Printer Friendly

"Please don't be Turk, be Greek, be Armenian": agency and Deixis across virtualized Turkish imagined community.

While engaging in Twitter analyses, researchers have taken several different routes. As Moe (2012) points out, most Twitter analyses have used large quantities of tweets. Indeed, since its launch in 2006, Twitter users have been supplying researchers with immense amounts of data. So far this has resulted in generating user models and network analyses employing tens of thousands to millions of tweets (Abel, Gao, Houben, & Tao, 2011; Moe, 2012). Using smaller data sets is quite uncommon, and little research has used corpora analysis (Kouloumpis, Wilson, & Moore, 2011, p. 538). While Twitter analyses often lead to networks between groups and users, they also have an underexplored dimension. Treating tweets as conversations (Boyd, Golder, & Lotan, 2010; Honey & Herring, 2009; Huang, Thornton, & Efthimiadis, 2010) is an emerging research path in Twitter analysis.

Framing tweets as conversations helps us engage them linguistically, but this new path has several issues. Conducting critical discourse analysis (CDA), this article addresses some questions regarding language use in a virtual space that might be taken for granted in ethnographic or other text based research. These issues relate to the immediacy of the context and include the location of spatial deictic markers in a virtual space without borders; the necessity of using a deictic center for spatial references; the issue of transference of morality as well as dominant discursive structures that arise with the use of a deictic center and empathetic deixis; and the contestation of agency and national borders within a virtual moral geography.

However, treating Twitter messages only as conversations is also problematic due to the number of users and the nature of interaction. Although users respond to each other using the @ sign and engage in what appear to be conversations (Honey & Herring, 2009), these conversations on Twitter involve a much larger number of participants. The data set for this research is a discussion between more than 50 users, generating a total of 65 tweets. Moreover, Twitter brings together people from very different sociocultural and geographical backgrounds and allows them to engage in a discussion on identity, culture, and conflict. Twitter fosters discussion between people who would not have come together to talk about politics in any other way. Therefore, Twitter bends certain rules we take for granted in ethnomethodological or other interaction-based research such as time, space, social hierarchy, and social proximity. As users respond to one another's tweets and take sides, instead of a polysemic discussion, the voices of more than 50 users are reduced to two conflicting discourses. Through this reduction, users who identify on the same side create virtual imagined communities in which their tweets can be read as a single text.

This article addresses these issues by analyzing a Twitter discussion revolving around an intercultural conflict: a heated debate between Turkish nationalists and liberals over the murder and the court case of an Armenian Turkish journalist. The research questions this article aims to answer are: How does the online context shape the linguistic properties of public debate, and how does public online interaction sustain, reproduce, or challenge discursive power structures? By examining online interaction through CDA, this study contributes to our understanding of linguistic mechanisms in an online context as well as their connection to power and agency.

The article first provides a brief historical background of the case study and expands on concepts such as nationalism and liberalism to contextualize the discussion and the analysis. Next, I explain the methodological and theoretical lens through which I conduct the study. Then, I present the analysis and discuss the practical as well as theoretical implications of the findings.

Background and Context

On January 19, 2007, Hrant Dink, a journalist of Armenian descent and a Turkish citizen, was assassinated in front of his office by a teenager named Ogun Samast and Samast's mentor, Yasin Hayal (Tavernise, 2007). International organizations such as Human Rights Watch (2007) and Amnesty International (2007) suggested that the assassination was a reaction to Dink's outspoken nature and his writings, which supported the Armenian stance regarding the 1915 Armenian genocide debate. After the assassin was caught, however, the press uncovered pictures and videos showing that the police glorified Samast as a hero after the capture. The state commenced a court case and additional investigations to solve the case of Dink's assassination. Investigations brought to light Hayal's alleged links with the Turkish Intelligence Agency, lost evidence and camera footage, and contradictions in the case files. These discoveries raised doubts regarding a wider conspiracy that included the state officials' involvement in the assassination (Turgut, 2012). On January 17, 2012, however, the case was closed, and the court decided that Ogun Samast and his mentor, Yasin Hayal, were solely responsible for Dink's murder.

Immediately after the assassination, those who condemned the murder gathered to protest. They held signs that read "we are all Armenians." The same signs and slogans were used to protest the court decision five years later. On January 18, 2012, after the Hrant Dink case was closed, Ahmet Hakan, a political journalist and television host, tweeted to his followers that he was going to host the lawyer of Hrant Dink's family to discuss the murder. As a response, Hakan received tweets from nationalist as well as liberal followers. The central question among the nationalists revolved around one particular slogan that was present in the protests: "Did you also become an Armenian?" After Hakan tweeted that he indeed became an Armenian, he received a host of protest tweets from nationalists and a few from liberal supporters. Ahmet Hakan retweeted some of these nationalist and liberal supporter tweets.

This analysis examines all the tweets that Ahmet Hakan retweeted. As a follower on Twitter of Hakan, I was able to observe the retweets as he was sending them in real time. Without interfering with or participating in the discussion, I collected all the tweets starting from Hakan's first tweet that initiated the discussion until his last for that day--a total of 65 tweets. With these tweets, participants contest and define the borders of being Turkish, Armenian, Muslim, and Christian. Through their discussion we can observe the linguistic location of agency of "being" and "becoming" Armenian within competing discourses. In the process, the exchange also raises some larger questions: Who has the power to be or to assign identities? Is this solely within the reach of the dominant discourse? Or can the contestation live up to a moment of resistance--and, if not, what does it live up to? The analysis, therefore, examines an instance of "talking back" in Turkey. In addition to its linguistic features, the discussion offers insights into the nationalist imagery of a nationalist order of discourse and the sociolinguistic construction of social borders in social media.

