TOWARDS A REALIST SOCIAL CONSTRUCTIONISM

Social constructionism has often been seen as incompatible with realist approaches to the social world. This paper argues that critical realism is thoroughly compatible with moderate versions of social constructionism and indeed provides stronger ontological backing for it than the anti-realist approaches that are often associated with more extreme versions of social constructionism. The paper illustrates the argument by offering a realist account of how discourse may underpin processes of social construction. This is then applied to the case of the subject, an application that is framed as a critique of Judith Butler’s performative account of the subject.

The paper begins by saying a little more about critical realism's general approach to ontology and social ontology, and developing the argument that it is indeed compatible with constructionism.Indeed it provides us with tools that enable us to develop a more coherent form of constructionism: more coherent because in this form the causal mechanisms that lie behind social construction can be analysed and understood, and only constructionist claims consistent with those mechanisms are accepted.The second section illuminates and concretises this argument by considering the case of discourse, a force that is often regarded as contributing to the process of social construction.Engaging with Michel Foucault's account of discourse, this section examines how we might develop a realist causal account of its influence.Finally, the third section concretises the argument in a different dimension by considering one of the many types of thing that have been seen as socially constructed: the subject.It criticises Judith Butler's account of the construction of the subject and offers an alternative realist model of the construction of the subject.This is a model that is compatible with recognising the reality of agency: the reality of human beings as causally significant in their own right.The paper thus combines and highlights some key elements from my recent work, and more detailed versions of many of its arguments can be found there (notably in Elder-Vass, 2007;2010;2011;2012).

Realism and constructionism
There is space here for only the briefest of introductions to critical realism.Perhaps the best starting point for such an introduction is critical realism's understanding of causality and its significance for the social sciences.For realists the social world, like the rest of the natural world, is driven by causal processes, and therefore social science is at least partly concerned with seeking to explain the causal interactions that produce social events (Bhaskar, 1998(Bhaskar, [1979]]; Elder-Vass, 2010).Critical realists argue that all events are caused by multiple interacting causal powers (Bhaskar, 1975), including for example the powers of individual persons and the powers that we attribute to social structures (Archer, 1995).Even this brief statement already produces quite a different sense of causality than the positivist or empiricist account of cause, because the combination of powers that is present on any one occasion is always contingent, and so causal powers are only ever tendencies: their characteristic effects are not always realised as they may sometimes be overcome by the influence of other contingently present powers.Causal powers, therefore, rarely and perhaps never produce the exceptionless empirical regularities that empiricist thinkers like David Hume identify as the form of causality (Bhaskar, 1975: 33-35;Elder-Vass, 2010: 40-43;Hume, 1977Hume, [1748]]).
For realists, these causal powers are emergent properties.As I interpret this claim, this means that they are properties or powers that are possessed by things, using this word rather loosely to include any persistent assembly of parts.They are emergent in the sense that they are powers that would not exist if the parts concerned were not organised into a certain type of whole.They are therefore a product of the particular organisation of the parts that is characteristic of wholes of this type (Elder-Vass, 2010: chapter 2).This point is most simply explained by using a non-social example, before we move on later to more sociological cases.Consider the example of a laser pointer, composed of a plastic case, a button, an electric circuit, a battery, and a laser diode.When these parts are organised into the form of a laser pointer, the pointer has the power to shine a focussed beam of light at a distant point when the button is pressed.This is a power that depends on the parts but also on the organisation that turns them into a laser pointer, and because of the latter it is a power that is not possessed by the parts unless they are organised into a laser pointer.It is therefore an emergent power of the pointer, a new power that is produced when the parts are organised into this form.
Philosophers of science have debated emergence at some length because it offers an alternative to reductionist ways of thinking about causality (id., ibid.: 53-58).Eliminative reductionists argue that there are no emergent properties because the apparent powers of higher level wholes are always really produced by the effects of their lower level parts.Some emergentists have gone to the other extreme and argued that emergent properties cannot be explained at all.I offer an intermediate view.On the one hand, I argue, emergent powers can be explained by learning the mechanism that produces them: the process of interaction between the parts that produces the power.On the other hand, however, such explanations are not eliminative reductions: if we can explain a power in this way it does not cease to be emergent, it does not cease to be a power of the whole thing, and it does not become a power of the parts, since it is power that does not exist unless the parts are organised into the form of the larger entity -in this case the laser pointer.
