Truth after post-truth: for a Strong Programme in Discourse Studies

Contemporary post-truth discourses put the constructivist foundations of Discourse Studies to a test. According to critical observers, discourse analysts have been playing into the hands of Trump, Brexit and right-wing populists by politicising scientific knowledge and undermining the idea of scientific truth. In order to respond to these concerns, this article outlines a Strong Programme in Discourse Studies. While the Strong Programme insists on truths as discursive constructions, in no way does it claim that all ideas have the same truth value or that an idea can become true because somebody wants it to be true. The Strong Programme makes the case for discourse research that is constructivist (it asks how truths are constructed practically) without being relativist (all ideas do not have the same normative quality). Taking inspiration from debates in Science and Technology Studies of the 1970s, the Strong Programme formulates principles for discourse researchers dealing with conflicting truth claims. Discourse analytical explanations of truths of first-order participants and of second-order observers should be symmetrical, heterogeneous, multi-perspectival and reflexive. The Strong Programme discourse research is grounded in the founding traditions of “French” and “Critical” Discourse Studies, which have struggled over questions of truth and reality since the beginning. While critically interrogating the structuralist heritage of these strands, the Strong Programme insists on the practices of making and unmaking ideas through language use no matter whether they appear as true or false to participants and observers. Discourse Studies are encouraged to critically reflect on how hierarchies between knowledges are not only represented but, through their representation, also constituted through discursive practices.

Introduction: truth in the age of post-truth G enerations of discourse researchers have engaged in critical reflections on truth as a weapon of the powerful. From Nietzsche's idea of truth as "a mobile army of metaphors" (Nietzsche, 1997, p 880) to Foucault's investigations of the nexus of truth, subjectivity and discourse (Foucault, 2017), many have pledged to deconstruct 'Western' claims to objective and universal truth. Yet if truth is nothing more than a product of discursive struggles, would one not have to accept all ideas as equally true? And how do discourse researchers defend the truth claims of their own research vis-à-vis the discourses they criticise?
According to critics, discourse theorists have gone too far in questioning reality since the advent of mass media society (e.g., Flyverbom and Reinecke, 2017). Thus, observers from within Discourse Studies, as well as from outside have denounced French discourse theories as 'postmodernist' (Habermas, 1993;Eagleton, 1996), even as a threat to Western democracy (Ferry and Renaut, 1988). Thus, for these critics, 'postmodernism' supports the idea that anything goes in moral affairs, that truth is nothing but an expression of power relationships and that an idea is true because people want it to be true. And they blame postmodernists for discrediting the idea of scientific truth.
One recalls the scathing attacks against French discourse theory and Science and Technology Studies during the Sokal affair, which erupted in 1996, when a nonsense article on quantum physics was accepted for publication in Social Text, a Cultural Studies journal from North America, known for its 'postmodernist' inclinations. Alan Sokal, the physics professor who authored the article, blamed discourse theorists for a perceived lack of intellectual rigour and warned against politicising the knowledge produced by the natural sciences (Sokal and Bricmont, 1998).
Now that populist leaders such as Trump reject scientifically established facts like climate change, STS scholars and discourse researchers once again need to defend constructivist approaches to scientific knowledge. How do they respond to those unlikely 'friends' in the political arena who, like Brexit champion Michael Gove, claim to "have had enough of experts" (3rd of June 2016 on Sky News) and who, like Kellyanne Conway (22nd of January 2017 on NBC), justified Donald Trump's claims about his inauguration as "alternative facts"? How can discourse researchers go on with critically interrogating truth and reality if their research may serve propagandists of post-truth and their ideological agenda? Some liberal commentators from the media sphere have traced populist aberrations back to French discourse theory, for instance Guardian columnist Matthew d'Ancona (2017, chapter 4), Casey Williams in the New York Times (2017) or social media activist Helen Pluckrose (2017). But why are theorists such as Foucault and Derrida, who worked in France in the 1960s and 1970s and never made specific claims about contemporary politics in the U. S. or the UK, held responsible for political problems today? The attacks against discourse theory are even more surprising given that the young urban intellectuals who like to refer to discourse theory today generally have no sympathy whatsoever with the ideological agenda of right-wing populists.