Orders of Discourse

To further contextualize the event, I investigate the discursive power behind the tweets. The Twitter discussion cannot be analyzed separate from the discursive power structures that enable their reproduction. In this case, the question is: What are the discursive power structures that encourage such reaction among some of the followers of Ahmet Hakan? To make the connection between power, discourse, and the tweets under consideration, it is essential to outline how power and discourse are related. Discourses, and Twitter messages, are more than just social practices. Discourse "also serves particular ends, namely the exercise of power" (Jager & Maier, 2009, p. 35). Van Dijk suggests that power should be understood as "preferential access to, or control over, public discourse" (Van Dijk, 2008, p. 14). This access and control over the public discourse asserts the influence of power because it frames and limits the way thought, talk, and behavior are shaped. Fairclough (2001) expands this idea of power as control on two dimensions: power acts both within and behind the discourse. "Power in discourse is to do with powerful participants controlling and constraining the contributions of non-powerful participants" (Fairclough, 2001, p. 39). These constrains include content, relations, and subject positions. Power behind discourse suggests that "the whole social order of discourse is put together and held together as a hidden effect of power" (Fairclough, 2001, p. 46). These overt and covert relationships between power and discourse influence the discourse itself. Discourse with this effect of power conforms to particular conventions. Borrowing from Foucault, Fairclough (2011) calls these conventions "orders of discourse ... social structuring of semiotic difference, a particular social ordering of relationships between different ways of making meaning--different genres, discourses, and styles" (p. 11).

I classified the 65 tweets in this analysis into two such conventions: the nationalist order of discourse (the NOD) and the liberal order of discourse (the LOD). I classified those who reacted against Ahmet Hakan under the banner of NOD, because throughout the tweets, they consistently refer to themselves as nationalists. I classified those who supported Hakan under the convention of LOD, but not because they display an order-of-discourse solidarity similar to their nationalist counterparts. While those who support Hakan do not define themselves as explicitly as nationalists do, Hakan declares himself a liberal. Therefore, for the sake of the analysis, and without imposing an identity, I named the discourse that supports Hakan's liberal position as LOD.

Those who associate with nationalism comply with the NOD, which has some generic discursive aspects such as elitism and gendered ethnocentrism based on a conservative "blood and soil" rhetoric (Choudhary, 1988, p. 41; Erlank, 2003, p. 655; Parekh, 1995; Sorenson, 1996). Within the tweets, however, those who ascribe to the NOD further relate with a specific, local version of this order of discourse--namely, Kemalism. Kemalism is a discourse named after the founder of the Turkish Republic. This discourse is authoritative, nationalistic, gendered, and ethnically and religiously exclusive (Ergundogan, 2012; Ozcer, 2012; Solgun 2012; Tayman, 2012). Kadioglu (1996) describes the roots of Kemalist discourse as such:
   Turkish nationalism contained the premises of an Enlightenment
   mentality as well as a brand of Romanticism. It purported to
   synthesize the materialism of the West and certain indigenous
   cultural traits such as Islam, as well as pre-Islamic Turkic
   traditions. (p. 191)

This synthesis between ethnocentrism, Eurocentric materialism, and religious fanaticism influenced the nationalist discourse in Turkey as well as reproduction of this discourse through institutions such as education, economy, culture, and bureaucracy (Smith, 2005). The NOD, therefore, is a result of this reproductive practice, and the power behind its discourse is the support of the state as well as its historical discursive structure.

Those who attempt to oppose the NOD do not have a unified self-definition similar to the users of the NOD. Instead, they show a sense of solidarity with Ahmet Hakan, the journalist who started the discussion. Hakan claims to be a "liberal"--adhering to principles such as "the minimal state, individualism, social differentiation, the rule of law, and human rights and liberties" (Caha, 2001, p. 35). While those who assist Hakan might believe that they benefit from a power position because they tweet with a famous journalist, neither Hakan nor his supporters have a power position other than the fact that they think they are talking back. Consequently, the Twitter discussion takes place with a power imbalance.

The NOD is a discursive position that is endorsed by the state. In addition, the NOD uses a discursive structure that is well established over decades. The LOD, on the other hand, has a perception of self-efficacy and does not have any discursive or linguistic structure alternative to the NOD. This results in an interesting juxtaposition of seemingly opposing discourses, as the analysis reveals.


Although different scholars emphasize different aspects of discourse, certain underlying principles can be found across definitions. Some define discourse "as socially constructed ways of knowing some aspects of reality" (Van Leeuwen, 2009, p. 145) and a semiotic modality that mediates between "structure" and "events" (Fairclough, 2011, p. 11; Reisigl & Wodak, 2009). In these accounts, structure refers to larger, abstract social practices as well as axiomatic patterns and taken-for-granted assumptions. Events, on the other hand, are tangible and visible experiences in daily life, such as a Twitter discussion. Because they are socially constructed, discourses are also social practices. This means that discourses have a "dialectical relationship between particular discursive practices and the specific fields of action (including situations, institutional frames, and social structures) in which they are embedded" (Wodak & Reisigl, 2003, p. 382). However, scholars point out that discourses are not mere reflections of social practices. "Discourses not only shape but even enable (social) reality. Without discourses, there would be no (social) reality" (Jager & Maier, 2009, p. 36). From these accounts, discourse can be understood as social practices mediating between social structures and daily interactions, which not only reflect but also mold the social reality.

One aspect of discourse that needs further elaboration is its role as a mediator between social structures and daily interactions: how it operates between what are called the macro and the micro levels. Van Djik (2009) refers to these social structures as "semantic macrostructures" and to daily interactions as "local meanings" (pp. 68-69). He suggests that "local meanings are a function of the selection made by the speakers/writers in their mental models of events of their more general knowledge and ideologies" (Van Dijk, 2009, p. 69). While analyzing local meanings might outright reveal their sources of influence--for example, via lexicon--linguistic location of semantic macrostructures might not always be as apparent. Sometimes we might need to reconstruct how an event implies those semantic macrostructures.

