference between them is that NU Online uses the social-religious groups under it while the Portal Islam mentions more government agents. NU Online states that it relies on organisational bodies which are suitable for certain segments. The words universities, schools, and mosques indicate the place of their movements. Portal Islam mentions several institutions such as KPK, TNI, president, century, MPR, and Golkar, which shows partiality or opposition to these institutions. Portal Islam actively provides comments on current issues, especially those which attack the government.
We also present a series of verbs to show their actions and determine how they are intervening in society through various activities. In the NU Online, we found words such as moderate, ward, understand, build, strengthen, and fortify. Meanwhile, on the Portal Islam, we found negative verbs such as corner, urgent, and violate. We also found verbs like involve, correct, influence, and sustain. Furthermore, we group words that denote an ideology defined as a particular

Collocation
This section focuses on the collocation analysis of the three words, which are most commonly found in both corpora, such as radicalism, Islam, and Indonesia. We focus on the verbs, adjectives, and nouns, which accompany the selected words. All collocations are categorised into several themes conducted by manual concordance analysis.
TABLE 2 lists the collocations for three frequent words in the two corpora, like radicalism, Islam, and Indonesia. There are three categories in radicalism, including adjectives, ideologies, and social groups. Adjectives, in this study, show how the words radicalism, Islam, and Indonesia are interpreted. Ideology is a system of thought which underlies individuals or organisations to act, while social groups refer to specific institutions.
In the collocation of radicalism, the words accompanying radicalism in NU Online include origin, virus, transnational, hatred, threat, and global. The words virus and origin indicate that they intend to prevent radicalism as the root of terrorism. They also describe radicalism as a global threat and transnational crime to show the severity of radicalism. We found the word fabrication and danger on the Portal Islam, showing that Portal Islam views radicalism as an issue which is not as big as what is reported by the media; it is only a political spin.
To show the word radicalism's ideology, we found words such as separatism, egoism, transnational, heretical, ideology, danger, intolerance, global, hatred, and threat in NU Online. NU Online illustrates radicalism as a dangerous, heretical, transnational, and global scale-like act of separatism, which is an act of spreading hatred and threats. Meanwhile, on the Portal Islam, we found words such as intolerance as a word which follows radicalism. This shows that Portal Islam interprets radicalism as intolerance.
In social groups, we found words such as PMII and teachers in NU Online, while on the Portal Islam, we found Ristekdikti, campuses, faculties, HTI, caliphate, police, and ministers. PMII and teachers in NU Online are the two central institutions of the deradicalisation movement, while Portal Islam continues to criticise the government by condescending Ristekdikti, campuses, and the police. The emergence of words such as the caliphate and HTI used as their attempt to clarify that the current state threat is not HTI and caliphate but communism.
In the collocation of Islam, we divided it into two categories, namely sects and social groups. Sects refer to the belief in a particular school of thought, which then determines how they worship, while the definition of social groups remains the same as the previous definition. NU Online talks a lot about washatiyah, grace, goriban, sharia, archipelago, caliphate, radical, and moderate, while on the Portal Islam, the word Islam is followed or follows the words allergy, archipelago, and radical. We also found social groups that accompany or are accompanied by Islam on NU Online, such as Islamic Organisations, Ministry of Religion, State, Media, FPI, and Universities. On the Portal Islam, words such as professor, officials, HMI, Golkar, terrorist, ministry, terrorism, TNI, KPK, and HTI are found to be used before or after the word Islam.
In the collocation of the word Indonesia, we found words such as separated, my country, and mubalighin on NU Online, while on the Portal Islam, we found words such as leading, above all, intolerance, terrorism, and radical. Lukman.) NU students are reminded to be aware of religious and secular radicalism The same opportunity to add a sign of the entry of radicalism in campus mosques is the activity Religious understanding has mushroomed, which has led to radicalism that threatens the Republic of Indonesia. Liberalism The BNPT will continue to monitor the issue of radicalism and terrorism in order to maintain the integrity of the nation Indonesia and Pancasila are following Islamic law, so do not impose and (apply the system to increase the understanding and insight of moderate Islam for these students by applying to jointly block the spread of understanding of radical Islam through the socialisation of tolerance and moderation in their respective fields while spreading friendly Islam and counteracting radicalism Its current develops critical thinking, while exclusive Islam tends to be doctrinaire, closed the middle of society. Negara Kesatuan Republik Indonesia is our big home that must be Terrorism nets still have their influence in Indonesia, from Darul Islam, Islamic State to Al. Rahmat stated that the level of radicalism in Indonesia has entered a yellow light or is alert, so that as well as terrorism, forbidden in Indonesia . "Santri must guard religion, not do it because at the same time, Indonesia is being swayed by a situation of radicalism Source: processed by authors, 2019 NU Online intends to show Indonesia as a country that is currently facing radicalism, while Portal Islam states that radicalism in Indonesia is just political spin. Like other collocations, we found social groups in the collocation of the word Indonesia. In NU Online, we found the words PKI, GNKRI, communist, HTI, Pakistan, and TNI, while on the Portal Islam, we found the words LIPI, HTI, Indonesia, Tempo, and universities.