Yet whatever the seriously flawed arguments that some journalists throw against academics and intellectuals, I find it legitimate to ask whether right-wing populists, who tend to reject science and scientific reason that does not serve their agenda, are emboldened by constructivist critiques of truth that one can find among poststructuralists and postmodernists, in STS scholars and discourse theorists. While constructivism has been the object of a long critical debate (Hacking, 1999), leaving the question of truth to philosophers and political theorists would be a mistake. Discourse researchers should be attentive to the questions of philosophers such as Arendt (2006), Habermas (1990) or Rancière (1995). One may not share their predisposition for abstract, acontextual and universalist thinking. Yet discourse researchers should consider that all truth claims have the same value. Nor are ideas true if they are accepted by a majority.
Discourse researchers can subscribe to the idea that there are discursive struggles over truth but not all truth claims have the same normative quality. There are claims about realities, which may become true through the force of large social groups (think of the dynamics of mass media discourse) in the exchanges between few select specialists (e.g., small academic and professional communities). Some truth claims are about social realities, others about non-social realities. And often there is a conflict between different types of truth claims emerging in games mobilising different resources and following different rules.
While discourse is a practice mobilising linguistic, as well as non-linguistic resources, truth claims are made in and about the material world. Discourse theorists have been interested in how social and cultural phenomena are turned into 'matter' through discursive practices (Butler, 1993). Discourse theorists can agree with theorists of the New Materialism that there are no one-way causal relationships between discursive practices and nondiscursive matter (Frost, 2011). Humans do not have a monopoly of knowledge and agency over a material world seen as a passive surface waiting to be shaped by human inscriptions (Barad, 2003). Discourse researchers should have no problem with recognizing the social and non-social constraints on the representations people make of reality. A lake is not frozen because people say it is frozen and people feel hungry no matter what is said about their bodies. Discourse communities may indeed accept the idea as true and real that the lake is frozen and that you are not hungry even though the lake is not frozen and you are hungry.
While discourse researchers should accept that there is a world which is beyond discursive reach (physics, biology etc. are not merely social constructions as STS scholars have always reminded us, Vrieze, 2017), it would be problematical to understand whatever happens between language users in terms of physical, biological or other such non-discursive laws. At the same time, the social world is a heterogeneous space of articulated elements, things, practices, bodies, which should not be reduced to one allencompassing power game which explains it all. Therefore, what one can take from both political philosophers and the theorists of the New Materialism is that claims gain specific truth values in a heterogeneity of practical expertise that constitute the specific normative quality of a truth claim.
To counter the politics of post-truth, discourse researchers, therefore, do not have to return to Truth-i.e., to the assumption that some ideas are inherently better than others, that only trained specialists can have access to them, that some ideas are true before and outside discourse as it were. Discourse researchers can distinguish between truth claims with higher and lower normative quality without betraying their fundamental constructivist orientations. Hence, in the following, I will outline a Strong Programme that makes the case for discourse research which is constructivist without being relativist. It formulates principles that allow discourse researchers to deal with truth claims of first-order participants and second-order observers. While the Strong Programme pleads for symmetrical explanations of true and false knowledges, it recognizes that not all knowledges are equal. Some knowledges have more truth value than others. Yet all truths are entangled in social dynamics and political struggles as a result of which not everything is accepted as equally true and valuable knowledge. I will start with a discussion of two major traditions in Discourse Studies, namely "French school" discourse analysis and Critical Discourse Analysis (CDA). Both these traditions are struggling with the heritage of structuralism, which sometimes leads them to adopt asymmetric (i.e., 'weak') explanations of true and false discourses. I will then look into debates in Science and Technology Studies over the social nature of scientific knowledge. In this debate, political and epistemological questions over the authority of the observers were raised similar to those discourse researchers and social researchers are struggling with today. I will conclude with the observation that a 'strong' epistemology (constructivism) does not imply a 'weak' politics-i.e., science pretending to be neutral or unable to take political positions. Whenever language users enter discourse, they participate in struggles over truth, which cannot but be political.