This reconstruction occurs through the concept of context. What makes context so important is that the aforementioned reconstruction of the context can reveal essential ties between the text and social structures. According to Gumperz (2003), Goffman established this link by suggesting the concept of "interaction order"--referring to "a distinct level of discursive organization bridging the linguistic and social" (Gumperz, 2003, p. 216). This article constructs those ties by exploring how the social context manifests itself in tweets. Context can be defined as both the physical environment in which the event takes place and the taken-for-granted "shared cultural knowledge" (Paul Gee, 2011, p. 6) used within the text. Van Dijk (2008) adds a cognitive dimension to context and suggests that it can be "defined as the mentally represented structure of those properties of the social situation that [is] relevant for the production or the comprehension of discourse" (p. 90). Context is any structure that surrounds the reproduction of text--physical, cultural, or mental. These surroundings are often axiomatic and are not spelled out completely by their users. Certain tools of CDA, explained below, allow us to forge connections between the text and the context.

Context, through its connection to dominant discourses and other axiomatic beliefs, falls squarely within the realm of CDA, whose mission is to analyze "the way social power abuse, dominance, an d inequality are enacted, reproduced, and resisted by text and talk in the social and political context" (Van Dijk, 2003, p. 352). CDA attempts to expose such forces by examining the relationship between semantic macrostructures, local meanings, and how one influences the other: "It is precisely in these macro-micro links that we encounter the crux for a critical discourse analysis" (Van Dijk, 2009, p. 83). Through examining this relationship, CDA questions the assumptions people often make. These assumptions, and their manifestations in discourse, are one of the most important ways in which a social group reproduces power abuse.

Tools of Analysis: Deixis, Metaphor, and Agency

To connect the event, the Twitter discussion, with the implied context as well as larger social structures, I employ certain tools used in CDA. I study the use of place and person deictic markers and metaphors to investigate the linguistic construction of agency, national solidarity, and morality of the national territory. These tools will draw connections between text and context, since "the sociopolitical force of ideologically constructed social boundaries give[s] symbolic meaning to linguistic practice" (Bucholtz & Hall, 2004, p. 477). These tools will also help us address the linguistic problems stated in the introduction--such as the sociolinguistic location of agency, the use of place deictic markers within a virtual realm and the transference of morality that is attached to such use--and expand our understanding of language in virtual space.

Space Deixis

Deixis is the principal tool used in this analysis. Deictic markers are "words whose reference must be determined from context. Common deictic words fall into the categories of person ... place ... and time" (Paul Gee, 2011, p. 8). According to Levinson (1983, p. 54), deixis is the most important way in which the relationship between text and context can be visible. Brown (1995) explains the reason behind this function of deixis:
   Deixis encodes features of the context of utterance, hence its
   interpretation requires the listener to be able to appreciate, and
   if necessary to construct, the relevant context. In cases where the
   entities, events and locations are not previously known to the
   listener, the listener can use the deictic features of the
   utterances to create an appropriate prototypical context. (p. 109)

In short, "deictics tie speech and writing to context" (Paul Gee, 2011, p. 9); in the virtual realm, however, the context is a bit more complicated. If a researcher is interviewing a person about U.S. politics in a park in Washington, DC, then the context could be constructed with a little effort. But what happens when people discuss Turkish politics on Twitter? As a researcher living in the United States, I have no information about their physical contexts. The possibility of a geographical context is completely useless, because people might be tweeting about the subject from another country and still talk about Turkey as "this country." The best guess I can make as a researcher is to suggest that any reference to a space marker in the virtual realm is that user's mental construction of the referred physical space. In other words, in the virtual realm, the only useful definition of context is the one put forth by Van Dijk (2008, p. 90) as mental constructions.

To make more sense of the spatial deictic markers in the virtual space, two important concepts are deictic center and empathetic deixis. The deictic center is "an egocentric locus in which the Speaker (and the speaker's place and time) provide the parameters in relation to which a reported event is situated" (Schiffrin, 1990, p. 249). Since an online discussion does not take place in a physical space, the only spatial contextualization we can create depends on the users' linguistic construction of the deictic center. With the idea of deictic center, however, comes the concept of empathetic deixis, which

arises when the speaker appears to transfer the deictic center, at least for some period of time, to the listener, and it appears to be a product of the movement out of the canonical situation of utterance, when the speaker speaks of some entity or event not in the here and now, and not mutually visible. (Brown, 1995, p. 112)

Talking about a country in a virtual space is an "entity not in here and now, and not mutually visible" (Brown, 1995, p. 112). Therefore, in a virtual space, spatial markers need to be understood as deictic centers--since a physical context is not visible--and same mutual invisibility makes these deictic centers empathetically transferred.

Transference of the deictic center through empathetic deixis becomes an important issue when the conversation revolves around differences of identity. When a user tweets about Turkey in the midst of a political discussion, that spatial reference is not free of normative statements. This makes the empathetic deixis carry an extra load of morality in addition to the deictic center. In other words, the deictic center in a virtual discussion is not free of attempted impositions of dominance, hegemony, and morality.

Connections between geographical space and concepts such as morality, hegemony, and dominance invoke the idea of moral geography (Hill, 1995; Leap, 2010; Modan, 2007). Moral geography refers to the process through which "spaces [are] associated with ... values" (Hill, 1995, p. 111). Modan (2007) expands on Hill's definition and suggests:

A moral geography is an interweaving of a moral framework with a geographical territory. Through the use of various discourse strategies and themes, community members create alignments and oppositions among people and places. These alignments and oppositions are then evaluated positively or negatively in relation to various value and belief systems circulating the community. (Modan, 2007, p. 90)

Although Modan writes about a specific community situated in a neighborhood and in a city, I use her description analogously. Just as I take references to space in terms of deictic centers connected to morality, I adopt the idea of community in terms of use of plural personal pronouns and discursive positionings.

