Partly driven by social media, the mediatisation of society has spawned new forms of political communication in the public sphere, accompanied by high hopes for more deliberative discourse along the Habermasian model of deliberative democracy (Habermas 1962/1990). Habermas sees the public sphere as a discursive space in which individuals and groups congregate to discuss matters of mutual interest and, where possible, reach a solution shared by all. This ideal constellation of public discourse is based on certain normative conditions: (a) formal and discursive equality; (b) reciprocity and rational critique as core qualities of the discourse; (c) ideal role taking, meaning that participants must remain hermeneutically open and sensitive to how others (the public) define themselves and the world. It is frequently argued, that the internet offers new options to realize this kind of deliberative sphere of discourse. Critics, however, see the internet not as a place for new democratic structures or more deliberation, but as a digital environment, which increases the fragmentation of society by supporting an elitist discourse. Taking up the Habermasian model, this paper explores on the basis of empirical research, how patterns of interaction and argumentation on twitter influence online discourses. The analysis is based on a theoretical model of Twitter as a functional discursive practice, which forms a "Discourse Universe", structured and determined by a semiotic sign system. This semiotic system consists of language signs, images, and main signifiers. These main signifiers are the following four operational signs, which are connected to discourse specific functions: (1) @ Addressing/Mention (at) (2)
Organizing/Tagging (hashtag) (3) RT Sharing/Distribution (re-tweet) (4) http:// Linking to other sources (link/short URL) These twitter-specific communicative functions of addressing, linking, tagging and republishing offer new opportunities for citizens to participate in the political discourse. The tweets have a contextualized communicative function within a larger framework of a discourse network. This discourse network is based on specific functions of twitter as a tool for interpersonal exchange. This network can be described by the following functions, based on the Twitter signifiers (@, , RT, http://) (Thimm et al. 2011): * Intertextuality: references between Tweets, coherence via @, structure via * Transmediality: topics of discourse are discussed and commented in different media via http:// * Intermediality: multimodal expression of discoursive content in photos etc. These communicative functions are realized by multimodal intertextual compositions of language signs, pictures, films and sound. Based on this functional model, the main research questions put forward can be focused on four levels of discourse: * Interactive Level: Is Twitter a tool for a more direct, more immediate exchange between citizens and politicians? If so, in which contexts? * Functional Level: Which communicative strategies and functions are used in these exchanges? * Performative level: Which linguistic means are used, which speech acts are enacted? * Societal and Political Level: Does Twitter enable deliberative democracy, according to the principles put forward by Habermas? To follow up these research questions, an extensive body of Twitter messages ("Tweets") was documented during five state elections in Germany. For the analysis, a total of 375.000 Tweets were systematically collected three weeks before and one week after the elections in the years 2010 and 2011.These tweets include: (1) tweets by selected candidates of each party, (2) tweets by print media twitter portals, as well as (3) by citizens posting about the election and connecting up with politicians. The overall methodological process applied can be characterized as triangulation, combining qualitative with quantitative measures. Qualitative measures included interpersonal interaction (via @replies and RTs), semantic analysis (structuring function via hashtags), argumentation and narration (via links, twitpics and videos) and speech acts (e.g. metacommunication). Quantitative measures consisted of content analysis, frequency profiles, and interactional topic profiles, based on an analysis with help of the QDA Miner Software. The quantitative and qualitative measures were combined to extract Twitter-profiles of the politicians, including results of their activity, interactive styles, argumentation patterns as well as interaction structures (e.g. types of reference, topic management). The analysis revealed different strategies on the level of interaction and responsivity on the part of the politicians. Some focused on campaign issues and went ahead with their personal agenda ("presentational type"), while others engaged directly and intently with citizens and used twitter as a dialogical tool ("interactive type").The content analysis showed an unusual mix of private and political topics on the part of the politicians, a mix, which is only partly accepted by the wider public.* So far, Twitter is often a one-way medium: people are following politicians, but don't necessarily interact.* Most recently, however, more citizens have started using the interactive functions to voice their interests and concerns. It will be argued that Twitter is a complex and highly condensed platform ("discourse universe") for information diffusion, interpersonal exchange, argumentation and mobilisation. In respect of the idealized concept of deliberative democracy, the data demonstrate new options for political debates and political participation. Overall, Twitter will be conceptualized as a "staying in touch" and "gatewatching" medium, which not only combines a control and information exchange function, but enables citizens to participate and influence politics more than ever before.