Arguments like this become useful to social scientists only if and when they can be applied to the social world.For example, realists have argued that we can explain individual human agents as entities with emergent causal powers of their own, powers that arise from their physical structure: an argument that becomes sociological when we recognise that our physical structures (most obviously our neurological structures, but also muscles, for example) are themselves influenced causally by our social experience (idem, 2007; 2010: 89-98).This is an argument that will be relevant to the discussion of subjects in part three below.
Equally important is the case of social structures.Consider two types of social structure: organisations and norm circles (the latter concept is developed in Elder-Vass, 2010: chapter 6; and 2012: chapter 2).Both are social entities, composed of people (and potentially other parts too), which have powers that those people would not have if they were not organised into these entities.An orchestra, for example, is a relatively simple organisation composed of musicians and instruments.An orchestra has the power to produce harmonious music, a power that those musicians would not have as individuals, a power they would not have without the set of relations and commitments that turns them, collectively, into an orchestra.This is therefore an emergent causal power of the organisation.Similarly, I have argued that we can best explain normative social institutions as the product of social entities called norm circles: groups of people who share a commitment to endorse and enforce a particular norm.Norm circles will be discussed in a little more detail in section two below, as part of a realist causal account of discourse.
This kind of account of social structures is valuable as a response to methodological individualism, the form of eliminative reductionism with which we are familiar in sociology (idem, 2010: 83-84).It does not, however, lead to an equally unbalanced methodological collectivism.Instead, in combination with the broader critical realist account of causality, it enables us to recognise that both social structures and individual agents have emergent causal powers, and that social events, on the model of causality introduced above, are the product of multiple interacting causal powers, including the powers of both individual agents and social structures, and indeed other material objects.
A characteristic feature of social constructionist arguments it that they assign a critical role in the process of construction to language, discourse, and/or culture.Often, invoking some variation of the interpretive tradition, this is linked to the claim that these cannot be given causal explanations.Instead, it may be suggested, they can only be interpreted or understood using hermeneutic methods.Realists, by contrast, are open to the need for interpretive work for making sense of the meanings of language, discourse, and culture, but do not see this as posing an obstacle to including these forces in causal accounts.A realist social constructionism, in other words, would see language, discourse and culture as products of interacting causal powers and also, potentially, as causal forces themselves.This opens up the prospect of seeing social construction as a real causal process, or a family of such processes.By developing a social ontology of language, discourse, and culture we can then develop an understanding of the entities, powers, and mechanisms at work.This in turn should help us to distinguish between viable constructionist claims that are compatible with plausible accounts of such causal processes and non-viable claims that are not.The present paper is a contribution to such an enterprise.
I should stress, however, that mine is by no means the first contribution.Bhaskar, for example, stresses that social structures are concept dependent (Bhaskar, 1998(Bhaskar, [1979]]: 38), thus linking interpretive to structural questions, and has questioned suggestions that there is a conflict between realism and constructionism (idem, 1993: 186).And a range of realist thinkers have argued that realism is compatible with moderate forms of social constructionism while rejecting the anti-realist forms of constructionism referred to above (including, for example, Joseph and Roberts, 2004: 5;Mingers, 1999;Sayer, 2000: 62-63;Sewell, 1992: 12;Sismondo, 1996: 2;Smith, 2010: 119-122).Broadly speaking, this paper builds on the distinction that these authors have developed between moderate forms of social constructionism that are compatible with realism and more radical forms that are neither compatible with realism nor causally plausible.