Person Deixis

Person deictic markers in an online discussion raise questions for linguistic analysis. In the Twitter discussion under examination here, users often engage in the discussion from first-person plural and second-person plural points of view. It is quite remarkable that the more than 50 users participating in the discussion probably do not know each other in person and likely never will. Yet, as they share the same discursive position, they write "we" and "you all" repeatedly. This phenomenon is especially interesting, since a Twitter user most likely is tweeting alone using his or her computer or mobile device.

As mentioned in the introduction, the discussion does not fit the usual CDA structure, because it is not an exchange between two individuals. There is a remarkable cohesion between users who take the same discursive position. Tweets of either discursive position--the NOD or the LOD--could almost be read as a single, unified, and cohesive text.

Similar to the use of spatial deictic markers imploring the idea of moral geography, cohesive use of first-person plural deictic markers--especially in the context of nationalism--recall the concept of "imagined community" (Anderson, 2006). Through a historical analysis, Anderson suggests that the "convergence of capitalism and print technology on the fatal diversity of human language created the possibility of a new form of imagined community, which in its basic morphology set the stage for the modern nation" (Anderson, 2006, p. 48). Anderson's emphasis on the juxtaposition of technological development and language is not misplaced. Another technological revolution centered on language, Twitter, creates yet another level of imagined community: a virtualized imagined community. As shown in this analysis, the cohesion created through first-person deictic markers--especially across NOD tweets--conjures the sense of national solidarity. Urban (2001) suggests that this phenomenon of reproducing the imagined community using the first-person plural can be seen as "metaculture." He suggests that the "we" in a nationalist text assimilates the agency of the proposed and possible participants to the conversation. As such, the assimilating "we," according to Urban (2001), is at the heart of constructing this imagined community.


Metaphorical analysis draws mainly from Semino's (2008) interpretation of Lakoff and Johnson (2003). While they define metaphor simply as "understanding and experiencing one kind of thing in terms of another" (Lakoff & Johnson, 2003, p. 5), Semino (2008) applies their ideas to contesting and conflicting discourses. She claims that different uses and juxtapositions of metaphors represent a specific version of reality, and such representation is situated within a specific discourse (Semino, 2008, p. 90). Similarly, metaphors that are the focus of the present analysis provide gateways toward understanding the underpinnings of linguistic representations.


One of the main investigations this study undertakes is about agency and its relationship to hegemony and dominance. As Ahearn (2001, p. 111) explicitly states, there is an inseparable interconnectedness between language and power. One way this interconnectedness manifests itself is through attribution of agency. In Ahearn's (2001) terms, agency is "the socio-culturally mediated capacity to act" (p. 112). The dominant discourse, in its attempt to impose a set of universal truths, imposes a set of identities as it sees fit and frames the sociocultural capacity to increase its own dominance. While such imposition is observable in the NOD tweets, those who resist NOD struggle to define who they are and liberate themselves from the constraints of the NOD.

The Analysis

In this section, I engage in a close reading of the 65 tweets and employ linguistic analysis to show how the use of deixis and metaphor by the NOD and the LOD define national as well as social borders of being Armenian, Turkish, and Muslim. (2) I also discuss how these linguistic markers devise agency--the discursive contestation between the NOD and the LOD for self-determination of identity. I demonstrate how the LOD tweets, though trying to create a stance of resistance, enforce the dominant discourse.


The nationalist order of discourse operates within a nationalist genre--though it certainly is not the only genre to which the NOD belongs. The NOD has elements from religious fundamentalism to sexism, from ethnocentrism to xenophobia. Through the use of third-person deictic markers, the NOD not only constructs a virtualized imagined community but also frames the liberal order of discourse as an incompatible identity. Moreover, through the use space deictic markers, the NOD imposes a moral geography regarding the deictic center, Turkey. Even in the first several tweets, it is possible to witness these dynamics in play:

1. this evening in unbiased zone there is hrant dink's lawyer fethiye cetin. subject: "dink murder with all its aspects". (ahmethc)

2. What will happen when you do, did you also become an armenian suddenly, though we witnessed you turning 360 degrees but, I think this is a little much I don't love you any more.: ((ea)

3. to the one said "did you also become armenian?": I did, so what? is it hard on you? (ahmethc)

4. Ahmet Hakan, what is it to you, is it your job to announce the Armenian protest. Who knows what improper things he was doing in that newspaper (ja)

5. against all the racist fascists let us shout out with love: we are all armenian, have you anything to say? (ahmethc)

6. Mr Ahmet, no one is racist, only you and likes of you have a little large denomination, doubtlessly you DONT HAVE any country love! (ea)

Here, the 2nd, 4th, and 6th tweets are NOD tweets, and the others are tweeted by Ahmet Hakan. Although it is the same user, the use of first-person plural markers in the 2nd and the 6th tweets is a common pattern throughout the discussion. The "we" in the 2nd tweet and "no one" in the 6th tweet refer to the virtualization of the imagined community and creation of the nationalist solidarity against the Other. During the process of constructing in-group solidarity with the use of first-person plurals, the Other is further marginalized, and the lines between groups are further solidified. The differentiation between first-and second-person plural deictic markers establishes a mutually exclusive dichotomous relationship between the NOD and the LOD as well as between implied Turk and Armenian, which is picked up and repeated for the rest of the interaction.