Concordance
The third section focuses on the concordance of the two corpora. A concordance can express certain word associations in a sentence so that we use it to investigate certain discourses and knowledge which each corpus is trying to build. We show how the words radicalism, Islam, and Indonesia are used in the context of a news article in each corpus. Table 3 contains concordances for words such as radicalism, Islam, and Indonesia in NU Online. It is intended to investigate in what context the sentence appears. The word radicalism is described as a chaotic condition that must be abandoned and watched out for, currently wide-spreading, and the government continues to monitor it to maintain the integrity of the Republic of Indonesia. The word Islam in NU Online is often used to show moderate Islam, which is a friendly, inclusive, and tolerant religion. In NU Online, Indonesia is described as a country that is currently facing radicalism. Table 4 contains concordances for words like radicalism, Islam, and Indonesia in Portal Islam. We found that the Portal Islam seeks to confirm radicalism as a political game rather than a product of religion. Islam in the Portal Islam is described as a religion which is being cornered because it is considered to if it is correct, radicalism will disappear by itself. By the way, who All religions, which then conclude that radicalism and terrorism are not products of religion, moreover exposed to the radical thoughts of the chancellor. Radicalism is not based on ideology but on interests or corner Islam as a religion that produces radicalism and terrorism. This fact has been spat on Do not campaign against terrorism and being exposed to radicalism into terror and radical acts more sadly all Islamic based schools are suspected of being the birthplace of thoughts During the last decade Islam was cornered as the sole brand holder of the perpetrator Radicalism and terrorism do not belong to Islam because radicalism and terrorism are unfamiliar In order for the TNI to clash with religious Muslims, it must be removed with accusations Academics, scientific institutions, and Islamic religious institutions, fail to understand the three dynamics and concerns Got a place back even though career in Indonesia, the issue of radicalism is almost the same as the issue of terrorism to immediately leave Indonesia, according to him, there is no place in the hemisphere @prabowo is behind the Hizbut Tahrir Indonesia movement so far, his stance has always been on the side of them. radicalism to terrorism. He said Indonesia was facing a new threat through According to him, there is no place in any part of Indonesia for followers of ideology Source: processed by authors, 2019 deliver radicalism. Indonesia is described as a prominent country and should be proud of so that the issue of radicalism in Indonesia is not too severe.

Discussion
A discourse grows and lives in the corpus to construct cognition and legitimise certain actions for the reader. Each news portal re-contextualises a particular event by adding, changing, and removing some aspects of the event to suit their ideology (Karomani, 2004). This section explains how the corpus develops discourse on radicalism. Referring to van Dijk's (2003) approach, the discussion is divided into three levels which are at the first level, we dismantle the discourses in the text; at the second level, we explore the various knowledge arising from these discourses; and finally, we connect the knowledge into a mental representation of radicalism in society.
The results of the analysis of the corpus show that NU Online builds the meaning of radicalism as a virus that needs to be prevented and combated. The narrative is formed by combining several discursive strategies involving experts, bureaucrats, and opinion leaders. In this way, NU Online incorporates scientific authority, personal authority, and impersonal authority (Leeuwen, 2008) to re-contextualise the meaning of radicalism. Through the results of research, policy, and religious studies, NU Online forms a consensus that radicalism is dangerous because it threatens individuals and national integration.
Moreover, groups or people exposed to radicalism are translated into numbers and words such as "virus" to raise questions on their humanities. This labeling aims to generate negative sentiment towards radical groups and provide legitimacy and support for moderates to act according to the moral power given to their groups by the text (Baker et al., 2008;Leeuwen, 2008).
Portal Islam talks more about radicalism in the context of certain events. They build a narrative that radicalism is a political spin to gain public sympathy. Portal Islam builds a narrative that radicalism is fabricated to marginalised Muslims and is a black campaign against specific candidates in the election. In this narrative, Portal Islam suggests the public always to rethink when issues of radicalism emerge and to deem it will disappear by itself.
The discursive strategy used by Portal Islam is by re-contextualise various issues in a political context and then connect them to the issue of radicalism. Various issues ranging from the election of university rectors from abroad, remissions on corruptors, and presidential speeches are directed into radicalism issue, not to provide terror but to justify that radicalism is only a political spin and fear of radicalism is groundless. This is not conducted using the authority of experts. Portal Islam prefers a Twitter account with massive followers to provide information on various issues related to radicalism.
