Diversity discourses and the articulation of discrimination: the case of public organisations

ABSTRACT Public organisations are contexts that particularly further a contested diversity discourse. They have a long tradition of various equal opportunity policies and are characterised by an internally differentiated structure. Based on interviews in a university and two city administrations in Austria and Germany, this study shows that the ubiquitous ‘diversity turn’ is not intrinsically opposed to the articulation of discriminations and social inequalities. Rather, diversity discourses in public organisations are characterised by a ‘heterogeneous space’, which enables the recourse to discriminations of groups of people marked as diversity-relevant as well as to instrumental benefits of diversity-related activities. As a consequence, the organisational diversity discourses and their related practices are far less driven by one-dimensional motives of marketability than commonly assumed.


Introduction
More than 10 years ago, Ragins and Gonzalez (2003, 133) noted that 'the slippery construct of diversity has become "an empirical onion" that reveals more complexity with time and analyses'. This assessment appears even more true in view of today's research landscape, since open questions (e.g. the importance of business vs. social justice rationales) are still unresolved, and new fields of literature are emerging (e.g. context-sensitive studies of diversity policies/management in practice) (Tomlinson and Schwabenland 2010).
Following a non-deterministic, context-sensitive and empirically grounded approach to diversity (management) studies, this paper takes the perspective that the specific conception of diversity, its related practices and their potential effects, are not an 'a priori given' but subject to constant negotiations by more or less strategic and powerful organisational actors (see also Zanoni et al. 2010). By analysing how public organisations in Austria and Germany implement, shape and adjust the diversity discourse, 1 I argue that drawing on diversity rhetoric and claims for social justice are not mutually exclusive. Rather, the empirical investigation indicates that the diversity discourse is potentially permeablehowever not in an arbitrary senseand that different, seemingly contradictory, rationales can simultaneously be deemed relevant in organisations. As I will demonstrate, pointing out social inequalities and cases of discrimination, as well as referring to instrumental benefits of diversity management, are both part of diversity discourses in organisations. By focusing on the 'extreme cases' of public organisations, that is, organisations characterised by non-profitability and the claim to be representative of the population they serve, I aim to gain insights into what is 'sayable' and 'doable' under the 'diversity paradigm'. Before presenting the empirical data in detail, I reflect on current critical diversity (management) literature 2 and its pivotal lines of debate. In the next step, I clarify the theoretical framework that guided this study and specify my research questions. This is followed by an overview about the process of data collection and analysis. The presentation of results is rounded off by the discussion of their relevance for future diversity research.
The 'diversity turn' and its critics Despite manifold and long-standing research efforts, there is no consensus on whether the 'widespread "diversity turn" in governance and management' (Vertovec 2012, 287) is able to create substantial change for the inclusion of historically disadvantaged groups and for increasing equal opportunities in general. The criticism is multifaceted and ranges from reservations regarding the effectiveness of diversity policies and diversity management in practice to fundamental doubts about the usefulness of the diversity concept as such. With regard to the implementation of diversity-related practices, one major concern is that they might primarily aim at generating a 'happy image of diversity' (Ahmed 2012, 152) instead of actually transforming the organisational culture and its corresponding structures (see also Nkomo and Stewart 2006). This assumed lack of effectiveness is often traced back to the conceptionor more specifically: the 'non-conception'of diversity itself: scholars describe it as 'broad and ill-defined' (Noon 2007, 780) or as an 'empty container' (Ahmed 2012, 80), into which you can throw all sorts of arbitrary differences, which may result in neglecting inequalities based on group membership and/or power differentials (Linnehan and Konrad 1999;Ragins and Gonzalez 2003). This leads to an insufficient consideration of historically established power relations in organisations, which can result in the implementation of diversity-related practices that solely focus on individual organisational members (e.g. diversity training) instead of deeply intervening in core principles of organisational inclusion/exclusion policies (e.g. definition and evaluation of performance, work arrangements). This tendency of 'surface polishing' inherent to diversity policies/management is partly explained by the widespread assumption that 'most problems of management are problems of motivation rather than structure' (Kalev, Kelly, and Dobbin 2006, 591). As a consequence, a large proportion of diversityrelated change initiatives take the approach of changing the individual organisational member instead of organisational structures and processes (ibid.; see also Bell and Hartmann 2007;Zanoni et al. 2010). Further, scholars argue that the ineffectiveness of diversity-related practices is caused by the overemphasis on a 'managerial perspective' ; see also Lorbiecki and Gavin 2000;Thanem 2008), i.e. on how managers should handle or create a diverse workforce instead of concentrating on the needs of all organisational members at different hierarchy levels and work areas.