The idea of spatial deixis as an imposition of moral geography is introduced subtly in these tweets. In the 4th tweet, the user suggests that Hrant Dink was engaged in "improper things in that newspaper." "That newspaper" is an Armenian newspaper located in Turkey, which implies that there is a "proper" way of doing journalism in Turkey. The user's idea of being a proper journalist in Turkey stems from his or her conviction of morally acceptable behavior in Turkey. Therefore, by classifying an Armenian journalist as "improper," the user imposes his or her idea of Turkish morality. Through such infliction, the user further justifies the murder. When the user states, "who knows what improper things he was doing in that newspaper," it is not too much of a stretch to suggest, "so it is OK to murder him." Within the NOD, the Other is a threat to be eliminated. Therefore, the user's moral construction of Turkey classifies assassinations against the Other as morally justifiable, while journalism can be labeled as "improper."

Some other NOD tweets expand on the moral definition of being Turkish:

6. Mr Ahmet, no one is racist, only you and likes of you have a little large denominations, doubtlessly you DON'T HAVE any country love! (ea)

15. you being Armenian will not be hard on anyone. Especially if its not hard on you and disrespectfully you can say I am Armenian. (am?)

22. You possess a character more colorful than a chameleon and a wide denomination ruthlessly competing with the emptiness in space, it is normal for you to become Armenian. (md)

26. noo, its not hard on me it fits you well good luck with it human scrap (fd)

29. Immediately change the Glorious Muslim Turkish name that you are not worthy of. (oc)

32. as the grandsons of a great nation who ruled the world the only thing that can be said is HOW HAPPY IS THE ONE WHO SAYS I AM TURK. (v1)

37. We are not fascists, Ataturk Nationalists, we are not like you ignorant, sold and shifty ahmet hakan! (mg)

50. In this country every bite passing down your throat is ill-gotten!! If you have a little honor also let go of the identity in your wallet, and the Turkish Lira in your pocket. (bj0)

58. weren't you from Yozgat where did Armenian come from ... what you eat, what you drink even the air you breathe belongs to this nation. Don't lose your respect ... (ge)

The NOD tweets not only construct an image of a "good citizen" through the use of certain adjectives, they also prescribe sets of feelings and behaviors one should have and do in order to be accepted as a Turkish citizen. In the 6th tweet, the user suggests that Ahmet Hakan does not have any love for his country. According to this user, then, the correct way "to love" the country would be to ascribe to the NOD and operate within its framework. Therefore, the morality of the "country" is constructed in such a way that it permits only one kind of love, which is displayed through the NOD tweets. In the 15th tweet, the user further classifies Ahmet Hakan as "disrespectful." Because being respectful requires an understanding of morality, being disrespectful means that morality was transgressed. The same word is repeated in the 58th tweet. Being born on the land, according to the user, suggests that one "belongs" to the land and being born into a certain morality. Also in the 58th tweet, the user equates denying the original land with denying morality. In the 50th tweet, "ill-gotten" refers to a religious concept. An analogy would be eating nonkosher. The 50th tweet emphasizes the idea of imposed moral geography with reference to spatial deixis. Through the connection between "country" and "eating ill-gotten," the user dialectically juxtaposes being religiously and morally pure with being nationalist. The user highlights this by connecting "honor" and "letting go of identity."

Disagreeing with such moral imposition of the "country," Ahmet Hakan faces scrutiny, which can be observed in the contrasted use of adjectives. Being at odds with the country's morality, Hakan is portrayed as having a "large denomination" (in the 6th and 22nd tweets, having a large denomination is an expression that amounts to having no principles); "disrespectful" (15, 58); "human scrap" (26); "inglorious" and "unworthy" (29); "sad" (32); "ignorant, sold and shifty" (37); and "ill-gotten" and "dishonored" (50).

One important metaphor that needs to be explained is the notion of "turning." Being a turner can mean having no principles (similar to "having a large denomination"--in other words, no conviction), but it also has a gender component. Some of the tweets that reference turning include:

2. What will happen when you do, did you also become an armenian suddenly, though we witnessed you turning 360 degrees but, I think this is a little much I don't love you any more.: ((ea)

27. latest version of being a turner is becoming an Armenian ... no one will be surprised if you say I changed religion tomorrow. (mi)

42. Mr ahmet nowadays you turn a lot the other day you were turk, when did you become Armenian in two days you will be Israeli ... (yea)

60. you are a puppet if you don't say on live broadcast you are not a man everyone should understand who is what ... (p82)

65. It is hard on us Armenian. It is hard on us that a man like you is treated like a man in Turkish media. Please don't be Turk, BE GREEK BE ARMENIAN (ao)

Within this metaphor, and in public discourse in Turkey, "to turn" can refer to the act of coming out or to a male-to-female sex-change operation. This gender metaphor suggests that the "normal" is the patriarchic NOD, and positioning differently from this normality is considered "turning"--a perverse deviation from whatever is the imposed normality. Therefore, use of this metaphor not only binds the nationalist discourse as well as the "nation" with patriarchy, it also accuses any move away from the imposed patriarchic normality of the NOD--that is, being a man--with evil, perverse corruptness. This can be seen clearly in the 65th tweet, when the user says, "it is hard on us that a man like you is treated like a man in Turkish media. Please don't be Turk, BE GREEK BE ARMENIAN." This equates being a man with being Turkish; further suggesting being an Other is being a turner, or being a woman--with underlying implications of being weak, dirty, and unworthy.

NOD tweets accomplish several things. Through the use of spatial deictic markers, they impose a certain discourse of morality. This morality, while it permits murder, disapproves of claiming another identity if one is born into a land. This is consistent with the "blood and soil" obsession of the NOD. Through the use of personal deictic markers, NOD tweets both unite and divide. As they construct a virtual imagined community and national solidarity through the use of first-person deictic markers, they also clarify their borders with the Other through the use of second-person deictic markers. NOD tweets enhance this process of "othering" through the use of derogatory adjectives, which the users associate with being Armenian, Greek, and Christian. Finally, through the use of gender metaphors, NOD tweets associate their own discursive position with masculinity while identifying the Other as a weak, deviant perversion. NOD tweets temper the strong sense of morality, identity, and masculinity with the idea of being born into, and therefore "belonging" to, the nation.