Each corpus is more interested in fusing the information function subtly into the propaganda function. The information conveyed aims to disseminate their group's ideology to readers. This can be seen from the forms of knowledge which emerge from the article, for example, a contextual knowledge ("…the entry of radicalism in the campus mosque is the physical activity conducted by the congregation in it. These activities usually lead to militarism…" and "…in most cases, the aim of the supporters of radicalism is students who are innocent or do not have a strong religious background…"), a knowledge of certain social groups ("… NU students are reminded to beware of religious and secular radicalism."), an opinion ("…Republic of Indonesia and Pancasila are following Islamic law. So, don't impose (implementing the caliphate system, for example) …") as the ones in NU Online. On the contrary, Portal Islam also disseminates fragmented knowledge such as knowledge which comes from a personal opinion without the basis of certain studies ("…so that the TNI can clash with religious Muslims, it must be removed with accusations…"), a knowledge of future predictions ("… radicalism will go away eventually…"), and social group knowledge ("…this concept needs to be treated seriously. Specifically, the people are worried about the rise of communism…"). Based on this explanation, the knowledge built upon the two sites is partial. Radicalism is never told in its entirety context with the social, economic, and political conditions of the subjects exposed to radicalism. This partial knowledge is used as a base by the reader to understand the issue of radicalism. Therefore, information about radicalism is limited and adjusted to the ideology and agenda of each corpus.
Considering that the two portals have a niche segmentation, hatred or cynicism against radicalism has turned into a mental representation of the group forming a prejudice against the outsider. The mental representation of "us versus them" becomes a moral reference for their readers to act. This corpus directs their actions.
The power-knowledge then recontextualises radicalism into a new form, either as a virus or as a political spin. Then, the discourse turns into a mental representation of each group, and policies related to radicalism emerge through these mental representations. That knowledge becomes a bridge for the emergence of discriminatory cultures in society.
Therefore, it is essential to dismantling the verbs which are often used in discussing radicalism. Verbs are markers of the actions to take related to emerging discourses. Through verbs such as moderate, ward, understand, build, strengthen, and fortify, NU Online urges all circles of society to take part in doing the same for those exposed to radicalism. It is similar to Portal Islam, which utilises many verbs such as marginalise, violate, and force to encourage readers to think that radicalism is used to enclose Muslims, and it will disappear by itself so that extremism needs to be maintained.
Without a clear context, radicalism is only judged by its participation in certain ideological groups. Finally, this will lead to policies or policy drafts that are discriminatory against society. This policy stems from the discourse on radicalism, which is built from cognition about radicalism that has been adapted to certain media ideologies and their agenda. This contradiction in discourse not only confuses the public, but the narrative of strengthening the group also provides legitimacy to discriminate against outsider groups. The two portals should also provide nuances of the complexity of radicalism, the subject exposed to radicalism, social, economic, political, and historical conditions from the events which occurred so that they cannot merely consider radicalism as a virus or a black campaign. Instead of helping the public, a simplification discourse on radicalism, such as in the two corpora, is counterproductive and may backfire in countering terrorism in Indonesia.

CONCLUSION
This study is intended to understand the construction of radicalism by two spectra of Islamic organisations. We found that as the media massively preach about Islam as a religion linked to radicalism and terrorism, these two sites, in their respective ways, try to dismiss the issue by presenting an image of moderate Islam on NU Online and clarifying that radicalism is a political spin on the Portal Islam. Our study found that radicalism in NU Online is considered a virus that needs to be prevented and combated through various preventive measures. Radicalism is narrated as a threat to individuals and the integrity of the nation. On the contrary, the narrative built by Portal Islam is radicalism as political spin. Specifically, Portal Islam engages people to rethink the issue of radicalism by not connecting it with Islam. Both of these meanings are dangerous because they remove the social, political, and economic context of radicalism and simplify it to the individual level. If it left unchecked, it would lead to discrimination acts towards people or groups deemed unsuitable for their group.
Academically, this study contributes to the study of radicalism by presenting the struggle for the meaning of radicalism by moderate and extremist Islamic organisations. Practically, this study helps practitioners and stakeholders to understand on how both portals convey ideology through content on the websites, and what types of messages or words are used to build public trust. Furthermore, this study proves that CL through keywords, clusters, concordances, and collocations help researchers conducting critical discourse research. These four elements facilitate the authors to dismantle the discourse on the two corpora in this study.
This study has several limitations, which can be used as recommendations for future research. The first limitation is the data collection period of one year. Future study may overcome this limitation by observing the news distributed at different times to see the development of the word radicalism every year or based on certain moments which contribute to the narrowing or expansion of the meaning of radicalism. Another limitation is the utilisation of two news portals as a representation of moderate and extremist Islam in this study. It can be expanded using Maulana's research (2018) for future studies. Moreover, as news portals shape the meaning of radicalism through science, policy, and opinion leaders, other studies need to examine more deeply by using other discourse analysis such as genealogy on how the power of science shapes policy and opinion leaders.