Discourse after Foucault
The work of Foucault has sometimes been given a radical constructionist interpretation, but many realists have suggested that his work is in fact compatible with a realist view of the social world (e.g.Al-Amoudi, 2007;Hardy, 2010;Joseph, 2004;Pearce and Woodiwiss, 2001;Sealey and Carter, 2004: 46;and papers by Woodiwiss, Day, and Frauley in Frauley and Pearce, 2007).Rather than debating Foucault's intentions, however, this section will draw on Foucault's account of discourse and develop it in a realist direction.Foucault's work on discourse has two features that make it a promising starting point.First, although Foucault may have resisted use of the word "cause" there is a strong sense in his work that discourse shapes our social world in something like a causal way.Second, his reflections in The Archaeology of Knowledge remain the most thorough and coherent analysis of the nature of discourse to be found in the literature (Foucault, 2002(Foucault, [1969]]).
Nevertheless, it remains extremely unclear what mechanism might be responsible for the effects that Foucault claims for discourse, and in particular, it remains unclear how this might relate to the contributions of human actors.In a way what Foucault faces here is the classic sociological problem of structure and agency, a problem that has been a central focus of critical realist work in sociology (e.g.Archer, 1995;Elder-Vass, 2010;Porpora, 1998).What this paper seeks to do, then, is to reuse some of Foucault's materials in a realist theory of discourse that also draws on realist understandings of structure and agency, understandings that as we have seen are in turn informed by critical realism's broader ontology of causality.The product is a theory that recognises that discourse has a causal power, but also that subjects and other social structures have causal powers of their own, and a theory in which we can make sense of how these causal powers relate to each other.
For Foucault, discourse consists primarily of statements and discursive formations.Foucault approaches the definition of statements partly by considering their relationship to sentences.Statements bear some relation to the content of sentences, but they are not the same thing as sentences, and there is a certain ambivalence in Foucault's account of the relation between the two.The obvious reading is that a statement can be equated to the meaning of a sentence.Thus, for example, Foucault argues that if we see a notice that is translated into several languages, or if we hear a speech and its simultaneous translation, the original sentence and its translations do not constitute different statements but rather are simply different instances of the same statement (Foucault, 2002(Foucault, [1969]: 116-117)]: 116-117).What seems to be preserved between these different sentences is indeed the meaning, and I shall interpret the concept of a statement in this way, but it is worth noting that Foucault is somewhat resistant to this interpretation.This may be a consequence of a desire to avoid the complexities of hermeneutic methodology.Foucault is, primarily, studying past statements, but has no interest in potential multiplicities of interpretation of those statements (id., ibid.: 31).His interest is in how those statements were regulated in the period in which they were produced, and today's interpretations are irrelevant to this question; it is the question of how they were understood by contemporaries that is significant for the process of regulation, but it is the regulation that Foucault is interested in, not the interpretation.
Foucault's discussion of translation also prompts another important point: statements are not a linguistic phenomenon.They are of course expressed using language, but the language itself is not what Foucault is interested in.Foucault is interested in the ways in which what is said is regulated (the content of our statements), not in the ways in which how it is said is regulated (the language we use to express them), and these two issues are governed by two different sets of rules or norms. 3Linguistic rules govern features like vocabulary, grammar and pronunciation, but Foucault's interest is in discursive rules: rules that govern what can be said and what should not be said.
It is sets of such rules that constitute the second element of discourse: discursive formations.These are sets of rules about what statements can be made in a given social context, rules that Dreyfus and Rabinow (1983: 70) call "principles of a rarefaction" because they tend to cause the things that are actually said to cluster within a limited subset of the things that it would be linguistically possible to say.
Discourse, for Foucault, then, is some combination of the statements that are made and the rules that govern those statements.But what is the relation between these two components, and how could the discursive rules produce the effect of regulating the statements that are produced?Foucault has some clues to offer.These rules, he says, are to be found and operate "not only in the mind… but in discourse itself" (Foucault, 2002(Foucault, [1969]]: 69).This, of course, implies that they do operate in the mind, but not only there, so here he seems to be offering some place to the human individual in the process of reproduction of discourse, and yet at the same time he rejects a purely subjectivist account of discourse.These rules, he also says, operate "without reference to a cogito" but "in the totality of things said" (id., ibid.: 138).The implication of the first half of this statement would seem to be that they are lodged in our brains and operate on our discursive acts but without being under the control of an autonomous self.The implication of the second is that the mechanism of their reproduction is driven in some way by the accumulation of past statements.But this is as far as Foucault takes us in the direction of explaining the causal mechanism at work.Discourse itself, it seems (in an argument that parallels the work of Niklas Luhmann), affects the production of further discourse, but it is not at all clear how.