This sense of belonging is also the landscape of determining agency; belonging automatically forgoes any sense of personal agency. The territorial reflex of the nationalistic discourse is reflected within the will of the NOD as an utter and complete surrender, as well as assimilation, of its subjects. Any deviation is strictly corrected by agents who accept and willingly enforce the universal reality they live in to the others. If we recite Ahearn's (2001) definition of agency as "socioculturally mediated capacity to act" (p. 112), the NOD's attempt can be interpreted as bordering and constraining the sociocultural capacity to allow action and identification only within the narrow limits of being a heterosexual male Turkish Muslim. If one were to disagree with those conventions, as it says in the 50th tweet, it is suggested that one "let go of the identity in your pocket."


The liberal order of discourse, through its "talk-back," attempts to establish an opposition to the NOD. Of the 65 total tweets, 17 belong to the LOD. These 8 tweets are exemplars of the 17:

5. against all the racist fascists let us shout out with love: we are all Armenian, have you anything to say? (ahmethc)

7. how many yasin hayal, how many ogun samast are there in the country (ahmethc)

8. blocking blocking doesn't end these white berets (ahmethc)

17. to those say "why are you making these rt": will only I see the truth of the country, you see too. (ahmethc)

35. alas! fascism poisoning. this damned thing doesn't have an anti-venom. it is best I escape (ahmethc)

61. instead of being a Muslim Turk like Samast it is certainly much more valuable to be a Turkish Armenian like Hrant ... (sb)

63. you filled my screen with all the fascists who don't look at human as human, you cooled me from life I swear. (sz)

64. No one understands you, I am also Armenian if need be I am also turner. So what to who! (rs)

In terms of personal deictic markers, the LOD tweets are similar to the NOD tweets. The use of the first-person plural to refer to the self contrasts with the use of the second- and third-person plurals used for the Other. Unlike the NOD tweets, these users do not construct a national solidarity. They do, on the other hand, also virtualize an imagined community. This virtual bond, just like its NOD counterpart, strengthens the in-group tie while delineating the border between "us" and the Other. Similar to the NOD tweets, users of the LOD associate the personal markers that refer to the Other with derogatory adjectives. Users of the NOD are labeled as: "fascists" (5, 35, 63); Yasin Hayal, the assassin's mentor, and Ogun Samast, the assassin (7, in the 8th tweet, "white beret" refers to the hat the assassin was wearing on the day of the assassination); "poison" (35); "invaluable" (61); and "who can't see humanity" (63). These adjectives and descriptors, combined with dichotomous personal markers, enforce the incompatibility between the users of the NOD and the LOD.

The polarization is so engraved within the discourse that an attempt at equality cannot escape it. In the 54th tweet, the user writes, "I say I am Armenian but I am TURKISH no to discrimination RT." This tweet shows a reaction to differentiating by saying "no to discrimination." In the same tweet, however, the user claims, "I am Armenian but I am TURKISH." The utilization of hypotaxis, "but," points to a position difference between what comes before and what comes after. As the user displays the nature of this difference by the use of uppercase type, he or she glorifies the national identity of being Turkish over being Armenian.

The LOD agrees with the gendering metaphors used in the NOD. The 64th tweet states, "No one understands you, I am also armenian [sic] if need be I am also turner. So what to who!" The user employs the "turner" gender metaphor that is also used by the NOD. The NOD frequently uses this metaphor to refer to the Other who "turned" from the normal way of being. In his or her attempt at talkback, the writer of the 64th tweet agrees with the gender metaphor without trying to change the meaning and accepts the imposed patriarchy. By claiming "if need be" the user frames accepting the patriarchy as a necessary sacrifice to defend his or her position. This illustrates the lack of an alternative discourse, or language, to resist the NOD.

Ahmet Hakan (ahmethc) considers these discussions and the NOD as "the reality of the country" in the 17th tweet, when he says "to those say 'why are you making these rt': will only I see the truth of the country, you see too." The "truth" value and absolute imposition of this black-and-white perspective is emphasized by the use of the positivist "seeing is believing" metaphor. Moreover, in the 8th tweet, Hakan claims that "blocking blocking doesn't end these white berets." In the 35th tweet, Hakan also talks about the "damned thing," fascism, as a poison when he writes: "alas! fascism poisoning. This damned thing doesn't have an anti-venom. it is best I escape." While the poison needs to be eliminated with an antivenin, Hakan states that such an "anti-venom" does not exist. Through these metaphors, Hakan taps into the narrative of a relentless, unstoppable evil and further demonizes the discursive Other.

At first glance, the use of space deictic markers--and, therefore, contestation of a moral geography--might seem missing from the LOD tweets. The lack of reference to the land, however, is more of a submission to the norms of the NOD than a contestation. In the 17th tweet, Ahmet Hakan writes, "will only I see the truth of the country, you see too." The "truth of the country" refers to the N OD tweets. Accepting the NOD as the "truth" and tying it to the "country," Hakan confirms the morality of the NOD. This endorsement further imposes the moral geography in the NOD tweets and reproduces the dominant discourse.

The LOD tweets perform similarly to their NOD counterparts. In terms of personal deictic markers, they also refer to self and to the Other in plural pronouns. In addition, they associate the Other with negative adjectives. Similar to the NOD users, they virtualize their imagined commu nity of "liberals"--without the added touch of national solidarity--and differentiate themselves from the Other through these linguistic moves. Metaphorically, users of the LOD tweets perceive the NOD users as a threat that needs to be eliminated. In addition, they conform with the gender metaphors put forward by the NOD. Another aspect of how the LOD users conform with the NOD users is in the morality of the space. Ahmet Hakan acknowledges that the moral imposition of the NOD users are the "truth," and other LOD users contest neither the NOD's imposition nor Hakan's acknowledgement. Hence, while the LOD users do have an opposite and confrontational stance against the NOD on some aspects, they establish it through the same grammatical and lexical resources that the NOD uses.