Classical discourse studies: from the crisis of Truth to the many truths Discourse Studies is a recent field, which has resulted from the encounter of two lines of debate: discourse theory and discourse analysis. Discourse theory deals with questions in social, political and cultural theory around the role of language and communication in contemporary society (Laclau and Mouffe, 1985;Foucault, 1969;Butler, 1997). Also labelled as 'linguistic turn' in post-war philosophy (from Heidegger to Foucault and Habermas) or French Theory in the cultural and literary field, discourse theory concentrates on epistemological questions such as the limits of the speaking subject (Derrida, 1967), the dilemmas of representation in postmodernity (Jameson, 1991) or the discursive negotiation of normative claims (Habermas, 1981). While discourse theory shows a proclivity for philosophical and epistemological problems, discourse analysis puts emphasis on analytical methods that discourse researchers use to investigate social practices in view of producing insights into empirical objects (Angermuller et al., 2014). If discourse theory points to the intellectual challenges in Discourse Studies, discourse analysis reminds us of the crucial role of analytical models and empirical methods in Discourse Studies.
Whenever the theorists met the analysts turned out to be particularly productive (Angermuller, 2015). And two such moments have given birth to brands in Discourse Studies which are recognised internationally today: "French" Discourse Studies, which goes back to debates around Marxism, structuralism and psychoanalysis in France in the late 1960s, and "Critical" Discourse Studies, which has become prominent since the late 1980s in the Anglophone world.
The two labels may be somewhat misleading in that "French" discourse researchers usually see themselves as "critical" claiming a background in Marxist social theory. And some "critical" discourse researchers cite "French" discourse theorists (who are not necessarily French or even French-speaking natives if one includes the many commentators of Foucault et al. outside France). It needs emphasising that Critical Discourse Studies is not more "critical" than French Discourse Studies, both being rooted in Marxism and involving academics with a strong engagement in civil society (Dufour, 2013). What is more, even though the "French" school of discourse analysis, which one can trace back to a group of linguists around Michel Pêcheux, is a largely French-speaking phenomenon, it comprises scholars in many other countries, especially from the "Latin" world (Latin America, Spain, Portugal, Italy, Romania…). And the geographical base of "Critical" Discourse Analysis may not be too obvious, either. While centred in the UK, it is especially popular among a large community of English-speaking discourse researchers.
"French" Discourse Studies goes back to a first conjuncture of discourse theory and discourse analysis, of discourse theorists such as Michel Pêcheux and Michel Foucault (Foucault, 1969;Pêcheux, 1969) and the discourse analytical tools from linguistics, including corpus analysis (lexicométrie), distributionalism and pragmatics, most of which are less known outside France than French discourse theories, which have been translated into many languages. While Foucault made the label of discourse known across the disciplines and internationally, it was Michel Pêcheux, who established 'discourse' in linguistics by founding what is sometimes called the 'French' school of discourse analysis. Pêcheux and his collaborators emphasise the need for a linguistic analysis that accounts for the 'materiality' of language (Conein et al., 1981). In this view, language is not like a window to the external world or to an internal consciousness. It is perceived as a surface of opaque signs whose constraints on interpretive activities need to be decrypted.
The stance of 'French' discourse analysts is 'materialist' in at least two ways (cf. Beetz, 2017). For one thing, it rejects the spontaneous interpretive practices of hermeneutics and relies on rigorous formal analysis of material linguistic forms. For another, it articulates linguistic analysis of discourse with 'materialist' social theory. Thus, Pêcheux takes inspiration from Althusserian Marxism as discursive formations are embedded in their material 'conditions of production', i.e., class struggle. Pêcheux's materialist programme was basically shared by Foucault even though in a less radical way. Foucault's discourse analytical work is not Marxist even though he showed a keen, critical sense of the power relationships in which language use is inevitably bound up with. In his Archaeology of Knowledge, he resolutely goes beyond the abstract formalism of linguistic structuralism while referring linguistic practice to its sociohistorical context (Angermuller, 2014, 7ff.).
In France, such research has succeeded in establishing 'discourse' both as a transdisciplinary problem ('discourse' as a recognised theoretical problem across the social sciences and humanities -'French discourse theory') and as an object of a subdisciplinary field (i.e., discourse analysis as a specialized research practice within linguistics -'French discourse analysis'). Outside France, French discourse theory, often inspired by translations of Foucault's critical work, has become popular in the literary and cultural field and among many Anglophone and German-language social and political theorists. Within France, discourse analysis has seen a decisive move toward pragmatic questions while focusing on societal institutions (Maingueneau, 2014).