For a realist this story is frustrating: Foucault is telling us that discourse makes a difference to the social world, but stops short of explaining how this could possibly be the case.Can we improve on this story?Can we develop it into a coherent causal explanation?
This paper argues that the general points about ontology and cause discussed in the previous section can be applied to this case: that we can explain the causal influence of discursive rules as the consequence of an emergent causal power of social entities.Discursive rules are norms of a kind: norms about what we should say, write or think.It may therefore be possible to explain them in the same terms as other normative social institutions: in terms of causal powers of the social entities I have called norm circles (Elder-Vass, 2010: chapter 6).A norm circle is a group of people who are committed to endorsing and enforcing a specific norm.Every norm, I argue, has such a group of people standing behind it.They criticise and punish those who fail to observe the norm concerned, and they praise and reward those who do observe it.As a result of this sanctioning behaviour, individuals exposed to the influence of the norm circle tend to develop dispositions to conform to the norm.These dispositions, of course, are only tendencies.Like other causal powers, these ones may be frustrated on any particular occasion by countervailing powers.Thus, for example, my tendency to step aside to let someone pass may be frustrated if I am in a close-packed crowd that prevents me from doing so.Thus the norm circle, a social entity with an emergent causal power arising from a process of interaction between its members, operates through its members to exercise that causal power: the power to create a tendency for affected individuals to conform to the norm concerned.
Discursive rules may operate in just the same way as norms more generally: they have a causal impact on individuals as a result of the influence of a discursive norm circle, a norm circle committed to the discursive rule concerned (Elder-Vass, 2012: 153-157).Thus, for example, there is a set of discursive rules about what can be said and what should not be said in articles in academic journals, and these rules are causally effective because there is a group of people -primarily journal editors and reviewers -who are committed to enforcing them and who have the power to sanction academic writers in support of these rules.Similarly, there have been historically variable rules about how we may and how we should not speak about madness, which has had profound effects on the ways in which the mad have been treated in our societies, as Foucault has demonstrated (2001Foucault has demonstrated ( [1961]]).
This way of thinking about discursive rules, I suggest, offers a productive synthesis between Foucault's work on discourse and realist social ontology.In particular, it solves the causal problems that undermine Foucault's account of discourse.Now, the causal processes and mechanisms that make it possible for discursive rules to affect us are clear.There are groups committed to these rules that enforce them.Furthermore, the role of the individual subject is clear: individuals tend to act in conformity with the rules because the pressure from discursive norm circles leads them to develop dispositions to comply with the corresponding rules, but they only tend to act in conformity with them, because other causal powers also affect them (e.g.other norms, chemical influences on their emotions, their own prior decisions about their projects in life, etc.).Thus we get general conformity with the rules because there is a real social force that backs up those rules, but transgression and change are also possible, whether as a result of deliberate innovation or as an unintended consequence of non-compliance, because there are also other social forces (and indeed physical forces and individual reflexive agency) influencing our behaviour.Discursive constructionism, it turns out, is not just compatible with a realist ontology, but stronger and clearer as a result of being combined with it.

The constructed subject
We may examine the issues raised by discursive constructionism by considering some of the arguments that have been offered regarding the construction of one specific type of object.Let us take the example of the subject.There is a certain aptness in examining this case, not only because it has been a prime site for constructionist arguments, but also because of the significance of the subject's role in the debates discussed above on the nature of discursive power itself.
The concept of the subject is closely linked to the concept of agency, but its usage within the poststructuralist tradition has enmeshed it in a network of connotations that convey a certain scepticism about the traditional concept of agency.The idea that our subjectivity is socially constructed has come to be associated with the idea that our subjectivity is in some way inauthentic or compromised: that our subjectivity is not what it might appear to be; that the conventional notions of agency and the subject, in other words, refer to something that is not real.Yet, I would like to suggest, this implicit dismissal of the human capacity for agency on the basis of arguments for the social construction of the subject is not tenable.Instead, I will argue, we can develop a more moderate realist constructionism about the subject, a constructionism that is compatible with the possession of an element of autonomy by individual human agents.
One difficulty in this debate is that the concept of the subject is used to refer to many different aspects of the human individual.Elsewhere I have identified eight different usages that are relevant to this debate (Elder-Vass, 2012: 184-187); let us focus here on just four of these, and some of the relations between them.First, we have what we may call the agentic subject: the person who is capable of reflection and choice.This is a version of the concept of the subject that I will defend.Second, there is the Cartesian subject, the subject as a free-floating rational mind, more or less disconnected from the material world and with universal qualities that are independent of any social history.This is the concept of the subject that is commonly attributed to Descartes, a concept of the subject that is widely rejected.Third, we have the authorised subject: a person who is recognised as being authorised to make non-trivial decisions.Much of the feminist debate on subjectivity has been driven by the denial of this status to women.
Early attacks on the concept of the subject by Althusser and Foucault were concerned with the relation between such concepts and a fourth one, the political subject: a person who is subject to political power, a subject of a state or sovereign.Althusser and Foucault, in their different ways, argued that our sense of being agentic subjects, free agents, is produced by ideology (for Althusser) or a discursive formation (for Foucault) as part of a process of ensuring that we accept our status as political subjects (Althusser, 1994(Althusser, [1971]]; Foucault, 1983).Our sense of freedom, these arguments imply, is historically contingent, a recent product of an ideology or discursive formation that secures our complicity in our own domination by enticing us to believe that we choose it freely.It is this sense of subjectivity, Foucault seems to argue, that in some future historical period may be "erased, like a face drawn in sand at the edge of the sea" (Foucault, 2002(Foucault, [1966]]: 422).This is a very odd argument.Why would it be necessary to persuade us that we choose our subjection freely if there was not already some danger that we might refuse to comply with power?And if there is such a danger, doesn't that mean that our autonomy pre-exists the supposed moment of construction of our subjectivity?And indeed Foucault himself is ultimately committed to the need for resistance and to the existence of some capacity to resist.So perhaps the agentic subject is more than a discursively/ideologically produced illusion after all.We could still accept, of course, that the particular form of our understanding of subjectivity is socially constructed, but the undercurrent of scepticism about agentic subjectivity itself in the poststructuralist tradition cannot be sustained on this basis.
Perhaps the most important recent expression of that scepticism has come from Judith Butler.In her debate with Seyla Benhabib, Butler has attacked the idea of the agentic subject (Benhabib, 1995a;1995b;Butler, 1995a;1995b).Butler questions the possibility of knowing anything except through discourse and therefore questions the idea of a subject outside discourse (Butler, 2006(Butler, [1990]]: 202).Instead, she suggests, we should see the subject as a linguistic site or category (idem, 1995b: 135).People become subjects, she says, only by occupying the site of the subject, as speaker or addressee, in speech acts, which are therefore performative of being a subject (in a sense of performativity that is drawn from the work of J. L. Austinsee Butler, 1997: 10-11;1995b: 134).This perhaps makes most sense if we read subject in the sense of authorised subject defined above: people become authorised subjects (or not) on the basis of a sedimentation of discursive acts.Butler, as far as I am aware, however, tends to read the concept of the subject rather more loosely and widely than this, and I have suggested elsewhere that her critique of Benhabib depends heavily on various conflations between different concepts of the subject.In particular, her attack on the agentic subject is constructed largely by conflating it with the Cartesian subject, as if these two concepts were one and the same, when in fact Benhabib is defending the agentic but not the Cartesian subject (Elder-Vass, 2012: 192-193).
For Butler, then, the subject is the outcome of a linguistic process of performative construal that occurs during speech acts (Butler, 2006(Butler, [1990]: 195)]: 195).One implication of her argument seems to be that we are only subjects during these speech acts, and that the subject is recreated in every such act, as if our bodies were empty shells between one speech act and the next.But, she insists, there is also an element of iterability to such performatives: a permanent possibility of realisation of the subject position that is built into the structure of our language (idem, 1993: 13; 1995a: 47; 1995b: 134).This possibility, she argues, arises from the linguistic conventions that reserve a place for the subject in our patterns of discourse, rules that require us to refer to persons and locate them as actors in our statements (idem, 1995b: 134).But although Butler claims that this retains a place for agency in her social theory, it is an agency that seems to be stripped of any place for individual reflexivity or autonomy, and indeed it is hard to see what opportunity such agents could ever have for altering the discourses in which they are embedded.Butler does not seem to escape from the problem that faced Foucault in his earlier work: the problem of an account of subjectivity that is so much a product of external forces that the capacity for critique and resistance is inconceivable, while on the other hand her own critical political agenda depends upon the existence of such a capacity.

Realism and the constructed subject
For realists, the subject is something more than a position in discourse.For realists, we are real physical human beings capable of reflection and choice (Archer, 2003;Shilling, 2005).These capabilities are emergent properties of living human bodies, arising from the composition and structure of those bodies, including their brains (Elder-Vass, 2010: 89-93).These are capabilities that continue to exist from moment to moment independently of any need for discursive speech acts, though the form they take is certainly influenced by prior such acts.It is the plasticity of the human brain, the ways in which its neural structures are constantly reshaped by our experiences, whether discursive or practical, that makes it possible for prior discursive acts to influence subsequent discursive structures.As our brains are influenced by the discursive pressures exerted on us by the discursive norm circles that form an important part of our social context, we develop new or subtly altered dispositions, including dispositions to exert normative pressures ourselves.This is the transmission belt that enables prior practice to influence future practice, but any consistency in such influences depends on the existence of groups of people with systematically similar normative commitments.
The consequence is that we have a capacity to reflect and decide that exists independently of any particular socialisation, but that the workings of that capacity are influenced by our knowledge, beliefs and dispositions, which alter according to our experience.A further consequence is that as agents we can be influenced by social structure, although different theorists emphasise different pathways through which this can occur -for Archer, for example, the pathway is the knowledge we have of our social context; and for Bourdieu, it is the dispositions that we may acquire unthinkingly from our context (Elder-Vass, 2010: 99-108).The implication is that we are agentic subjects -reflective, decision making individuals, even if at times we act unreflectively -but not Cartesian subjects -we are embodied, shaped to some extent by our social context, and we are not free floating asocial rational minds.Now this, I suggest, is compatible with a moderate constructionism about subjects.Although we have the capacity to be agentic subjects independently of any particular social context, what kind of subject each of us becomes does depend on the processes of social construction.This is most striking when we consider the case of authorised subjects.The kinds of people we are accepted as, and the kinds of actions, whether discursive or practical, that we are authorised to perform, depend upon performative acts that enact social norms -norms of gender, class, and social role, for example.And here we may connect the argument back to what was said above about discourse.The place of the subject in statements, the issue of what may or may not be said about subjects and indeed of who may qualify as an authorised subject in discourse, is certainly influential in shaping the kinds of subjects that we are, the kinds of freedom for example that we feel able to enact.The place of the subject in statements is in turn a product of discursive formations -what Butler calls linguistic conventions -but this is not the end of the matter.Those formations or conventions are normative, and they do not exist in some free floating abstract form, nor in some miraculously similar form in a multitude of individual heads.They are the product of real social groups, of discursive norm circles, who provide the systematic pressures that sustain the norms concerned within the population.
This kind of constructionism still allows us space for critiques of hegemonic conventions, and thus for some at least of Butler's political agenda.But it also solves the problem of resistance in poststructuralist theory: as real material beings with the capacity to make decisions that are only partly shaped by our social context we have the capacity to do otherwise than might otherwise seem to be dictated by hegemonic conventions.Such possibilities multiply when other norms develop that may be in conflict with hegemonic ones, and we are obliged to make choices between them. 4The realist constructed subject thus has the autonomy of the agentic subject: not the pure autonomy of the Cartesian subject but enough autonomy to make resistance and change possible.The extent and form of that autonomy, of course, may vary depending on the historical conditions, which influence the range of options within which we may have realistic possibilities of making choices as well as the kinds of people that we believe ourselves to be.
In saying so, we open up further vistas.The causal chains at work in society do not start or end with discourse and the normative processes that lie behind it.The influence of norm circles itself is the result of a causal history, and when we investigate the relation between individual agents and norm circles we see only one part of that multiply determined causal history.In giving a fuller account of that causal history we must be conscious of the wider social powers that influence our normative institutions.Behind the influence of ideology, for example, Marxists see the power of capital.In abandoning ideology for discourse, Foucault sidelined the question of what social powers lie behind discourse, although he later took up this question from a rather different angle through his engagement with the power/knowledge relation.Perhaps there is an element of autonomy in the discursive circuit, as thinkers like Luhmann would have us believe, but our discursive norms are also heavily shaped by other forms of social powercapital, governments, media corporations, and social movements, for example.A truly critical constructionism will also need to engage with their influence.To do so in any detail, however, is beyond the scope of this paper.For an interesting discussion of the relation between "hegemonic norm circles" and "norm circles of resistance" see Cresswell, Karimova and Brock (forthcoming).

Conclusion
This paper has argued that social constructionism must be combined with a critical realist social ontology if it is to offer a coherent approach to developing critical social theory.If social constructionism is to be plausible we need an account of how social construction actually works, an account that is consistent with our understanding of the material world of which we are part.Imperialistic constructionisms that see everything as constructed and that turn away from the reality of our world are ultimately self-defeating, both because they are impossible to reconcile with what we know of our sheer materiality, and also because they make it impossible to justify substantive ethical views that enable us to criticise what exists.This paper has sought both to outline and to illustrate the view that we can productively combine a more moderate constructionism with a realist ontology.We can, for example, develop plausible and coherent causal accounts of the influence of discourse on our dispositions, beliefs and actions, by seeing that influence as a causal power of the discursive norm circles that endorse and enforce discursive rules.The ontology advocated here also enables us to make a clear and plausible connection between such social entities and the individual human agents that make them up.Those individuals are independently material people with casual powers of their own, yet they are also shaped and influenced by discursive pressures.Realists can therefore accept that subjectivity is socially constructed in the moderate sense advocated in this paper without denying the reality of the agentic subject, as more extreme constructionists like Butler seem to do. 5  The argument of this paper has focused on the cases of discourse and the subject, but similar arguments can be made about language, culture and knowledge as mechanisms of construction, and about many other potential objects of construction.Further discussion of many of these, and indeed more in depth discussion of some of the themes of this paper can be found in Elder-Vass (2012).In all of the cases I have considered, it appears to be both possible and productive to combine a realist social ontology with moderate constructionist theory.Realists, I suggest, should be social constructionists; social constructionists should be realists; and social scientists in general should be both.

Towards a realist social constructionism
Social constructionism has often been seen as incompatible with realist approaches to the social world.This paper argues that critical realism is thoroughly compatible with moderate versions of social constructionism and indeed provides stronger ontological backing for it than the anti-realist approaches that are often associated with more extreme versions of social constructionism.The paper illustrates the argument by offering a realist account of how discourse may underpin processes of social construction.This is then applied to the case of the subject, an application that is framed as a critique of Judith Butler's performative account of the subject.

Pour un constructivisme social réaliste
Le constructivisme social a souvent été vu comme incompatible avec les approches réalistes du monde social.Cet article soutient que le réalisme critique est totalement compatible avec des versions modérées du constructivisme social et que, en fait, il lui apporte un plus grand support ontologique que les approches antiréalistes, souvent associées aux versions plus extrêmes du constructivisme social.L'article démontre cette théorie en proposant une description réaliste de la façon dont le discours peut soutenir des processus de construction sociale.Cet argument est ainsi appliqué au cas du sujet, selon une proposition encadrée en tant que critique de la description performative du sujet, de Judith Butler.