Discussion and Conclusion

To answer questions at the intersection of linguistic analysis, online interaction, and reproduction of power, this study of a Twitter discussion found that certain elements of the interaction speak to different approaches in discourse analysis. For instance, the exchange is similar to a conversation. Users refer to each other, especially to Ahmet Hakan, and respond to other users. While traditional critical discourse analysis focuses on an exchange between two people, the analysis here is very much like a conversation between two competing positions. But the exchange also lacks certain aspects of a conversation, such as turn taking, because Ahmet Hakan controlled which tweets got retweeted. On the other hand, the NOD and the LOD refer to each other in the first-person plural, assert a common identity,

and construct a virtual imagined community. These properties enforce cohesion across users and enable a textual reading of the tweets. While there is not a developed method of Twitter analysis akin to critical discourse analysis or narrative analysis, it might be useful to start thinking about common discursive and linguistic patterns of online discussions. The practice of online interaction is growing, and distinct methods of analysis will most likely prove useful.

Primarily, this study analyzed language in a virtual space. I argued that in the virtual realm, we can interpret spatial deixis only through the deictic center, which carries itself through empathetic deixis. While empathetic deixis is an important marker in transferring the deictic center, in a discussion about national identity it becomes a tool to impose the morality attached to the referred space. In addition, the use of plural personal pronouns in reference to self and the Other can be yet another level of imagination of an already imagined community--a virtualized imagined community. This seeming cohesion between people who have presumably never met in person and most likely never will can point to the success of the internalization of the dominant discourse.

These aspects of language use in the virtual realm have important ramifications for the reproduction of power. Through the analysis, the dominant discourse was revealed in the grammatical structure as well as in the lexical construction of the NOD and the LOD tweets. While the users of the LOD try to resist the NOD, they sustain the power in and behind the dominant discourse that underlies both the NOD and the LOD. Here, the LOD as well as the NOD join together through their shared linguistic tools and connect themselves to the seemingly all-encompassing dominant discourse. Although each discursive pattern is supposedly oppositional, their sharing unites them within a larger discursive whole.

This discursive unison has significant political implications as it pertains to diminishing the prospects and possibilities of democracy within a system where the dominant discourse seems to have no counterdiscourse. Therefore, what is at stake is not solely linguistic mechanisms of online interaction. Understanding these dynamics also has practical implications. At first glance, the LOD might seem to be the counterstance to the NOD, but a second look reveals that the LOD is another reflection of the dominant discourse. Practically, this suggests that any attempt to oppose the NOD in Turkey needs to take the linguistic mechanisms of dominant discourse into consideration. As Fairclough (1993, p. 4) suggests, discourse and social practice are inseparable; therefore, no one can expect a different social result to be achieved through the same linguistic and discursive practices. Without establishing alternative linguistic and discursive dynamics, it will be impossible to forge a resistance--"talking back" will always be "talking in."


Abel, F., Gao, Q., Houben, G., & Tao, K. (2011). Analyzing user modeling on Twitter for personalized news recommendations. User Modeling, Adaption and Personalization, 1-12.

Ahearn, L. M. (2001). Language and agency. In W. H. Durham, E. V. Daniel, & B. B. Schieffelin (Eds.), Annual review of anthropology 30, 190-137.

Amnesty International. (2007, January 19). Turkey: Murder of a journalist deplored. Retrieved from

Anderson, B. (2006). Imagined communities (3rd ed.). London, UK: Verso.

Boyd, D., Golder, S., & Lotan, G. (2010). Tweet, tweet, retweet: Conversational aspects of retweeting on Twitter. In R. H. Sprague (Ed.). Proceedings of the 43rd Hawaii International Conference on System Sciences. Koloa, Kauai, HI (pp. 1-10). doi:10.1109/HICSS.2010.412

Brown, G. (1995). Speakers, listeners and communication: Explorations in discourse analysis. Cambridge, UK: Cambridge University Press.

Bucholtz, M., & Hall, K. (2004). Theorizing identity in language and sexuality research. Language in Society, 33(4), 469-515.

Caha, O. (2001). The inevitable coexistence of civil society and liberalism: The case of Turkey. Journal of Economic and Social Research, 3(2), 35-50.

Choudhary, S. K. (1988). Review: The "elitism" of nationalist discourse. Social Scientist 16(3), 41-46.

Ergundogan, Y. (2012, February 29). Talim ve terbiyeden gecenlerin sancili yuzlesmesi. Kureyel [The painful face-off of those who pass through training]. Retrieved from

Erlank, N. (2003). Gender and masculinity in South African nationalist discourse. Feminist Studies, 29(3), 653-671.

Fairclough, N. (1993). Discourse and social change. Cambridge, UK: Polity.

Fairclough, N. (2001). Language and power (2nd ed.). Essex, UK: Pearson ESL.

Fairclough, N. (2011). Critical discourse analysis. In J. P. Gee & M. Handford (Eds.), The Routledge handbook of discourse analysis (1st ed.) (pp. 9-20). New York, NY: Routledge.

Gumperz, J. J. (2003). Interactional sociolinguistics. In D. Schiffrin, D. Tannen, & H. E. Hamilton (Eds.), The handbook of discourse analysis (pp. 215-228). Malden, MA: Blackwell.

Hill, J. H. (1995). The voices of Don Gabriel: Responsibility and self in a modern Mexicano narrative. In D. Tedlock & B. Mannheim (Eds.), Dialogic emergence of culture (pp. 97-147). Champaign, IL: University of Illinois Press.