The other major international brand is Critical Discourse Studies, which began in the UK and in countries of the Commonwealth (van Leeuwen, 2008), in the German-speaking world (Wodak et al., 1998) and through Teun van Dijk's work also in the Latin world (van Dijk, 1985). Just as French Discourse Studies, Critical Discourse Studies has its base in linguistics and it has developed many links to other fields such as sociology, history, political science, education, psychology, anthropology, philosophy (cf. Unger, 2016). Critical Discourse Studies is an umbrella label for a broad range of theories and methods at the intersection of language and society. While French Discourse Studies is inspired by Continental theoretical strands like structuralism, Marxism and psychoanalysis, Critical Discourse Studies is more eclectic in its theoretical orientations and is sometimes influenced by Halliday's systemic-functional linguistics. "Critical" discourse researchers are often driven by humanist concerns over social injustices with the explicit desire to respond to social problems and to bring about positive social change. Therefore, almost any discourse analytical methods and tools can be and are used within Critical Discourse Studies, including quantitative corpus analysis and the whole array of qualitative methods of social research, e.g., document analysis, interactional approaches or ethnography.
Whereas Pêcheux and Foucault are two major theoretical reference points in French Discourse Analysis, one may see the systemic-functional linguistics of MKA Halliday (1978) and Basil Bernstein's sociology of education (1971) perhaps as majors inspirations of CDA theorists. Halliday's theoretical project is centred on the semiotic resources that allows language users to realise the functions of language (such as the ideational, interpersonal and textual ones). Emphasis is put on the social contexts in which the meaning potentials of semiotic resources are realised. Halliday worked in close collaboration with Bernstein, who studied the role of language among pupils from lower and upper classes in Great Britain. Even though Bernstein's distinction between the restricted code of working-class pupils and the elaborate code of upper-class pupils has always caused controversial reactions, his influence on the way many British linguists conceptualised the nexus of language and society should not be underestimated. 1 Since the 1980s, the label CDA has come to designate language-related research on social problems, more specifically to research on how inequalities between large social groups (including relations of race, class and gender) shape and are shaped by the use of language in larger communities. Fairclough's social context model (1992, p 73) is an instructive example of how the functionalist (Hallidayan) approach to language is articulated with a macrosociological (Bernsteinian) perspective on the social. It puts the text (i.e., oral and written manifestations of language) centre stage. The text is surrounded by the context in two circles as it were, a first circle comprising processes of meaning negotiation between the discourse participants ("interaction") and a second circle designating society's institutional structures ("context"). Fairclough, therefore, testifies to a structuralist understanding of the social context in which language is used. Language use is embedded within the constituted structures of society. As a consequence, the social is posited as the hard ground of reality to which linguistic activity ("text" + "interaction") relates. A similar tendency of taking the social as a given rather than as the empirical problem can be observed among other representatives of Critical Discourse Studies. Van Dijk's sociocognitive approach (2008) asks how cognition mediates between language and society. Society is perceived as the mute outside of linguistically organised cognition. Wodak and Reisigl, Wodak (2009), too, define Critical Discourse Studies as linguistic research that focuses on such given social problems. Accordingly, rather than examining how discourse participants negotiate what counts as a social problem, Wodak and Reisigl start from a social problem and assess how language use relates to the problem thus stated.
Discourse researchers from both the French and Critical traditions have been crucially interested in how truths are produced and established through language use in discourse communities (e.g., Guilbert, 2008;Reisigl, 2008). 2 Yet one can observe that they tend to apply different accounts to the two types of social realities they deal with, namely to the social reality of the first-order participants (SR 1) and to the social reality of the second-order observers and discourse researchers (SR2). Why is such an asymmetry problematical?
It is a common question in Discourse Studies to ask how discursive practices, i.e., the social uses made of language, not only represent SR 1 but, through its representation, also constitute SR 1. A classic example is how social problems (SR 1) are constructed in public discourse (cf. Miró, 2017;Pechtelidis and Stamou, 2017). Yet, to account for the construction of SR 1, discourse researchers typically investigate how language is used in its social and historical contexts, which one may designate as SR 2: the communities, the institutions, the economic resources which are mobilised in discursive practices. The question is how to deal with contradictions between SR 1 and SR 2, which critical research often provoke. SR 1 may be false from the researcher's point of view, i.e., a mere opinion that the participants of a discourse have accepted as true. SR 2, by contrast, is the idea the researcher has about the social context, which she or he holds as true. While SR 1 may be a result of systematic discourse analytical investigation, SR 2 often refers to the theoretical framework or background knowledge the researcher assumes to be true. In other words, how does one deal with the problem that the social reality that Trump accepts (SR 1) will normally not accept the reality that discourse researchers presuppose in order to account for Trump (SR 2)?
Discourse researchers usually make truth claims about SR 1 and SR 2, which is precisely what they are supposed to do. What needs to be problematized, however, is a tendency to account for SR 1 and SR 1 differently: namely SR 1 as a discursively constructed reality (which therefore may be 'false') and SR 2 as a reality which is not constructed in discourse (which therefore cannot be false). Such an asymmetric account is weak epistemologically speaking for why could not SR 2 be challenged with the same arguments that are used to deconstruct SR 1? And it may not constitute a strong political position either if it invites post-truth propagandists to simply turn the table and place science, the media and the 'establishment' in the position of SR 1. SR 1 is then revealed to be a discursive construction from the point of view of SR 2, i.e., the world of America First and Brexit Britain.
'French' and 'Critical' discourse research, indebted to the heritage of structuralism, often struggles with such inbuilt asymmetries between the truth claims of first-order discourse participants and the truth claims of second-order observers. Such asymmetries have been a problem that 'poststructuralist' developments have tried to solve by rejecting any primacy of SR 2 over SR 1. It has indeed become difficult for post-Weberian social scientists to claim a god's eye view in face of the many competing truths in society.
While the social sciences have seen a turn towards the actor during the 20th century, the actor is mostly seen with suspicion in both Critical and French Discourse Studies, and with good reasons! Throughout the 1970s, Pêcheux and his collaborators held out against attempts to take pragmatic lessons and bring the actor back in, which they perceived as preparing the ground for a return to an insipid humanism (at best) and even to bourgeois liberalism (at worst). A similar tendency can be observed in the Anglophone and German-speaking world where critical discourse analysts were long pitted against conversation analysts as can be seen in the controversy between Billig (1999) and Schegloff (1997). For Schegloff, Critical Discourse Analysis relies on external theories of context, i.e., on ideas and theories the critical discourse analyst has about social inequality rather than on the knowledge the discourse participants flag out as relevant. Conversation analysts in turn have been suspected for not taking power and inequality into account and for seeing society as a mere illusion of left-leaning ideologues.
While these clashes, it seems, have pushed both discourse and conversation analysts back into their respective specialised niches, with a structuralist account of Truth for discourse researchers and a pragmatist account of a plurality of truths for conversation analysts, such an opposition is neither necessary nor productive (cf. Taha, 2017). Yet upon closer inspection, Schegloff's argument was meant to be a methodological, not an ontological one. Language use refers to specific practices of making some context relevant, i.e., to contextualisation as an ongoing activity of participants in discourse where nobody can claim to grasp 'the' context as such and thus take a position of Truth (cf. Porsché, ARTICLE PALGRAVE COMMUNICATIONS | DOI: 10.1057/s41599-018-0080-1 2018). What Schegloff challenges is the idea that discourse analysts can see society from a privileged viewpoint which is more objective than the ones other discourse participants take. The conflict between discourse and conversation analysts, therefore, prolongs struggles over who has the true expertise and authority to make true and legitimate claims about the social: the conversation analyst typically privileges the expertise of the actors (and perceives their truths on the same par as his or her truths) whereas the discourse analyst takes aims to reveal what the participants cannot see (which places her or him in a position of Truth). The danger for discourse researchers is to claim an absolutist epistemological position concerning the social.
To deal with this problem, French and critical traditions have been switching uneasily between two arguments which are difficult to reconcile. On the one hand, they like to see discourse as being constitutive of the social: discursive practices do not only represent the social. Rather, through representation, such practices bring forth the relationships and structures that make up the social. On the other hand, discourse researchers usually place discursive practices within constituted configurations of power and inequality, which are not 'just' discursive constructions. As a way out, I will invite discourse researchers to consider the Strong Programme, which conceives discourse as a situated practice of making and unmaking truths through the uses members make of language in a discourse community. The Strong Programme rejects philosophical accounts of Truth in favour of reflexive investigations of struggles over truths.

Towards the Strong Programme in Science and Technology Studies
At this point we will need to look into the lessons that discourse researchers can draw from Science and Technology Studies (STS). 3 The Strong Programme was formulated in STS in the 1970s, when the ideal of science as pure knowledge production came under attack. While earlier sociologists of science still clung to the idea that there was Truth out there, untouched by society as it were, which scientists could reveal under certain circumstances, a new generation of more radically constructivist scholars in STS felt that all scientific knowledge needed to be seen as a product of social, political and economic dynamics. Commonly associated with a group of philosophers and sociologists based at Edinburgh under the leadership of David Bloor, the Strong Programme emerged from the critical interrogations over the social nature of scientific truth.
Bloor's objective was to formulate basic epistemological principles which allow scientists to reflect on the truth claims made by others in relationship to their own truth claims. Bloor (1991, p 3) argued that contemporary sociologists of knowledge and science betrayed their discipline if they followed an implicit hierarchy-or 'asymmetry'-between 'true' scientific knowledge, which is true on its own account and needs no sociological explanation, and 'false' knowledge, which is in need of a sociological explanation. For Bloor, 'true', as well as 'false' knowledges result from the interplay of social, as well as an array of other ('material') factors. Hence, the Strong Programme explicitly acknowledges that truth and reality may be made not only from social practices. In a similar vein, the Strong Programme will invite discourse researchers today to apply discourse analytical insights symmetrically to both 'true' and 'false' assumptions about social reality. Strong programmers investigate the discursive practices, processes and mechanisms of constructing social order. But they do not claim that what is real and true in such constructions is a function of discursive practices only. Reality and truth are built from social stuff which can be discursive (involving language) or non-discursive (non-linguistic actions) and also from non-social stuff (which can be human bodies or non-human matter). The reality of the social conceals no inherent rationality that would entail a given, 'objective' way of describing it through discourse. Nor does it come with universal values which would demand a 'just' judgement. Strong programmers take into account whatever can help explain the making of truth and reality. They are materialists after all.
In his Knowledge and Social Imagery (1991, 5ff.), Bloor lays out the four principles that sum up the Strong Programme in the sociology of scientific knowledge: 1. Causality. Rather than supporting the causalistic logic of the natural sciences, Bloor points to the specific conditions and causes, which can include non-social ones "which will cooperate in bringing about belief" (1991, p 7). 2. Impartiality. Bloor rejects the idea of there being a priori hierarchies between knowledges. While Bloor abstains from privileging truth over falsity, rationality over irrationality, success over failure, one may take such hierarchies as the object to be accounted for. 3. Symmetry. Explanations would need to identify the same types of cause to "explain say, true and false beliefs". Truth, in other words, would not be something to be explained philosophically and untruth to be explained sociologically. 4. Reflexivity. The sociology of scientific knowledge, too, is involved in the production of truths and untruths. Therefore, explanations would also need to be applicable to the knowledge claims of the Strong Programme: "It is an obvious requirement of principle because otherwise sociology would be a standing refutation of its own theories." (1991, p 7), The Strong Programme has been widely received as the theoretical symptom of a practice turn in the sociological and historical research on science in the 1970s and 1980s. The implicit adversary was rationalistic and positivistic ideas of science as pure knowledge, as well as the classical ('institutionalist') sociology of science, going back to Merton (1968). Bloor and his associates perceived Merton to be tacitly indebted to a rationalist agenda. While Merton insisted on the sociocultural foundations (the 'scientific ethos') of the science system, his work aims at identifying the social conditions that are needed for true knowledge to emerge. In terms of the conditions that 'real' science needs, Merton sees Western liberal democracies, especially the USA during the 20th century, as superior to Nazi Germany or the Soviet Union under Stalin. If it is perhaps not a question for debate that Northern American scientists found vastly better conditions (which not only include economic resources) than many of their counterparts in Europe, the top-down, God's eye view on academia and the broader social system has turned out to be problematic for the more empirically minded researchers of science as a social practice. Once real scientific practices can no longer be subsumed under such large umbrella concepts, the epistemological authority of the sociological observer starts to be challenged by other experts and scientific practices need closer empirical scrutiny.
In this context, the Strong Programme articulated a set of theoretical ideas that crucially contributed to the emergence of Science and Technology Studies (STS). The Strong Programme has been associated especially with the qualitative, constructivist, microsociological strands, such as the Laboratory Studies (Latour and Woolgar, 1979;Knorr Cetina, 1981), the sociology of scientific knowledge (Hicks and Potter, 1991) and academic discourse analysis (Myers, 1985). 4 What are the general tendencies in the research that has been developing in the wake of the Strong Programme? Firstly, Strong Programme research typically focuses on material practices, which cannot be explained by intentional actors or determining structures (cf. van Eeden, 2017). Secondly, it prefers empirical case studies of complex arrangements of specifically interlaced practices and processes. Thirdly, it considers truth and reality as the non-necessary result of dynamics among heterogeneous elements.
Discourse Studies and Science and Technology Studies do not necessarily deal with the same objects and challenges: one does not 'use' language in the same way as one 'uses' a technology. While Bloor made the radical claim that even mathematics should be accounted for symmetrically (i.e., socially), most discourse research deals whose profoundly social nature nobody would contest. And presenting a research programme for the field of STS, which did not exist at the time, is different from taking lessons for Discourse Studies, which has been developing for decades.
Therefore, I have adapted the Strong Programme from STS to the special circumstances of Discourse Studies today. From STS, it takes the fundamental idea that true and wrong ideas are constituted in ways that are not fundamentally different, i.e., in discursive practices. While 'weak' strands have recourse to discourse analytical explanations which can only account for 'false' representations of social reality, the strong programme investigates 'true' and 'false' ideas through the same set of methodological and theoretical tools of social research. Its first and major principle therefore is symmetry of explanation, applying the same logic of explanation to true and false representations of the social. The other three principles are: 1. Heterogeneity of factors. The Strong Programme recognizes the heterogeneous factors -discursive, non-discursive but also non-discursive non-social ones -that are mobilised in the construction of reality (cf. for a linguistic perspective on heterogeneity, see Maingueneau, 2017). It does not subscribe to the idea that the social is nothing but a product of linguistic practices. Nor does it believe that truth reflects one underlying power structure. It firmly places discursive practices in the material lives of discourse participants, who mobilise social, linguistic and physical resources to produce meanings, structures and truths. 2. Multi-perspectivality. It abstains from a God's eye view on the social, which privileges one reality over all others. Rather than deciding on what is the one true perspective on the social, it takes stock of the various, competing truths and realities and ask how they are constructed. While there is no epistemological primacy of one perspective over the other, the Strong Programme recognizes that not all ideas are equal and the question is how their truth value can be accounted for through the situated discursive and non-discursive practices that constitute them (Haraway, 1988). 3. Critical reflexivity. It invites discourse researchers to consider their activity as being part of the discursive objects they study. As a discourse about discourses in discourse, it acknowledges the critical effects scientific discourse may have on non-scientific practices (cf. Zienkowski, 2017). Scientific discourse cannot be neutral; it is always political because it is tied up in struggles over truth within academic communities, as well as in the broader social space (Celikates, 2018;Herzog, 2016). Discourse researchers, therefore, should reflect on their own role in social struggles over truth (cf. Nonhoff, 2017;Parker, 2015).

Conclusion: for a strong epistemology and a strong politics
The Strong Programme invites discourse researchers to deal with truths as a real-world problem. Whenever people use language, they enter struggles over truth. And in many cases one needs to deal with contradictory truth claims of first-order participants and of second-order discourse researchers. While discourse researchers cannot claim a privileged position that allows them to reveal a Truth that is hidden to others, there is no reason why they should not defend the ideas that they think are true, valuable and coherent against those ideas that are of lesser value to them. The Strong Programme invites discourse research to critically interrogate the struggles over truth in the light of the practice turn that has taken place in the social sciences (Schatzki et al., 2001;Boltanski, 1990). Rather than taking scientific truth claims as something categorically different from the truth claims of nonscientific language users, the practice view of discourse looks into the processes through which truths are constructed over time in a discourse community. By following the practice turn, the Strong Programme critically interrogates two major strands in the social sciences: a) the structuralist heritage, which one can find especially in classical strands of discourse research and which perceives the social as a reality which is before and outside language as it were, b) actor-centred approaches in sociology, which take meaning as a product of intentional and strategic language users.
The Strong Programme defends a third position which considers socially established truths as an effect of discursive practices. While it recognizes that not all ideas have the same truth value, it favours symmetrical accounts of the truth claims of discourse participants and those of discourse observers. The Strong Programme does not make discourse researcher to choose between universalism and culturalism, realism and relativism. Rather, it outlines principles for how to make truth claims in academic discourse with respect to the truths and realities constructed and established by other language users in their discourse communities. While Strong Programmers refrain from adjudicating on the truth, untruth or post-truth of any particular claim, they investigate the practices and processes that make some claims more true and valuable than others.
What are then the political implications and consequences for discourse research following the Strong Programme? Embracing a constructivist orientation does not lead to a normative anything goes and moral relativism. There is no reason why discourse researchers should think that the denial of climate change or the holocaust is just as acceptable as any other claim. The Strong Programme certainly does not make the case for 'fake news' or 'alternative facts'. On the contrary, it perceives the social world as a world of people and things, bodies and ideas which do not have the same value.
While neither populists nor intellectuals can claim access to Truth, the type of truths political propagandists and specialised experts produce are of a different order and quality. Promoted by billionaires from the oil industry, the idea that there is no climate change does not have the same normative quality as the ideas that a community of specialised climate researchers have produced through scientific procedures. Specialists produce truth claims through intimate knowledge of their objects in discourse communities which have developed trusted expertise. They may also have certain quality standards that are applied to distinguish between more and less valued knowledges. Highly valued ideas usually emerge from valuation practices and discursive dynamics over which individual language users usually have little intentional control. Why should specialists accept that the ideas valued in trusted expert communities are subordinate to the truths of non-specialists whose claims may be just personal whims? Not all knowledges have the same truth value.
Strong Programmers know that there is no Reality and Truth that can be accounted for by Reason. While they accept the plurality of truths and conflicting value hierarchies within and between discourse communities, they also know that truths are precious things that need time and labour, resources and practical know-how. Just like other language users, they are tied up in struggles over what counts as true and real in academic or nonacademic communities. These are struggles where discourse theorists can claim no special position unless they enter the very games they deal with and intervene in them practically. No matter whether truth claims are made by discourse participants or professional discourse analysts, the value of their ideas needs to be realised through the real effects their practices have on the social world. And this is an insight that one can perhaps take from the first critical theorist of discourse: "Man must prove the truth, i.e., the reality and power, the this-sidedness (Diesseitigkeit) of his thinking, in practice… Philosophers have hitherto only interpreted the world in various ways; the point is to change it." (Marx, 1969). Received: 29 August 2017Accepted: 21 February 2018 Notes 1 Whereas many North American researchers tend to conceptualise discourse as a symbolically mediated turn-taking process in a face-to-face situation, British discourse analysts show more awareness of wider societal constraints on language use such as class and capitalism. Unlike North American interactionists, who tend to focus on the negotiation of social order in face-to-face interactions, British discourse researchers are more interested in the articulation of discourse and class struggle. 2 And there are many other strands in Discourse Studies (Angermuller, 2015) that have dealt with the question of truth in discourse (one example is argumentation, van Eemeren, 2004;Amossy, 2005; another example are interactional approaches to discourse, e.g., Potter, 1996). 3 Another Strong Programme was formulated by Jeffrey Alexander and Philip Smith in cultural sociology (2010). By putting meaning centre stage, Alexander/Smith make the case for what could be called a discursive turn in social theory. While they do not specifically deal with the question of truth and reality, their framework gives a place to all socially and politically oriented discourse researchers no matter whether they are 'strong' and 'weak'. 4 An in-depth discussion would reveal that Bloor and Latour, who originated Actor-Network-Theory, always insisted on their differences. While Bloor castigates Latour for establishing an asymmetric vantage point above nature and society as it were (Bloor, 1999, p 85), Latour sees Bloor as being "too social" and as reproducing the "modernist" asymmetry between nature and society (Latour, 1999, p 127).