Apart from individualistic approaches, there are also group-based conceptions of diversity, which, however, are equally subject to criticism. This is because the groups that are marked as diversity-relevant are mostly characterised by assumingly fixed, 'naturally' given and one-dimensional categories (e.g. gender or race/migration background). Some scholars point out that the individuality of the group members is overshadowed by their function as 'representatives of a social group distinguished by a common sociodemographic trait, the repository of a "true", essential identity' (Zanoni et al. 2010, 13). Further, differences within the diversity-relevant target groups (e.g. the various experiences and needs of people with different migration backgrounds) would become marginalised whereas the distinctions between groups might gain hyper-visibility (e.g. people without vs. people with any form of migration background) (Litvin 1997; see also Stone and Colella 1996).
Another feature of such homogenised and essentialised constructions of diversity target groups is that they mainly refer to 'those who "look different"' (Ahmed 2012, 33) such as women, people of colour or with 'obvious' migration/ethnic background (see also Ely 1995). In contrast, white males 'do not have to be labelled. They are neither racialised nor gendered. White men are not represented as a group but as individuals.' (Nkomo and Stewart 2006, 533) In fact, by making those who are marked as diversity-relevant an explicit management responsibility, 'the distance and difference between those who manage and those who are managed is maintained' (Thanem 2008, 583) or even reinforced. The 'diversity turn' may result in the unintended consequences of further confirming the congruence of already powerful groups (e.g. male, white, able-bodied, 'mainstream-religious', heterosexual managers) with dominant organisational values (e.g. working full time vs. part time) and norms (e.g. working on weekends, socialising after work) instead of questioning them. Litvin (1997) aptly summarised this criticism when she diagnosed already about 20 years ago that the diversity discourse navigates between an individualism, which is insensitive towards inequalities, and a potentially essentialising collectivism. Both perspectives were not paying enough attention to power relations. Some scholars take their criticism even further and identify the 'diversity paradigm' and its related practices as one of the causes of the increasing inequalities in contemporary societies, rather than as a potential solution to them (see e.g. Ahmed 2012;Bell and Hartmann 2007;Michaels 2006). This deep scepticism is connected to the widespread assumption that the current diversity discourse is permeatednot to say determinedby demands of marketability (see e.g. Eveline, Bacchi, and Binns 2010;Noon 2007;Prügl 2011). Altogether, the critical diversity literature assumes an insurmountable gap between concerns of social justice and business rationales.
In contrast, Kamp and Hagedorn-Rasmussen (2004, 530) have argued that at 'this general level' of debate 'the authors tend to overestimate the stability of [the diversity] discourse' and to undervalue the dynamics and potential unpredictability of its encounter with specific organisational discourses embedded in societal frameworks. In addition, only few studies actually trace the effects of specific diversity-related measures (e.g. diversity staff position, mentoring programmes, diversity trainings) on the inclusion and promotion of historically disadvantaged groups (e.g. Dobbin, Kim, and Kalev 2011;Kalev, Kelly, and Dobbin 2006). This lack of empirical work can be explained by the fact that longitudinal, empirical data on the composition and development of the workforce in conjunction with practices of diversity policies and management are hardly available and difficult to collect, partly due to personal data protection.
Nevertheless, there is a growing body of empirical studies that are based on a nondeterministic, context-sensitive approach: For instance, Kamp and Hagedorn-Rasmussen (2004) studied the implementation of diversity management in a Danish municipality and found that the applied diversity discourse did not only differ across departments but also builds on and merges with a social responsibility discourse already in place. This shows that even if the diversity discourse and its related practices were exclusively driven by business rationales, their encounter with a specific organisational context may result in a reframing 'under the influence of established institutions' (537). Tomlinson and Schwabenland (2010) extracted from their study of diversity management in non-profit organisations in the UK, that key actors tried to reconcile 'moral and business rationales through re-inscribing utilitarian arguments within an organisational commitment to social justice' (101). However, they concluded that this reconciliation did not work seamlessly but still left room for discursive agency embedded in daily working routines. Similarly, by comparing practices of diversity management in a hospital and a technical drawing company in the Flemish Region, Zanoni and Janssens (2007) illustrate that minority employees are not helplessly exposed to managerial control attempts but rather 'draw on managerially inspired discourses in as far as such discourses allow them to construct a positive identity' (1394). This suggests that diversity discourses in organisations are permeated by multiple rationales that create a heterogeneous field of what is understood as diversity ir-/relevant.

Theoretical framework: a discourse analytical perspective
In order to analyse how public organisations adopt the diversity discourse and adapt it to their local conditions, it is crucial to elaborate the theoretical framework and its respective epistemological foundation, which informs the diversity concept of this study. This is important because 'there is diversity even in the way diversity is defined' (Ragins and Gonzalez 2003, 133), and researchers thus need to make their underlying understanding of diversity explicit. Furthermore, this definition of diversity reveals the position from which the empirical diversity discourses in public organisations are evaluated. The definition is carried out in two steps: First, I outline the specific approach towards diversity, where a context-sensitive, relational perspective plays an important role. Second, I illustrate the epistemological foundations in which this particular conception is embedded.

Diversity as twofold relationality
My approach towards diversity is characterised by the basic assumption of its relationality (Ely 1995;Pries 2013;Puwar 2004). This means that the differences, for which diversity functions as an umbrella term, cannot be explained by any form of substantial essence. For instance, being categorised as 'female' or as having an 'ethnic or migrant background' is inextricably linked to its assumed and simultaneously excluded opposite: the attribution as 'male' or 'white'/'non-ethnic'. However, this mutual constitution of elements, which forms the basis for the act of differentiation, does not imply that all of them have socially equal consequences. As Puwar (2004, 57-58) points out, some elements enjoy the privilege of 'invisibility': With "whiteness" defined as an absence of colour' (…), whiteness exists as an unmarked normative position. Similarly the male body is invisible as a sexed entity. Its absence of gender entitles it to take up the unmarked normative locale. (…) Left unnamed and unseen, invisibility in this context is clearly a place of power.
Consequently, a relational conception of diversity needs to take the specific power relations into account, which are inherent in the constitutive conditions of differences and constantly (re-)produced, shifted and thereby potentially transformed by every act of differentiation. Against this backdrop, the diversity concept presented here is characterised by a twofold relationality: On the one hand, each 'diversity dimension' is relationally constituted. On the other hand, the umbrella function of diversity for these elements is relational, as well. In this respect, diversity can be defined as 'the collective amount of differences among members within a social unit' (Harrison and Hock-Peng 2006, 196;emphasis in original). By emphasizing the context (social unit) it becomes clear that diversity must not be interpreted in terms of personal traits, but rather as a relationship between variables marked as diversity-relevant. Against this background, I consider diversity to be relative and not absolute. Thus, depending on which dimensions are defined as diversity ir-/ relevant, a group of people can appear differently 'diverse' (e.g. by pointing to different regional origins in a case where all group members are of the same nationality). Such a relational conception of diversity does not focus on certain categories of people; rather, it allows bringing the power relations within specific contexts to the fore (see also Ely 1995).

Diversity as constituted by discursive practices
This relational and 'power-conscious' concept of diversity, however, should not be misunderstood as the 'correct' definition of diversity. Rather it represents a specific understanding of diversity within the context of current (discursive) practices labelled as diversity management. The epistemological foundation of such a definition is based on the assumption that what we perceive as 'diversity' or 'diverse' is neither an objective fact nor an ontologically given. Rather, by following post-structuralist assumptions and in particular Foucault's (1972, 49) discourse analytical reflections on the connection between knowledge, discourse and power, I understand diversity as constituted by discursive 'practices that systematically form the objects of which they speak'. This means that there 'is no diversity-in-general; the conditions under which diversity becomes an object to be known (and critiqued) is a result of particular and specific circumstances of knowledge production' (Ahonen et al. 2014, 269). This production of 'diversity knowledge' can be more (e.g. advice literature) or less (e.g. organisational diversity discourse) intentional and also diversity research itself plays a part in its production.
With respect to the investigation of how public organisations deal with the 'diversity turn' in practice, it is of interest what body of knowledge organisational actors rely on when talking about diversity-related issues and how they distinguish between legitimate and illegitimate diversity. In this way, we may gain insights into whether the 'diversity paradigm' furthers only 'happy talk' and exclusively pays attention to 'marketable' differences or whether pointing out social inequalities and finding a voice for marginalised (groups of) people is (also) possible. From a discourse analytical perspective, this does not mean to investigate individual actors' motivations or hidden meanings underlying the organisational diversity discourses (Foucault 1972, 162). Rather, the focus is on identifying the particular 'conditions' (ibid. 28) that enable the actual articulation of statements in other words: what is 'sayable' and thus perceived as 'doable'within a given context. The 'conditions' are understood as the 'rules that made the statements possible and that simultaneously already reside in the (system of preceding) statements.' (Diaz-Bone et al. 2007, 6) With respect to the diversity discourse, 'following the rules' means, for instance, to intelligibly build on knowledge that is already marked as diversity-relevant and at the same time implicitly or explicitly differentiated from what is perceived as diversity-irrelevant. Hence, it allows insights into the potential as well as limits of the diversity discourse and the inclusions/exclusions that are thereby produced. Taking such a theoretical perspective, the specific research questions addressed in this paper are as follows: Which conditions enable the articulation of statements within the organizational diversity discourses? What does this tell us about the distinction between legitimate and illegitimate diversity and thus the recognition or denial of different claims for inclusion? Before answering these questions, I briefly present the methodological approach and the data set.

Methods and data
The data collection was guided by the concept of 'theoretical sampling', which is based on theoretical and problem-oriented assumptions instead of statistical representativity (Glaser and Strauss 1967). The exclusive focus on public organisations 3 was inspired by Eisenhardt's (1989, 539) advice to choose 'extreme situations' that are 'likely to replicate or extend the emergent theory'. The initial assumption was that by looking at how public organisations justify, frame and apply the 'diversity turn' in practice, we could gain insight into 'the maximum' of its inclusionary potential apart from questions of marketability. Accordingly, the three cases were selected because the two city administrations (Citymin A & B) and the university (Unimin) presented themselves as diversity affine, which was rare at the time of the data collection in study context (July 2011-May 2012. Further, all three organisations have long and outstanding traditions of equal opportunity policies with respect to different disadvantaged groups (e.g. women, migrant people as well as lesbian and gay people). And the organisations had established either a separate diversity staff position and/or a diversity network across various departments. Both the history of equal opportunity policies and the structural basis for the diversity-related activities are assumed to be essential prerequisites for the effective implementation of diversity management (Dobbin, Kim, and Kalev 2011;Kalev, Kelly, and Dobbin 2006).
I conducted a total of 16 semi-structured interviews with organisational members, who had been more (e.g. diversity manager) or less (e.g. representative for disabled people) involved in explicitly diversity-related practices. By interviewing organisational members with various kinds of involvement with diversity management, I intended to get an empirically grounded and broad impression of what is 'sayable' and subsequently thought of as 'doable' within the organisations. However, this does not mean that I was able to take a look at the actual discursive practices in everyday life or that the interview-based statements represent passive reflections of the organisational reality. Rather, I perceive them as part of a general, macro-level diversity discourse, which organisational members both (re-)produce and thereby potentially shift. Against this methodological assumption, I consider the 'artificial' situation of the interview as an opportunity, rather than a problem, because it allowed me to gain insight into particularly typical statements of the diversity discourse: 'In an interview situation, typically an encounter between two strangers, (…) in order to say something that the interviewer can grasp relatively easily, the interviewee must rely on established cultural resources for describing issues at hand.' (Alvesson 2003, 20) The organisations had different departmental structures. Thus the choice of interview partners was adapted to the particular organisational context (see Table 1). The interviews lasted one hour on average, were conducted at the workplace of the interviewees, tape recorded and fully transcribed. The interview guidelines focused on three main areas: (1) establishment of diversity management in the organization and the respective meaning of diversity/diversity management (e.g. What does diversity mean to you? What are the components of an appropriate diversity concept?); (2) the actual activities and measures that were marked as diversity-relevant and their embeddedness into every day working life (e.g. What are the activities associated with diversity management in your organization? Did you have come in contact with diversity management in your previous workweek?); and (3) the desired and undesired characteristics of prospective employees and colleagues (e.g. What tips would you give a person intending to apply for a job in your organization?).
The data analysis was based on the method of open coding (Strauss and Corbin 1990), which means to break the interview material up into (research-)relevant units of meaning and categorise them with codes indicated by the empirical data at hand. The coding process followed a two-step procedure. First, I identified statements that described different meanings of diversity and related practices, and analysed how these statements were patterned by continuously comparing established codes within and across interviews. Second, I analysed each pattern with respect to the conditions that enable its articulation as an intelligible component of the diversity discourse. In this context, the guiding questions of the analysis included: What are the necessary elements of a statement that mark it as diversity-relevant? Which components of a statement would need to be changed or removed in order to become diversity-irrelevant? What are the differences and similarities of these conditions underlying the diversity ir-/relevance of statement patterns across the interviews? In the following section I describe in detail the conditions of discourse participation that I reconstructed with the help of these heuristic questions.
Three conditions for participating in the diversity discourse of public organisations Altogether, I identified three conditions of statements that guide the diversity discourses across all three public organisations under investigation. The analysis shows that organisational members need to refer to some form of deviation of people marked as diversity-relevant or to demonstrate any instrumental benefit of diversity (management) in order to intelligibly participate in the organisational diversity discourse. This corresponds with results of studies exclusively conducted in profit-oriented companies (Lederle 2008;Zanoni and Janssens 2004) but also to a certain degree with Tomlinson and Schwabenland's (2010) case study of non-profit organisations in the UK. Moreover, the data indicate that also the recourse to various forms of discrimination represents a valid condition in order to enunciate statements of recognised diversity relevance. By subsequently describing each condition in detail, I show that the interview-based statements provide a much more heterogeneous and far less predetermined room for articulating diversity-related issues than the scientific debate on the seemingly insurmountable differences between the business case and claims for social justice would suggest.

Condition of statements: recourse to deviation
The first condition for intelligibly participating in the organisational diversity discourses is based on the recourse to any form of assumed deviation. This deviation appears attached to groups of people, who are characterised as 'different' and thereby labelled as diversityrelevant. Actors define these differences by referring to mostly sociodemographic characteristics which they perceive as naturally given, decoupled from each other and embodied.
The actual selection of these characteristics is largely influenced by European and national non-discrimination laws, which explicitly aim at preventing discrimination linked to race/ ethnicity, gender, religion/belief, disability, age and/or sexual orientation. The importance of these legal guidelines is reflected in the statement by the Staff Member of the Department for Integration at Citymin B, who explains that building on the attributes that are legally protected by non-discrimination laws can be useful for a comprehensive equality strategy: I agree with it [the list of legally protected attributes], I mean, of course, the academic debate about them is much, much broader, keyword intersectionality or keyword class membership. But for the purpose of a strategy, I think, the attributes are suitable.
Similarly, the Department Head for Disability at Citymin A argues that she 'would not like to evaluate' the different dimensions covered by the non-discrimination laws: 'I think, you should look at them together and do not rely on figures. This I would find very difficult and the issue of disability could be in danger of being pushed into the background.' Here, the interview partner underlines the equal relevance of all six legally covered dimensions, probably also due to the fact that disabilityher specific working areais often perceived as a niche topic and thus threatened by getting too little attention. Generally, the characteristics listed in the non-discrimination laws represent a common ground for defining the umbrella function of diversity. However, which attributes become diversity-relevant in everyday organisational discourse and its related practices depend on the objectives of the organization and working priorities of the organisational actors. This is illustrated by the Department Head for Quality in Higher Education at Unimin, who describes that the university 'felt obliged to do something for the non-traditional students', to help them develop an adequate 'academic habitus' and thus decided to focus on 'two core dimensions of diversity, namely the social and ethnic background'. Likewise, the Department Head for Gender Equality at Unimin describes the ethnic and socio-economic background as important diversity dimensions, but emphasises that 'gender'according to her working areais equally crucial. At Citymin B, the Member of the Works Council does not refer to the social background as particularly meaningful, but stresses the significance of gender and national origin as core dimensions of diversity. She explains this selection with reference to her experiences in a previous job at a social services department in the municipality where she was confronted with people from 'Arab countries' and from 'Turkey', who in her view held a 'different attitude' towards 'gender issues' because the 'man has a different position there'. She further elaborates: I tell you, the peculiarities and the particular, which are brought along by these people, they do not play any role [at the social services department]. (…) But I have to take this into account, when such a person comes into an office and a woman sits opposite to him, that he will evaluate her totally different than we are used to.
Here, the interview partner states that the municipality needs to change in order to adequately respond to the different cultural backgrounds of particular groups of migrants. The Department Head for Gender Equality in the same city administration argues in a similar direction, however, not regarding fundamentally different attitudes of migrants but rather with respect to their potentially insufficient qualifications: 'I think that the administration should be open to all people, especially to people with migration background, (…) but on the basis of "how can I enable them to fulfill the requirements?" and not by lowering the demands.' She underlines that it is necessary to 'give them real opportunities and offer support or whatever is necessary to enable people in their diversity so that they can carry out administrative tasks'. This quote demonstrates that not the municipality itself should change in order to become more inclusive, but rather it should provide training opportunities for people with migration background so that they can meet the organisational demands.
This overview of statements relating to the condition of assumed deviations reveals three interesting aspects: First, it shows that the diversity relevance of certain groups of people is created through the assumption of exclusively negative deviations. The 'non-traditional students' with migration backgrounds and a particular socio-economic status are allegedly lacking an adequate academic habitus. Men of specific national origins are perceived as having attitudes that differ fundamentally and migrant people in general are associated with low qualifications. On the one hand, this confirms assumptions in the existing literature about the tendency to homogenise and essentialise diversity target groups: the diversity-relevant subjects are first and foremost characterised by their migration background, which overshadows all other aspects of their individuality. On the other hand, it becomes apparent that the focus is not on 'marketable differences' but rather on characteristics assumed to negatively influence the opportunities to participate in the public organisations or society as a whole. The price for increased attention is that these groups are marked as deficient.
Second, this construction of diversity target groups is accompanied by another side effect. The high relevance of some form of deviation does not only result in a depreciation of the group members perceived as diversity-relevant but also in an affirmation of those from which they are distinguished: the traditional students with the adequate academic habitus, the people with the culturally appropriate attitude and the good standard of education. They represent the implicit but at the same moment omnipresent ideal of organisational members, who appear 'naturally' congruent with the dominant organisational culture and its corresponding structure. Here, it is interesting to note that although groups marked as diversity-relevant seem unambiguously constructed as 'the other', the interview partners articulate multiple ways of how to 'deal with them'. They mention that the organisational requirements should remain untouched and, instead, the diversity target groups enabled to fulfil them. But they also argue that the organisational cultureby means of training staff members at service providing departmentsshould change. Consequently, the burden of inclusion is not only exclusively placed on the shoulders of 'the others' but also on the organization itself.
Third, by illustrating the importance of the working areas for the prioritisation of specific diversity dimensions, I do not want to overestimate the actors' individual power to influence and shape the organisational diversity discourses. Rather, according to the discourse analytical perspective, the organisational diversity discourses allow and consist of multiple diversity understandingsalthough guided by the non-discrimination lawsand do not represent a one-dimensional stream of knowledge. As we will see below, public organisations in particular provide room for different and even opposed understandings of diversity (management).

Condition of statements: recourse to discrimination
The second condition of taking part in the organisational diversity discourses is based on the recourse to forms of negative discrimination. The main difference to references to assumed deviations is that the concern with actual or potential discriminations can both lead to extending the core diversity dimensions of the organisations as well as to an internal differentiation of each dimension. Additionally, the focus on discriminations has the effect that the significance of specific diversity dimensions is not determined in advance but rather the result of a problem-oriented approach. This is shown in the statement of the Department Head for Diversity & Equal Opportunities at Citymin B, who claims that 'people can be discriminated in various contexts because of totally different dimensions'. She underlines that the 'social background' is of particular importance: For instance, being a man is only in rare cases a ground for discrimination in contrast to being a woman. But if a man has a socially disadvantaged background, we suddenly deal with an extremely increased discrimination rate.
She further elaborates that apart from the social background also regional origin as well as educational background play key roles with regard to experiencing discriminations and that she and her team generally try to take 'multi-dimensional or multiple discriminations' into accounteven though none of this is covered by the non-discrimination laws. This dynamic concept of diversity, which focuses on the (negative) effects of certain 'diversity constellations' instead of on a priori defined dimensions, is similarly described by the Department Head for Fairness in Higher Education at Unimin. She emphasises the fact that it is necessary to adequately deal with any form of discrimination no matter what diversity dimensions might be involved: 'I don't need any list with any dimensions, rather it would be quite clear to me that if I thought that someone could be facing a disadvantage, that you take care of it.' The quote indicates a diversity concept which is adaptable to occurring discriminations and thus not restricted by a predefined set of diversity dimensions. Furthermore, such a dynamic, problem-oriented understanding of diversity allows some form of internal differentiation of the commonly homogenised diversity dimensions. For instance, the Department Head for Diversity & Equal Opportunities at Citymin B points out that people with disabilities are 'one of the most discriminated against groups', who get the worst qualification offers at the job centre: '[A]nd if you are a single mother with a disability, then it is beyond hope'. Here, we see that the interview partner describes a specific constellation of gender, dis-/ability and caring responsibilities as particularly prone to discriminations and thus as diversity-relevant. The Department Head for Social Policy at Citymin A also addresses the phenomenon of multiple discrimination and explains that 'mixed or crossed activities', which aim at groups such as 'people with disabilities and migration background' or 'women with migration background', represent an 'excellent way' how the 'diversity approach should be done'.
The statements above that refer to some form of discrimination indicate a hardly noticed angle of the 'diversity phenomenon'. They clearly show that the diversity discourse at least in public organisationsallows the coupling of diversity-related issues with the topics of discrimination and exclusion. In these cases diversity is conceptualised more context-sensitive and thereby open to additional, commonly neglected diversity dimensions such as socio-economic background or to 'unusual' diversity-relevant subjects such as men when occupying an especially vulnerable position. Moreover, the focus on actual or potential discriminations seems to enhance the visibility of individual life situations and thus reduces the likelihood that one dimension takes over a 'master status' (e.g. the group of 'women' vs. 'the single mother with disabilities').
However, this approach, which is individualising and simultaneously considering social disadvantages, may result in the stigmatisation of the (groups of) people concerned. Being marked as diversity-relevant depends on the additive linking of various grounds of discrimination, which leads to a deficit-oriented perspective on the diversity target groups constructed in this way.

Condition of statements: recourse to instrumental benefit
The third condition which allows participating in the organisational diversity discourses is the recourse to some form of instrumental benefit. This is based on the assumption that a 'diverse' team composition increases the quality of performance. This is demonstrated by a quote of the Department Head for Diversity & Equal Opportunities at Citymin B, who mentions that the 'success' of her department is based on the heterogeneity of her staff: [H]ere is everything that is conceivable as human diversity and this is a great wealth of which many new things arise (…), this does not mean that a homogeneous, supposedly homogeneous team can't be diverse or is not able to implement diversity, but it [a diverse team structure] definitely helps, if I may say, it promotes it.
Likewise but with a focus on gender, the Department Head for Quality in Higher Education at Unimin argues that he prefers a 'heterogeneous climate' in his working environment and that a 'female-dominated institution is as bad an exclusively male-dominated one'. His colleague, the Department Head for Fairness in Higher Education at Unimin points in the same direction by saying: I am in favor of mixing it, because I think that women have sometimes a different approach than men, it doesn't have to be, but it is generally said of them, and (…) I have actually always very much liked to work with teams, where both sexes are represented.
In all three cases teams or the whole organization are supposed to perform better when having a heterogeneously composed workforce. However, the interview partners conceptualise the heterogeneity differently: The statement of the Department Head of Fairness in Higher Education indicates that she assumes that women and men are essentially different. The quote of the Department Head for Diversity & Equal Opportunities reveals a more constructivist understanding: she uses the phrase that a 'supposedly homogeneous team' might not be able to perform diversity-related work as well as a heterogeneous one. Here, we see that diversity is not understood as a fixed characteristic of a person, group or setting but rather conceptualised as dependent on which diversity dimensions are perceived as relevant in a specific context. This could imply that the diversity discourse is not only inherently based on an essentialised idea of diversity but also enables the articulation of different, more constructivist and relational understandings.
In addition to linking a diverse workforce with an increased quality of performance, the data exhibit another stream of statements related to the occurrence of instrumental benefits. Here, diversity (management) is characterised as an additional value for all groups marked as diversity-relevant. For instance, the Department Head for Social Policy at Citymin A outlines that the municipality had been 'pretty advanced' concerning individual activities for various disadvantaged groups, and with the diversity concept they tried to develop a concept, which 'brought them together': 'We found out (…) that (…) all groups have common interests, it is about participation, it is about antidiscrimination, it is about visibility but not being "hyper-visible" and so on (…).' Likewise, her colleague, a Staff Member at the Personnel Department, explains that diversity represents a concept, which 'comprehensively integrates' the individual areas of activities. The Department Head for Gender Equality at Citymin B also points to the shared interests of different disadvantaged groups and emphasises the strategic factor of relying on the diversity concept within the municipality: [R]egarding the resource of social recognition, significance, so to speak, ranking in the political agenda, (…) these are points, which (…) play an important role for all of us, because whether gender or migration or disability, we always have a marginalized position within the municipality (…). I see great opportunities in the diversity process so that we (…) can achieve more together than alone, that is for sure. This quote illustrates clearly that the diversity concept and its related practices are associated with the benefit of lobbying more assertively for different disadvantaged groups. In both cases, with respect to the increased performance quality or the enhanced representation of interests, the recourse on a potential benefit of diversity (management) is not coupled with economic exploitation but rather bound to an 'end in itself' for the diversity-relevant target groups. This means that in public organisations as well, the articulation of some form of instrumental benefit of diversity (management) plays an important role, but is not underpinned by a business logic as such.
At this point, it is important to note that the just outlined benefit of diversity (management) is a relatively contested statement pattern. For instance, the Department Head for Gender Equality at Citymin B argues also in the opposite way: I see it rather as a threat for gender to get lost in it [diversity] (…). We are, so to say, relatively strong regarding staff, (…) but we don't have what diversity brings along, such a factor of the new and the hip, we don't have that regarding gender, thus we are competing with it [diversity] for social attention and significance, and (…) I think we need to make sure that we present each other together as well as separately in order not to get lost.
Likewise the Department Head for Gender Equality at Unimin expresses the concern 'that the equality between men and women (…) could be pushed into the background because of the huge task of diversity.' However, she also states that it is necessary to 'try to reasonably interlock both areas', keeping in mind that gender equality is still not achieved in many areas of life. In contrast, the Staff Member of the Department for Integration at Citymin B points the other way and formulates: [A]part from that I don't understand the logic why the one thing [diversity] would sweep the other [gender] under the carpet, one of my main criticisms was and is that 'the genderwomen' have not yet paid attention to female migrants, but always ignore them and look at everything from a white perspective.
All three quotesalthough with a different contentrefer to the relationship between diversity and gender and reveal that whether the assumed benefit of diversity (management) for several disadvantaged groups occurs, is highly contested. On the one hand, already existing gender equality policies are portrayed as in danger of losing attention and other resources because of the 'diversity turn'. On the other hand, the dominant position of gender in comparison to other diversity dimensions is strongly questioned because one-dimensional gender policies can (re-)produce exclusions themselves. By interpreting this disputed topic from a discourse analytical perspective, it is not really significant whether the interview partners articulate contradictory statements. Rather, it is interesting that the diversity discourses in public organisations open up a heterogeneous space which may include different views that do not necessarily have to be reconciled.

Conclusion
The conditions of statements for intelligibly participating in the organisational diversity discourses represent neither an exhaustive list nor do they reflect actual discursive practices in the organisations under investigation. Rather, they result from a (re-)construction work (Diaz-Bone 2006), which aimed to expose the regularities by which the diversity discourse in public organisations functions. The conditions are not mutually exclusive and thus the organisational members can draw on various combinations in order to take part in the organisational diversity discourses. The effect is a heterogeneous 'discursive space', which is on the one hand characterised by references to some form of instrumental benefit alongside motives of marketability (see also Tomlinson and Schwabenland 2010). On the other hand, the available 'discursive space' isto a larger proportiondefined by the recourse on deviations. Either groups of people are marked as diversity-relevant because of an assumed deficit in skills and qualifications, or they are perceived as discriminated against and thus vulnerable. The context-sensitive focus on the vulnerability of certain (groups of) people can lead to a multi-dimensional and more individualised construction of diversity target groups but is simultaneously accompanied by the labelling as 'particularly different'.
In this context, I would like to stress one aspect that seems especially noteworthy: The epistemological foundation of the attributed differences is not necessarily based on positivist assumptions of 'naturally' given, embodied and fixed characteristics as indicated by some studies (e.g. Litvin 1997;Zanoni and Janssens 2004). Although the data show that essentialised understandings of diversity are prevalent (e.g. the perception of women and men as fundamentally different), it also exhibits alternative interpretations, which describe diversity as a relational and dynamic phenomenon (e.g. the 'supposedly homogeneous team'). Furthermore, the linking of diversity with actual or potential discriminations brings questions of causes of and solutions to social exclusion to the fore instead of focusing only at the individual dimension of diversity issues. Against this background, it can be concluded that the diversity discourse is neither inherently based on ideas of essentialised differences nor intrinsically opposed to addressing issues of social disadvantage and exclusion. Rather, it is more likely that the organisational context including the key tasks, the composition of the internal workforce as well as external target groups and the tradition of equal opportunity policies are highly relevant for the quality and limits of the actual, diversity-related discursive practices.
Other scholars (Dobbin, Kim, and Kalev 2011;Janssens and Zanoni 2005;Tomlinson and Schwabenland 2010;Zanoni et al. 2010) have already emphasised the organisationally contingent nature of diversity management. However, the context of public organisations allows some specific insights contributing to a more differentiated picture of the diversity discourse. First, it demonstrates that the legal and national contexts are of great significance for the diversity-related policies in public organisations. All of them rely on European and national non-discrimination laws in order to define their understanding of diversity and use it as a common reference framework. However, none of the organisations under investigation limits its actual diversity policies to the six diversity dimensions (race/ethnicity, gender, religion/belief, disability, age and sexual orientation). Rather, they follow a problem-oriented approach. This results in a flexibilization of which (constellations of) dimensions become relevant when dealing with diversity-related issues. For instance, the organisations identify socio-economic status as well as educational backgroundtwo dimensions that are commonly neglected in for-profit organisations (Plummer 2003)as essential for the opportunities to participate in their services or society as a whole. This indicates that organisations build on and develop knowledge about social exclusion or structural inequalities that is not covered by the mainstream business-oriented diversity discourse, but might simultaneously feed back into the overall, macro-level diversity discourse and thus slightly expand its canon of knowledge. With respect to the national context it is important to note that in Austria and Germany, where the study was conducted, gender equality policies hold a special position within public institutions (e.g. affirmative action and other legal regulations) (see also Ferree 2008). This is also depicted in the data, which does not only point to a competitive relationship between issues labelled as gender-or diversity-relevant but also between the topic of gender and 'old questions' of migration/ethnicity. The statement by the Staff Member of the Department for Integration at Citymin B illustrates this fact by emphasising the ignorance of traditional women's politics towards concerns of migrant women. The issue here is not whether this perception is true or false, ratherfrom a discourse analytical perspectiveit is interesting that the diversity discourse encounters an already existing and institutionally well-established gender equality discourse in public organisations. Against this background, the question arises whether the diagnosis (Ahmed 2012;Bell and Hartmann 2007), that the diversity discourse functions as a potential silencer for the articulation of racial inequalities in the Anglo-American region, is transferable to the Austrian and German context. In contrast, the data indicates that the diversity discourse could be a door opener for diversity-relevant concerns in addition to already addressed gender issues. It is unlikely that the diversity discourse simply replaces the gender equality discourse, but rather that they coexist in a more or less competitive relationship and thereby almost inevitably exchange and appropriate their underlying knowledges. As a possible result the diversity discourse could gain complexity and thereby adopt and extend its conditions of intelligible participation.
This assumption of a constantly changing and thus disputed diversity discourse relates to the second aspect I want to emphasise with respect to the specific nature of public organisations. In fact, public organisations represent particularly contested spaces for diversity-relevant issues. The reason might be, quite simply, that they are often very large organisations with departments on several hierarchy levels and equipped with different competences. These departments and their individual responsibilities (e.g. gender equality, disabled people, people with migration background) have grown over decades and simultaneously established a specific body of knowledge (see also Kamp and Hagedorn-Rasmussen 2004). As a consequence, the entering of the diversity discourse into public organisations is not to be imagined as a unilateral top-down process, but as the establishment of a widely branched net that is constantly (re-)produced and altered by different, discursively powerful actors.
In the light of these findings, how can we answer the research question concerning the recognition or denial of different claims for inclusion? The investigation of public organisations that describe themselves as diversity affine reveals that pointing to discriminations and societal inequalities is not intrinsically opposed to a diversity discourse. Public organisations recognise multiple claims for inclusion, although mostly based on an essentialised and deficit-oriented construction of diversity target groups. However, this dilemma of recognising and simultaneously affirming differences does not represent a unique feature of the 'diversity phenomenon' but also applies to most other identity-based discourses (see, e.g. Alvesson and Billing 2002). Possibly, it is distinctive of the diversity discourse that it has a potential for inclusiveness of various differences and is adaptable to knowledge and practices already in place in specific contexts. This may be viewed either as further proof of the fundamental arbitrariness of the diversity concept itself or as a sign for its contested and thus 'open-ended' nature. Vertovec's (2012, 307) assumption that through 'the extension and reproduction of "diversity" discourses, policies and programmes, it could be that people are increasingly able to think about and perceive society in terms of more, if not more complex, social categories' is not yet proven right or wrong, but it definitely represents one of the paths the diversity discourse could take. Since this study did not investigate diversity-related practices in actu, but rather re-constructed them with the help of interviews, future research, which is more ethnographic and also involved in cross-country comparisons, could further contribute to a more fine-grained picture of the inclusionary potential of the 'diversity turn'.