Honey, C., & Herring, S. C. (2009). Beyond microblogging: Conversation and collaboration via Twitter. In Proceedings of the 42nd Hawaii International Conference on System Sciences. Waikoloa, Big Island, HI (pp. 1-10). doi:10.1109/HICSS.2009.89

Huang, J., Thornton, K. M., & Efthimiadis, E. N. (2010). Conversational tagging in twitter. In Proceedings of the 21st ACM Conference on Hypertext and Hypermedia (pp. 173-178). New York, NY: ACM. doi:10.1145/1810617.1810647

Human Rights Watch. (2007). Turkey: Outspoken Turkish-Armenian journalist murdered. Retrieved from

Jager, S., & Maier, F. (2009). Theoretical and methodological aspects of Foucauldian critical discourse analysis and dispositive analysis. In R. Wodak & M. Meyer (Eds.), Methods of critical discourse analysis (pp. 35-62). London, UK: SAGE Publications.

Kadioglu, A. (1996). The paradox of Turkish nationalism and the construction of official identity. Middle Eastern Studies, 32(2), 177-193.

Kouloumpis, E., Wilson, T., & Moore, J. (2011). Twitter sentiment analysis: The good the bad and the OMG! In Proceedings of the 5th International AAAI Conference on Weblogs and Social Media (ICWSM 2011) (pp. 538-541). Barcelona, Spain: ICWSM Retrieved from

Lakoff, G., & Johnson, M. (2003). Metaphors we live by (2nd ed.). Chicago, IL: University of Chicago Press.

Leap, W. L. (2010). Homophobia as moral geography. Gender and Language, 4(2), 187-200.

Levinson, S. C. (1983). Pragmatics. Cambridge, UK: Cambridge University Press.

Modan, G. G. (2007). Turf wars: Discourse, diversity, and the politics of place. Malden, MA: Blackwell.

Moe, H. (2012). Who participates and how? Twitter as an arena for public debate about the data retention directive in Norway. International Journal of Communication, 6, 1222-1244. Retrieved from

Ozcer, A. (2012, February 28). Resmi ideoloji ile yuzlegmek [To face the official ideology]. Resmi ideoloji sempozyumu [Official ideology symposium]. Retrieved from

Parekh, B. (1995). Ethnocentricity of the nationalist discourse. Nations and Nationalism, 1 (1), 25-52.

Paul Gee, J. (2011). How to do discourse analysis. Oxon, UK: Routledge.

Reisigl, M., & Wodak, R. (2009). The discourse-historical approach (DHA). In R. Wodak & M. Meyer (Eds.), Methods of critical discourse analysis (pp. 87-121). London, UK: SAGE Publications.

Schiffrin, D. (1990). Between text and context: Deixis, anaphora, and the meaning of then. Text & Talk, 10(3), 245-270.

Semino, E. (2008,). Metaphor in discourse (1st ed.). Cambridge, UK: Cambridge University Press.

Smith, T. W. (2005). Civic nationalism and ethnocultural justice in Turkey. Human Rights Quarterly, 27(2), 436-470.

Solgun, C. (2012). "Resmi ideoloji" ile yuzlegmek sorumlulugu [The responsibility of facing official ideology]. Resmi Ideoloji Sempozyumu [Official ideology symposium]. Retrieved from

Sorenson, J. (1996). Learning to be Oromo: Nationalist discourse in the diaspora. Social Identities: Journal for the Study of Race, Nation and Culture, 2(3), 439-468.

Tavernise, S. (2007, July 3). Trial in editor's killing opens, testing rule of law in Turkey. The New York Times. Retrieved from

Tayman, E. (2012). Resmi ideolojiyle yuzlegme igin sempozyum [Symposium for facing official ideology]. Radikal. Retrieved from = 1079858&CategoryID =77

Turgut, P. (2012, January 18). The murder of Hrant Dink: A Turkish court denies a wider conspiracy. Time. Retrieved from,8599,2104666,00.html

Urban, G. (2001). Metaculture: How culture moves through the world. Minneapolis, MN: University of Minnesota Press.

Van Dijk, T. (2003). Critical discourse analysis. In D. Schiffrin, D. Tannen, & H. E. Hamilton (Eds.), The handbook of discourse analysis (pp. 352-371). Malden, MA: Blackwell.

Van Dijk, T. (2008). Discourse and power. New York, NY: Palgrave Macmillan.

Van Leeuwen, T. (2009). Discourse as the recontextualization of social practice: A guide. In R. Wodak & M. Meyer (Eds.), Methods of critical discourse analysis (pp. 144-161). London, UK: SAGE Publications.

Wodak, R., & Reisigl, M. (2003). Discourse and racism. In D. Schiffrin, D. Tannen, & H. E. Hamilton (Eds.), The handbook of discourse analysis (pp. 372-397). Malden, MA: Blackwell.


Howard University

Ali Ersen Erol:

Date submitted: 2012-07-28

(1) I would like to thank William Leap, Bill Starosta, Teun A. Van Dijk, Joris Gjata, and Efe Sevin, and the reviewers for the time they dedicated to share their thoughts and comments on previous versions of this article.

(2) I analyzed the tweets in Turkish, then translated them to English. The translations were checked by other native Turkish speakers to ensure accuracy. The names of the users are changed to ensure anonymity--except for Ahmet Hakan, ahmethc, who is central to this discussion. The numbers next to the tweets refer to their place within the chronological order of the 65 tweets.
COPYRIGHT 2013 University of Southern California, Annenberg School for Communication & Journalism, Annenberg Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Erol, Ali Ersen
Publication:International journal of communication (Online)
Article Type:Report
Geographic Code:7TURK
Date:Feb 15, 2013
Previous Article:Relevant but long since absent: re-establishing a political economy of the Dutch media.
Next Article:Programming and editing as alternative logics of music radio production.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters