UvA-DARE ( Digital Academic Repository ) Playing language games : higher education quality dynamics in Dutch national policies since 1985

Higher education quality is a vague, ambiguous, multiple, and essentially contested concept. Quality’s contested character involves endless disputes about its proper use which makes it problematic to handle in governmental policies. Wittgenstein’s notion of language games is used to understand how, through time, higher education quality is enacted in Dutch governmental policy texts, and how its uses are related to each other. The analysis depicts various quality games interacting with different policy contexts, which show multiple enactments of quality as a unified concept alongside more differentiated uses. In the policy games, quality is not the focal notion. The games center around the steering relationship with the institutions, which are placed ‘at distance’. Through time, the games respond to increasing societal complexity and competition, and foster further flexibilization of institutional policies regarding quality and accessibility. In this management discourse with the institutions, quality is not used contrastively. It is concluded that quality’s vague and contradictory enactments and valuations are not problematic in the institutional steering relationship. Recent policy texts however relate quality’s ‘proper use’ to activities that enhance the student’s learning process. This draws attention to paradoxes for a distancing government in its role as a universal actor with societal responsibilities.


Introduction
'What the hell is quality?' researcher and policy advisor Christopher Ball asked in 1985 in a much-cited essay (Ball 1985). When Ball raised his question, front-runners like the Netherlands and the United Kingdom developed a quality framework for higher education. Thirty years of policies and analysis have not provided a univocal answer to Ball's question. In the 1990s, research showed that formal quality conceptions did not match situated meanings held by educational professionals. Quality became conceived as elusive, vague, ambiguous, multifaceted, and without an essential core. It is what Gallie (1956) calls 'an essentially contested concept', which involves endless disputes about its proper use by its different users (Harvey and Green 1993;Harvey and Newton 2007;Harvey and Williams 2010;Lips 1996;Newton 2010;Wittek and Kvernbekk 2011).
While academics extensively debated the higher education quality concept in the 1990s, the question 'what quality is' is currently understudied (Stensaker 2007;Westerheijden, Stensaker, and Rosa 2007;Wittek and Kvernbekk 2011). Quality's multiplicity, multidimensionality, and subjectivity are broadly accepted (Brockerhoff, Huisman and Laufer 2015;Harvey and Green 1993), as well as the conception that it can be classified by the way it is used following different rationales. The question 'what quality is' however remains apt as it is still also used as a unified concept (Wittek and Kvernbekk 2011). There is now a variety of discourses which use and define quality, and it is questioned whether all enactments can be seen as quality. Following Wittgenstein's notion of 'family resemblances', multiple and contradictory quality definitions and enactments can all be seen as quality. These are part of the same quality family, and family members do not need to resemble each other in their features in order to be part of this family (Wittek and Kvernbekk 2011;Wittgenstein 1953).
The contested character of quality is still present in daily practices, as it is interrelated with numerous enactments and valuations by varying family members operating at different social levels. This renders higher education quality a complex and ill-defined social problem which cannot be easily addressed using traditional problem-solving methods, especially at the macro-societal level (Krause 2012;Westerheijden, Stensaker, and Rosa 2007). Moreover, the discussion about its proper use remains, and Harvey and Newton currently argue that classifications such as 'quality as fitness for purpose' are empty categorizations without conceptual gravitas. Instead, a Marxist reconstitution of the quality concept is proposed that evolves around its essential goodness (Harvey and Newton 2007). Such a reconstitution however will not work, as the problem with contested concepts is that this essential character remains disputed, and differences cannot be finally settled. Even when one quality perspective becomes morally and socially dominant, there are always other perspectives, and we do not have pre-given standards to value which quality is the true one. To use an essentially contested concept is 'to use it against other uses and to recognize that one's own use has to be maintained against these other uses' (Gallie 1956, 172). To understand how the quality concept relates to wider social and political contexts, we need to study how it operates and functions as a prominent concept in decision processes, especially the macro-oriented policy process.
Wittgenstein's notion of language games is used to understand how higher education quality is enacted and used in governmental policies, while interacting with different actors and policy contexts. The notion of 'language games' connects 'family resemblances' with social constructions of reality and assumes that language is woven into action. Constructivist approaches acknowledge that the reality we know is interpreted, constructed, enacted, and maintained through language. In a constructivist discourse, language itself is a construction, and different language games will give multiple constructions, understandings, and assessments of higher education quality in interaction with its contexts. Language games enable us to study quality is in all its vague, ambiguous, and contradictory uses (Berger and Luckmann 1966;Wittgenstein 1953). To understand its uses, we pattern how higher education quality is 'played' in Dutch national policy plans published since 1985. The following research question directs the study: How and where are quality differences and unity created in governmental politics in interaction with changing actors and social contexts, and what does this mean for the policy process relating to higher education?
The patterning of language games however not automatically involves an understanding of quality's uses.The theory section 'Language games as theoretical and practical perspective' therefore addresses the possibilities and limitations of the language games perspective, and introduces Hall's work on 'articulation' to understand how and when quality is enacted as a contested concept in social formations. We then describe the case study on quality in Dutch policy texts through time. The "Results" section analyzes governmental policy documents on higher education quality and steering relations published between 1985 and 2015. This section ends with a wrapup of differences and unity in the quality games and is followed by a discussion on what its different uses mean for the policy process and further research directions.

Language games as theoretical and practical perspective
Quality's contested, vague, ambiguous, and elusive character invites research approaches that do not aim to define, categorize, or substitute its meanings. The full political consequence of a contested identity is that it is a 'constructed identity' which cannot be grounded in any category, and therefore has no guarantees in nature (Hall 1996, 443;Gallie 1956;Wittek and Kvernbekk 2011;Yanow 2003). The paradox is, however, that such approaches hinder its operationalization. Current quality studies therefore address its vagueness but do not operationalize it (Giroux 2006). This section introduces and discusses Wittgenstein's notion of language games to elucidate how quality differences and unity are enacted in interrelation with different contexts, without catching or narrowing down quality's meanings. Categorizations and typologies can thereby provide a useful starting point to study quality enactments, if they are used as sensitizing concepts. This section ends with a description of how sensitizing concepts can provide directions about where to look, without using them as definitive categories (Blumer 1954).

Introducing language games
Wittgenstein introduced language games in his lectures published as the Blue Book and further developed the concept in his Philosophical Investigations (Wittgenstein 1953(Wittgenstein , 1958. Central to the idea of language games is that we lay down rules and techniques for a game. These games develop as abridgements of practices. When we follow these rules, things do not turn out as we assumed however (Mouffe 2000;Stern 2004). Wittgenstein notes that we are entangled in our own rules, and this entanglement is what we want to understand (Wittgenstein 1953, 50e).
According to Stern (2004), it is essential to this method that attention is drawn to the context in which our use of language takes place, highlighting the state of affairs before contradictions are resolved. The term language game is applied to almost any action in which language is involved in some way, any interweaving of human life and language; 'I shall call the whole, consisting of language and actions into which it is woven, a language-game' (Wittgenstein 1953, 5). This makes the number of possible language games countless. Wittgenstein notes that there are countless different kinds of use of what we call "symbols", "words", and "sentences". This multiplicity is not something fixed, given once for all. New types of language and new language games, come into existence, and others become obsolete and get forgotten (Wittgenstein 1953, n).
Language itself is multiple, and this enables the interpretation of plural and polysemous concepts like higher education quality in different uses and contexts. The notion of language games does not make the difficulties in interpreting vague concepts disappear. Whereas context can select one of the ambiguous meanings, their relation with their contexts is more complicated for vague concepts. Vague expressions can be context-dependent, but context does not provide clear conceptual boundaries (Giroux 2006;Keefe 2000;Wittek and Kvernbekk 2011).
Language games do, however, allow sensemaking to understand vague concepts in practice. We do not need sharp boundaries to do something with a concept.
Is it senseless to say: 'Stand roughly there'? Suppose that I were standing with someone in a city square and said that. As I say it I do not draw any kind of boundary, but perhaps point with my handas if I were indicating a particular spot. And this is just how one might explain to someone what a game is. One gives examples and intends them to be taken in a particular way. (Wittgenstein 1953, 34e) Giving examples is not an indirect way of explaining or defining, but an expression of the game. This is how games are played, and understanding is conveyed in language games. Patterning higher education quality games will not provide clarity or definition, but will help to elucidate how quality is enacted and interrelates with its changing contexts (Mauws and Phillips 1995).

Possibilities and limitations of language games
Although the notion of language games has been used to study organizational practices and strategizing (Mantere 2010), it has not yet been applied to the study of quality in higher education. There is, however, a wide range of textual and discourse analytic approaches to higher education quality. Several studies depict discursive struggles over meaning, whereby Foucauldian and Critical Discourse Analytic perspectives tend to dominate (Morley 2003;Saarinen 2008;Vidovich 2001). Vidovich (2001), for instance, argues in a longitudinal analysis of Australian policies that quality is chameleonlike and tends to meld into its contexts. This enables the government to launch institutional reforms. Others are more critical toward the 'captive powers of discourse', and note that academics apply different repertoires to negotiate their own position in institutions (Ball 1993;Trowler 2001). These agency-oriented analyses however do not specifically problematize the quality concept.
Several studies analyze how quality is translated as a management concept (Blanco Ramírez 2014;Giroux 2006;Stensaker 2007). To enable broad dissemination, it has to lend itself to various interpretations, and each party has to recognize itself in its own version of the concept. This equivocality allows different courses of action while maintaining a semblance of unity (Giroux 2006). Such studies however focus on traceable translations and do not operationalize its vague enactments in the open social domain.
With respect to different translation and discourse analyses, the language games perspective is more open to the unlimited range of vague and contradictory higher education quality enactments. It asks for an interpretive policy approach which focuses on situationspecific meanings, and enables the study of the interplay between language uses with a variety of contexts. Being rooted in practices, language games allow us to understand how and where quality unity and differences are enacted, situated, and strategized in governmental policies (Yanow 2007;Astley and Zammuto 1992;Mantere 2010).
The language games perspective however runs the risk of losing analytical power since its toolbox comes without preset rules or limitations to conduct the social analysis. Its openness can lead to Lyotard's conception of society as consisting in a plurality of incommensurable language games. Anything then seems to go, which makes moral and political rearticulations impossible. Furthermore, if the practice perspective is not taken seriously, it reduces reality to a spectacle of what is immediately there on the surface (Laclau and Mouffe 2014;Grossberg 1996).
In line with Hall, we therefore study the practices of 'articulation' to understand how the different contested quality uses are maintained and played in relation to each other. Articulation literally means to utter, 'to speak forth'. It is however also a temporal linkage, whereby two different elements like 'quality as essence' and 'quality as differentiated' are connected. Processes of articulation involve social formations, whereby things are related through their similarities as well as their differences. As Hall notes, the practices itself do not necessary lead to political articulations. There will always be language games which do not connect or confront different quality uses (Slack 1996;Grossberg 1996).
Hall's notion of articulation is used to further clarify how the games are played, and when quality is articulated as a contested concept. Wittgenstein was wary of what we would now call cognitivist interpretations of mental worlds, which originate meaning making in private, inaccessible minds (Potter 2001). As a critical Marxist, Hall differs from this perspective. To put it overly simple, Hall integrates identifications in the processes of sensemaking, which he understands as processes of struggle while interrelating with underlying structuring schemes. Contrary to earlier Marxist theorist, Hall however does not fall in the pitfalls of essentialism (Grossberg 1996). Hall thereby uses the games analogy as brought forward by the linguist Saussure. While both Saussure and Wittgenstein farewell the perspective that a word stands for its meaning, Saussure understands linguistic signs and language itself as biplanar (Harris 1988, 14).
In practice, Wittgenstein's language games focus on directions and sensemaking in the (inter-)act, whereas Hall's games relate with underlying structures and identifications the actors bring forward in interaction. It enables a better understanding of its contestation as it also addresses underlying identifications. To gain a full understanding of how quality is enacted in governmental policies, we first interpret the policy texts following Wittgenstein's language games perspective. The conclusion then specifically addresses different articulations and contrastive uses.

Interpreting the policy texts
The aim of the study is to elucidate where and how differences and unity in higher education quality are enacted in governmental politics, and to understand what its contested, vague, and ambiguous character means for the policy process. Wittgenstein's notion of language games is used to interpret 12 Dutch national higher education policy texts since 1985, and pattern how these games are played and change through time.
The interpretation concerns the 1985 governmental white paper Higher Education Autonomy and Quality (Hoger Onderwijs Autonomie en Kwaliteit, or HOAK) and subsequent strategic planning documents. Since 1988, 11 generic planning documents have been published. They all address higher education and research, but the 2015 Strategic Agenda focuses on education. 1 The formal author is the Dutch national government, represented by the Ministry of Education and Science. The documents were first named Higher Education Research Plan (Hoger Onderwijs Onderzoeks Plan, or HOOP) and in 2007 renamed Strategic Agenda (Strategische Agenda Hoger Onderwijs en Onderzoek).
We analyzed the final drafts of the documents, which have been sent to the Parliament. The documents are the result of governmental interactions and deliberations with various and changing actors, and aim to provoke planned change. The collectively written policy texts are carefully constructed. While they provide directions to the planning process, they remain vague and open to future deliberations (Czarniawska, 1997).
The content and form of the plans change through time, and they range from 200 to over 600 pages. Some plans are published with addenda containing indicators and statistics. As these indicators and statistics are expressions of language games, no selections have been made in the texts. Both the core documents and their addenda are interpreted.
The constant comparison of language games forms the methodological basis of the interpretation. Language games can take the form of new strategies relating to solving policy problems, giving new orders, using statistical data, and countless other acts. They range from simple, well-defined acts, such as the use of predefined quality indicators in argumentation, to abstract meta-levels, such as providing vague indications. The texts are therefore constantly compared on different, interrelating units of analysis, ranging from single utterances to the whole body of texts. Quality classifications function as labels to code the texts and provide a starting point in the identification of quality games. Interpreting language games implies identifying where in the policy texts quality is situated, who is concerned, and which courses of actions are proposed to solve the indeterminacy. The analysis was computer-assisted using the Atlas-TI program.

Quality classifications as sensitizing concepts
Quality studies identify and contribute to the categorizing and modeling of the higher education quality concept from different traditions, spanning management, and educational perspectives (Brockerhoff, Huisman, and Laufer 2015;Giroux 2006;Stensaker 2007). We use Harvey and Green (1993) as sensitizing concepts. Though these categories have later been identified as empty, they provide a starting point to identify different rationales and practices in the games. Harvey and Green have raised awareness for quality as a relative concept, in the sense that it is subjective and has to be compared or valued against standards (Brockerhoff, Huisman, and Laufer 2015;Harvey and Green 1993;Harvey and Williams 2010); • Quality as fitness for purpose relates to the purpose of a product or service and is judged from this perspective. Any product is good if it serves its purpose. Students' education, for instance, should match the requirements of work-life. This perspective is dynamic because purposes can change. • Quality as value for money is difficult to discern from fitness for purpose. It is explicitly linked to economic motivations and the measurement of quality in terms of profit and effectiveness. • Quality as excellence encompasses two different notions: ○ Quality as exception is quite remote from quality as fitness for purpose, as it is distinctive and elitist, and by definition exclusive. It requires constant modification of elitist standards to keep the distinction. ○ Quality as perfection or consistency focuses on specifications for how quality is to be striven for in every part of a process. The result depends on the quality culture. The focus is on the process, and it is characterized by checklists and procedures. • Quality as transformation relates to the transformative process that students go through and is often addressed as Bildung. The transformative process can be of higher or lower quality.

Results
The HOAK paper: fostering autonomy and quality In the 1985 HOAK paper, higher education quality is primarily played in two interrelated ways: as a central notion in the design of changing steering relations and as a new system of quality assurance. The gist of the HOAK paper is that current steering relations are no longer effective in adequately advancing the quality of higher education in a complex and changing society. 'The dynamics and unpredictability of social and scientific change call for the reduction of uniforming and centralized procedures to the absolute minimum' (MinEd 1985, 9). The proposed solution is increased institutional autonomy. This enables systems dynamization, which fosters flexibility and quality.
In this reasoning, the quality concept is paired with the concept of institutional autonomy. 'The paired concept of "autonomy and quality" fulfils a central role in this paper: it indicates policy directions' (MinEd 1985, 10). The Educational Council of the Netherlands remarks on the draft HOAK paper that there is no logical and evident relationship between those two concepts and that it cannot be simply assumed that increased autonomy will automatically lead to quality improvement. The government responds in the final paper that quality increase is not a reality that automatically comes with growing institutional space for policymaking, but that it forms a necessary precondition thereto. The realization of quality therefore becomes the primary responsibility of the institutions. They have to use the increased room for policy maneuver (MinEd 1985). The central position of the paired concept was not further questioned, and the paper was accepted without much political discussion. The government, the institutions, and their umbrella bodies were convinced that a new relationship between government and institutions was needed. No further argument was necessary, for who could be against more autonomy (Lips 1996;Mertens 2011)? Although quality and autonomy fulfill a central role in the HOAK paper, the elaboration focuses on the how of changing steering relations between the national government and institutions. The planning system is to change from detailed governmental prescriptions to institutional accountability. The government thereby strives for a situation of distanced and global steering, whereby the institutions themselves are responsible for their policies and interact with different societal subsystems. It is noted that this increases the space for institutional profiling and differentiation. The intent is to put more effort into communication and interaction in the steering relationship with the institutions.
The HOAK paper introduces two new planning documents to support the interaction between the government and the publically funded institutions. In their Development Plans, the institutions should formulate how they plan to interact with societal changes and respond to the governmental Higher Education Plan. In turn, this Higher Education Plan (which would become the HOOP) should interact with the development plans and contain the governmental vision on the higher education system.
The introduction of a good system of quality assurance is found essential to the functioning of the changing steering conceptions. The institutions and their umbrella bodies are to organize a system of quality assurance, and the inspectorate will have an additional, evaluative role.
To summarize, together with (institutional) autonomy, quality provides directions to new policies laid down in the HOAK paper. It supports new steering policies, whereby the distancing of institutions improves their flexibility in interacting with a complex and changing society. Higher education quality is thereby treated as an abstract though unified object, realized when the institutions use the improved space for policy maneuver.

1988-1992: coming into play
In the first draft HOOP, educational quality is not a central notion. Its introduction focuses on the role of the government and the HOOPs as communicative documents in the process of realizing new steering relations and related planning systems. The HOOPs should bring cohesion in governmental communication with the institutions and other parties.
What is written in the HOOP should be substantiated and motivated, and therefore legitimate, but not compelling […].
[…] an exchange of insights and arguments is needed, of meanings and attitudes, of intentions and foresights. We call this exchange dialogue. (MinEd 1987, 9-10) Dialogue is a central concept in the first three HOOPs. It is considered necessary to enable responsible decision-making by both government and institutions. The proposed dialogue is formal and procedural, and conducted by exchange of written documents as well as through deliberations with the institutions, their representative bodies, and other parties. Like in the HOAK paper, quality is supposed to be the result of actions performed primarily by the institutions, guided by governmental perspectives on developments in different societal subsystems. With the absence of quality as a central notion, this reasoning is, however, less explicit than in the HOAK paper.
When quality is addressed in the strategic sections, it is treated as a relative though unified concept, which can be compared with other countries' performances. The Dutch quality highlands are valued positively in comparison to more strongly proliferated institutional differences in the United States. Policies that foster excellence and quality peaks should not endanger this high quality standard (MinEd 1987).
In the strategic section of the first HOOP, the notion of excellence interrelates with the more prominent notion of quality as fitness for purpose. This labor-oriented perspective on higher education quality is also highlighted in the planning sections. Macro-societal developments interrelate with the functioning of the current educational system and scenarios for the supply of, and demand for, higher-educated people on the national labor market. Statistics for example address scenarios for supply and demand in specific sectors. In the dialogue, the government focuses on the planning of educational subsystems, not on individual institutional interactions.
Although we can identify sensitizing quality concepts like fitness for purpose and excellence, quality is not explicitly named or suggested in the texts. National labor market statistics and higher education statistics, for example, are not understood as quality indicators. The sensitizing concepts relate to higher education in general, not specifically to quality.
'Quality assurance implies an explicit opinion on what is understood to be the quality of education or research' (MinEd 1989, 319). In the second HOOP, it is noted that there are various perspectives on the scientific, social, and individual features of graduates, as well as on the meaningfulness and efficiency of the educational process. Several parties with varying interests and values mingle in the debate and are entitled to do so. Hence, a varied set of instruments is needed to gather and value the ordeals of students, educational professionals, graduates, and customers. The discussion on what is understood as quality is advanced further when the concepts used are specified. Operationalizing these concepts clarifies where value orientations diverge and can have the effect of the conversation becoming more pragmatic (MinEd 1989, 319).
The intent is to formulate valuable performance indicators with the institutions, and come to a shared language. This can have a disciplining effect on interactions with the institutions, if those involved agree on which indicators are valuable and acceptable (MinEd 1989, 319-320). A quality dialogue that makes the different value orientations explicit is found necessary to come to these performance indicators and serves the interaction with the institutions. This quality dialogue, however, is not played out in the HOOPs.
The first HOOPs exemplify a quest for how the planning and steering game is practiced in interaction with the institutions. In doing so, quality is played in at least three ways. The strategic policy game treats quality as a unified though relative object, which can be valued against other countries' performances. In the planning game, quality is not specifically addressed. We can however recognize several sensitizing concepts, whereby fitness for purpose predominates. Finally, the quality assurance game treats quality as a differentiated concept, whereby different value orientations and quality conceptions should be addressed in a government-led dialogue and shared language.

1994-1998: playing differentiation and selectivity
The strength of nations in social, economic, and cultural domains is increasingly determined by the capital that well-educated people represent. To build and maintain this human capital, higher education is of crucial importance. Higher education is the place where talent is fully developed, where young people are challenged to make the best of themselves, where researchers achieve better results by being challenged by their students. Excellent institutions are essential for the prosperity and welfare of our country. (MinEd 1995, 3) These first sentences of the 1996 draft HOOP exemplify how higher education and higher education quality are played in the introductory parts of the HOOPs. The current society and its relevant developments are sketched and related to desired future educational developments. We can recognize the idea of quality as fitness for purpose. Higher education creates human capital, which increasingly determines the social, economic, and cultural development of the nation. This quality-as-fitness-for-purpose frame interrelates with quality as excellence, because excellent institutions are crucial to Dutch prosperity and welfare. It is even possible to identify connotations relating to Bildung and individual talent development. Just as in the first HOOPs, quality is not explicitly named or suggested here. The sensitizing concepts are again related to higher education in general, not specifically to quality.
This game in the 1996 HOOP is vague and conceptual. The introductory sections are airier and more evocative than the labor-oriented notions in the first HOOPs and no longer contain labor market forecasts. The quality conceptualizations relate to generic social effects for higher education and research at an abstract level. As the earlier quotation illustrates, the level of conceptualization is so generic that the different notions of quality such as fitness for purpose, excellence, and Bildung do not clash with one another. The 1996 HOOP thereby practices what the HOAK memo preaches and leaves the concrete dealing with complexity issues to the institutions.
These HOOPs draw extensively on the frame that institutional differentiation is needed to deal with the balancing of different goals in a complex society. This draft HOOP sketches how the accessibility, quality, and affordability of higher education will be balanced in this decade. With the observation that […] first and foremost increased differentiation and selectivity are needed, this draft HOOP continues the path of the 1994 HOOP. (MinEd 1995, 10) The situated meanings of differentiation, selectivity, as well as quality, accessibility, and affordability, are expressed in relation to one another. Quality, and especially the quality assurance system, is valued positively in comparison with other countries. A perceived downside is that the system is not sufficiently differentiated to meet the diverse needs of the increasing number of students. There has to be more focus on student selection, and smaller universities should provide room for tailor-made education. The institutions should be more flexible in the types of studies they offer.
The 1994 HOOP is the latest document to address dialogue specifically. This then disappears, together with the idea that specification of the concept advances quality discussions. The 1996 HOOP builds on the governmental coalition agreement, and the quality system is successfully implemented without further reference to shared quality indicators or situated meanings of the quality concept. These changes are accompanied by a changing relationship with the institutions. Following successful implementation of the quality system, the focus shifts toward monitoring the outcomes of visitations (MinEd 1993(MinEd , 1099.
The second half of the 1990s can therefore be characterized as a period in which evocative and strategic language games become more prominent. In these strategic policy games, quality is contrasted with efficiency and accessibility, a tension that should be solved by the institutions balancing different needs. There is, however, less emphasis on the rules of the games and on how quality should be played in dialogue.
With the successful implementation of the quality system, this systems game becomes more procedural.

2000-2004: changing contexts, practices, and systems
In the first years of this century, there is a prolonged tendency for policy texts to become more strategic and evocative. The HOOPs and Strategic Agendas display an incremental development from detailed planning documents to strategic documents with an increased emphasis on social effects. It is difficult, however, to connect social effects with policy measures, and to decide what exactly contributes to these changing policy practices (MinFin 2004).
The 2004 HOOP identifies three social trends of (1) transition to a knowledge society, (2) growing importance of Europe and internationalization, and (3) increasing societal complexity. There is a shared ambition to belong to the top international knowledge economies in 2010, and the reasoning is that institutions have to change to improve and foster the transition toward a knowledge society. Increasing societal complexity means that education should be oriented toward societal demands and enable students to function in this society. More higher-educated people are needed to prepare for a future complexity. Policy solutions are maximal participation, increased efficiency, and challenging education (MinEd 2004). These notions can again be related back to accessibility, efficiency, and quality, and have to be provided by strong and flexible institutions.
In these conceptual and airy language games, the notion of excellence becomes more prominent, although not predominant. Institutions are given more space to offer more than basic quality, for example through individually customized education, international experience, or special programs for the talented.
The context changes from national in the early HOOPs to EU-regional in the mid-1990s and international around the turn of the century. With the start of the Bologna process in 2000, the policy texts relate to European processes on harmonization and the position of Dutch education in comparison with other countries. In the 2004 HOOP, the knowledge society is the main context.
The policy texts again express the idea that institutions have to provide flexible solutions in response to the increasing complexity of the knowledge society. They should cooperate in networks, retain maximum autonomy, and be more distinctive and profiled. The changing societal context is, however, also used to legitimize shifting quality assurance conceptions. In the 2004 HOOP, the steering relationship with the institutions is again a central notion. The rules of the steering game are changed, and the government delineates its own role in relation to the institutions. The institutions are denoted as organizations with a societal task, with not only a vertical but also a horizontal accountability relationship with their direct environment.
Performance agreements are introduced in 2004 as a key concept in the development of steering relations. To bring the worlds of politics and practices closer to each other, shared ambitions are to be formulated, and institutions are asked to deliver a contribution based on their own profile. The formulating is tacit and indirect here, and performance agreements are not being used to measure individual institutional quality. It is stressed that formulating is an incremental process that is to be conducted with care and in close consultation with the institutions and their umbrella bodies.
The institutions are meant to use collective indicators and ambitions to formulate their own policy goals in relation to the government and their environment. Quantitative benchmarks are part of the governmental publication Mapping Knowledge, which also contains indicators relating to governmental goals. The quality indicators are student/staff ratio and the distribution of scores at visitations. Other indicators relate to accessibility and efficiency.
In the first years of this century, the quality assurance system changes into an internationally comparable accreditation system. This systems change does not receive much attention however. The 2000 HOOP notes that the quality system is good, but that it can be further improved. The 2004 HOOP sees accreditation as an instrument to improve the European comparability of the Dutch higher educational system and mentions an accreditation agreement with Flanders (MinEd 2000;MinFin 2004).
The first years of this century show a growing importance of the international context. This context remains at a distance however, and national changes in steering conceptions are foregrounded. The international context and complexity tend to be used as abstract entities that legitimize changes in steering relations with the institutions.

2007-2011: average is not good enough
In the 2011 Strategic Agenda, the bar is raised to prepare students for a more demanding future.
In 2025, the study culture at colleges and universities is characterized by challenges, achievement, and making the most of one's own abilities. The bar is raised, and the student who cannot jump over it will have to adjust his ambitions. (MinEd 2011, 8) Whereas the first HOOPs value the quality plateau of the Dutch highlands, the last decade shows a tendency toward excellence, further differentiation, and quality peaks. Excellence is primarily used to make distinctions between institutions and further differentiate students. The government aspires to a leading position amongst knowledge economies, and 'Entrepreneurs, researchers, educational professionals, and students should be more challenged to excel' (MinEd 2011, 4).
This distinctive notion of quality as excellence is explicitly voiced in the 2007 Strategic Agenda The Highest Good and the 2011 Strategic Agenda Quality in Diversity. The recurrent argument that an increasingly complex society needs differentiation and flexibility now explicitly relates to quality as distinctive. 'Average is not good enough' (MinEd 2007, 5). The distinctive notion of excellence becomes proliferated in response to globalization and societal complexity.
In 2011, quality is positioned as a central notion, and the funding should be less based on student numbers and more on quality. Sharper profiling by institutions is needed to increase quality and to be recognizable to students and employers. Profiling leads to choices and topics at which one is good, and this also improves basic quality. Profiling furthermore increases differentiation, reduces fragmentation, and is needed to react to societal challenges.
To realize the desired increase in quality, several agreements are made at the central, sectoral, and institutional level. Institutions have to enter into individual performance agreements and make their own choices in profiling. Quality is not further defined. These individual agreements on performance indicators and institutional profiling differ strongly from the communal attempts to formulate performance indicators in the first HOOPs.
A different instrument, aimed at safeguarding basic educational quality, is applied for the universities of applied sciences, which enter into agreements to develop standard knowledge bases and central testing for core subjects.
The policies to increase and safeguard quality are based on recommendations by the commission Future-Proof Higher Education System. The advice to focus on profiling to increase educational quality holds a central position in their report Triple Differentiation (Veerman Committee 2010). The 2011 Strategic Agenda is primarily based on this report and on the governmental strategy to secure a leading position for the Netherlands at the top of knowledge economies. There is no explicit reference to the role of the institutions in this policy process, although the top is meant to be reached together.
In sum, the 2007 and 2011 Strategic Agendas put more emphasis on higher education quality improvement and focus on excellence in a globalized competitive context. The reasoning is that further differentiation raises the generic national quality level. Individual institutional arrangements replace the earlier dialogue, and less attention is given to institutional quality practices and valuations.

2015: situating quality
In the 2015 Strategic Agenda, The Value(s) of Knowing, 'a sharper view on the same horizon is taken than in 2011, with a greater awareness of the changing context of higher education' (MinEd 2015, 9). The local context is situated alongside a globalized context, and educational professionals and students are positioned as cocreators of an unpredictable future.
Dutch students and educational professionals widely share the belief that the future is not an abstract quantity that happens to us. The future is the result of today's and tomorrow's choices that we make together. Starting with a notion of the society we want to be. And which education is needed. (MinEd 2015, 1) Quality is related to the purpose of education, and higher education should allow every student to get the best out of him/herself (MinEd 2015, 22). The focus, therefore, should be not only on qualification for the labor market but also on socialization and personality building. The demands on students remain high however.
We demand more from students. More personal development, more of their academic or professional attitude, their autonomy, their ability to work together, their expertise, their effort and participation, their creativity and imagination. Educational professionals are the drivers of this learning process. I understand educational quality as all those (learning) activities that maximally contribute to this. (MinEd 2015, 22) This stipulative definition explicitly relates higher education quality to learning activities as well as to the goals of social development, socialization, and personality development. This perspective differs from the first HOOPs, in which sensitizing concepts like Bildung are not directly related to quality. Since the second half of the 1990s, excellence is named and linked to quality and so is Bildung. Although the quality concept remains vague, what it relates to becomes more profiled.
The tension in balancing the three goals of providing quality, access, and efficiency is addressed again. These three goals are repeatedly identified as competing, and in 2015 this tension is explicitly identified as a trilemma. 'More educational differentiation is also an answer to the trilemma, which means that we at the same time want to maintain accessibility to higher education, realize high educational quality, and spend [governmental] budgets efficiently' (MinEd 2015, 22). The reasoning remains that is not possible to achieve all three goals concurrently. Increased accessibility will lead either to higher costs or to a decrease in quality (Bronneman-Helmers 2011). The goals are treated as unified concepts; there has to be a trade-off somewhere, or the institutions will have to provide flexible solutions. The analysis of the policy documents shows that the preferred governmental solution is more flexibilization and differentiation.
In their situated uses, these three goals are not stable however. Different versions of quality, accessibility, and efficiency are played and juxtaposed in 2015, for example by contrasting meaningful learning communities with quantity and increasing student numbers. These notions differ from the meanings presented in the 1990s, whereby freedom of choice was valued over institutional excellence. Those situated meanings can all be related back to tensions between quality, accessibility, and efficiency, but they refer to quite different things. The situated meanings slide.
The 2015 Strategic Agenda differs from the 2011 Strategic Agenda in both its analysis and its positioning in the policy process. It is based on a higher education tour through the institutions, whereby educational professionals, students, managers, and others concerned are consulted. The agenda also reflects, however, the limitations resulting from previous practices and rules. The much-criticized performance agreements for example cannot be eliminated, as they have become a formalized part of the policy process. Previous quality practices recur.

1985-2015: wrapping up quality games
The HOAK paper starts with a meta-conceptual understanding of higher education quality. It follows the reasoning that the current steering relations are no longer effective in adequately advancing higher education quality in a changing and complex society. Systems dynamization and institutional autonomy are needed to enable flexibility, which in turn is a necessary precondition for the realization of higher education quality. Over time, this line of reasoning develops into a persistent and intensified flexibilization narrative, which understands quality as the result of changing steering conceptions, systems dynamization, and institutional profiling and differentiation. This understanding of quality is unified and rather abstract, as it is primarily related to the steering relationship with the institutions and their broader societal context. It changes however along this context, which becomes more globalized and competitive. This game can be directly related to New Public Management or businesslike steering conceptions, as they develop in a neoliberal society.
In the strategic policy sections, notions of quality as fitness for purpose, excellence, and Bildung are intertwined and vaguely played. They are not, however, undirected in their vagueness and change along dynamics in societal contexts that are considered relevant. The patterning through time displays a positioning that starts with quality as fitness for purpose in the Dutch labor market and develops toward quality as excellence in a competitive global context. The latest Strategic Agenda explicitly positions Bildung alongside this global context. The 2007 and 2011 Strategic Agendas explicitly name quality as excellence, and the latest document provides a stipulated definition of higher education quality. In these strategic games, sensitizing concepts like fitness for purpose and excellence relate to both quality and higher education in general. While several scholars originally related these notions to different quality rationales and practices, they are not prominent as organizing categories in the policy documents. Only the notion of 'quality as excellence' is explicitly foregrounded as a policy goal.
In the policy games, the strategy to respond with increasing flexibility, decentralization, and differentiation to growing societal complexity recurs. It is reflected in the policy solution to balance the trilemma between the conflicting policy goals of educational accessibility, quality, and efficiency at the institutions. These goals are treated as unified concepts, and the trilemma between these goals is explicitly solved by institutional differentiation, treating quality as well as accessibility and efficiency as abstract concepts. What these concepts mean in relation to each other however remains vague, and changes through time. As Stone notes, goals like quality, efficiency, and accessibility are treated as motherhood issues. Everyone is for them when they are stated abstractly, but the trouble begins when people are asked what they mean by them (Stone 2012, 14). From a governmental perspective, it makes sense to leave these trade-offs to the institutions.
Besides these prominent strategic games which are abstract and primarily treat quality as unified, the policy documents play numbers and indicators games. The first HOOPs contain labor market forecasts and sporadic international comparisons. There are attempts to set quality indicators that facilitate a meaningful policy dialogue between government and the institutions. This dialogue fades out in the 1990s however, without having reached shared indicators or meanings. In 2004, the addendum Mapping Knowledge is introduced, with different sets of indicators. The intent is to achieve shared sectoral goals and cautiously develop performance agreements with individual institutions. The 2011 Strategic Agenda, however, emphasizes individual performance agreements, without explicating their relationship with educational quality.
Another game played is the accountability and quality assurance game. The early HOOPs display intentions to relate the development of the quality assurance system to the formulating of policy goals. In the way that quality assurance is operationalized in the texts however, it is not related to the strategic games that give quality improvement pride of place. The policy documents devote small sections to the development of the quality system and address accountability-related notions only when the quality system changes or when there are problems with quality assurance. The government's delegation of responsibility for quality assurance to its monitoring bodies is a likely contributor to this.
Quality is thus simultaneously played by the government as both a unified and differentiated concept. These different enactments show divergent governmental practices interacting with changing social contexts. At times, the games are interrelated. For example, the current emphasis on excellence is reflected in the focus on institutional performance agreements. The concurrent unified and differentiated games however do not add up to a unisound governmental understanding of what quality is or should be.

Conclusion
In order to gain a better understanding what quality's contested and equivocal character means for complex, political and multiactor governmental decision processes, we have patterned how quality is constructed and enacted in Dutch governmental policy texts since 1985. The language games perspective enabled the identification of various quality enactments in governmental interactions with changing contexts. These patterns showed persistent uses of quality as a unified concept, as well as more differentiated enactments. These interactions changed toward a focus on institutional profiling and differentiation in a society which is understood as competitive. Its dual concurrence as a unified and differentiated concept however did not change, and did not add up.
How quality is enacted is rational from a governmental perspective. The different quality games probably all make sense according to the social rules developed in play and found appropriate for their institutionalized settings and contexts (March and Olsen 2006;Wittgenstein 1953). It is striking that the steering relationship with the institutions forms the pivotal notion in the governmental policies, and how quality is played and enacted relates to this steering relationship. Whether it concerns the abstract New Public Management game, the trade-offs of the evocative policy goals, or even the numbers and accountability games, they all somehow relate to this interplay with the institutions. Higher education quality is not the focal notion in these games.
With regard to the functioning of quality as a contested concept, we come to several interrelated conclusions. First of all, identity issues related with quality's use as a unified concept have become a shared responsibility with the institutions. In order to deliver quality to society, they literally have to develop their own identity in competition with each other, and solve the tensions between quality, efficiency, and accessibility in institutional practice. We also see that the interaction process with the institutions develops from an explicit and deliberative dialogue which aims to develop general indicators, toward individually negotiated performance agreements and concurrent development of standards to value these performances. Third, the government does not explicitly use quality against corporate institutional uses in its formal policies.
These conclusions are consonant with Lips' findings for the HOAK period and the first half of the nineties, that the relevant actors in the government, the institutions and their representative bodies did not come to shared underlying understandings of 'what quality is' (Lips 1996). They are however also in line with studies which show that both the government and institutional management easily recognize themselves in quality as a management concept (Giroux 2006;Stensaker 2007). In spite of the different enactments, this management discourse does not center around discussions on quality's true character, and the government and institutions do not tend to hold competing uses.
The interpretation of the policy texts however also shows that the relationship with institutional actors as academics and students changes through time. While the first documents barely address them as actors, the latest Strategic Agenda presents them as drivers of higher education quality, which is stipulatively defined as the student's learning process. This definition is in line with Harvey and Newton's initial plea to reconstitute quality's true character in educational practices, and to rule out distrusting bureaucratic requirements (Harvey and Newton 2007). The government seems to respond to this essentialist perspective. The overall perspective brought forward in the policy documents is however that the government distances itself from policies in the institutions.

Discussion
For the policy process with the institutions, the conclusions imply that it is not necessary to organize an initial dialogue to come to a shared language, underlying meanings, or quality standards before concrete policy negotiations are conducted. The policy process changes into practices, whereby the standards to value quality develop concurrently with its constructions, as they are individually negotiated. Discussions with the institutions about quality and performance agreements are just part of this management discourse, as it is played out in a competitive neoliberal society.
Democracy is however not confined to negotiations with the institutions on how the cake is cut (Mouffe 2000;Stone 2012). For a government who is responsible for assuring higher education quality, accessibility, and efficiency to society as a whole, this not only means assuring that deviant perspectives and other actors are represented but also that they are actively heard and attended to in the management discourse and situated institutional discourses. The analyzed policy texts do provide some insight whether this is the case. They are however vague, and do not explicitly articulate concrete emancipatory issues and their translation in concrete policy measures. The study ability of study programs by different groups of students has for example gained a prominent position in the performance agreements, though this issue is only vaguely mentioned in the policy texts.
Strategy concepts are enacted as different concepts across discourses (Seidl 2006), and the relatively closed discourse with institutional management indicates that this is the case for higher education quality. Further research is needed to understand whether and how quality is articulated in the institutions, and whether governmental policy measures are seen as contributive or hindering the student's learning processes.
The point is that we do not have external or pre-given guarantees to value whether the different and contradictory perspectives are actively heard and attended, and neither does the government as a universal actor. A universality living in an unresolvable tension between universality and particularity makes it contaminated, and this is problematic for a government which is understood to act on everybody's behalf (Laclau and Mouffe 2014, xi). A full understanding of the articulations at stake in the boundless educational domain may be an interesting academic project, but it is not feasible for the government as a universal actor. As such the current development toward a multiplicity of quality practices paradoxically runs the risk of fostering further bureaucratic distrust, since these multiple policies are paired with formal accountability demands. Its vague and elusive character thereby remains. The good news is however that the Dutch government and the institutions are currently reworking the quality assurance system and its practices, in order to reduce these multiple accountability demands.
The growing domain of valuation studies focuses on the concurrence of evaluative practices with constituent practices like the educational process, in relation to decisionmaking (Lamont 2012(Lamont , 2009Heuts and Mol 2013). Though quality's vagueness is not specifically attended, this brings in new possibilities to further inquire language games as an open and flexible perspective and toolbox to study how quality is played in multiple interactions and constellations. How academic middle managers value and realize quality, and what they thereby identify with in their decision processes is for example an interesting question which contributes to the understanding of the games at stake. Let's play. Note 1. The Netherlands has a binary higher education system with universities (universiteiten) and more practice-oriented universities of applied sciences (hogescholen).

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Kasja Weenink has worked as a strategic policy advisor in several organizations in the Dutch higher education and science domain. She is currently conducting a PhD project in communication science that focuses on how higher education quality gains and creates meaning in interaction with different contexts.
Noelle Aarts is Professor of Strategic Communication at the Amsterdam School of Communication Research (ASCoR), University of Amsterdam, and Professor of Communication and Change at Wageningen University & Research. Focusing on conversations, she studies interhuman processes and communication and how space for change is created in these.
Sandra Jacobs is an assistant professor working on corporate communication at the Amsterdam School of Communication Research (ASCoR), University of Amsterdam. In her research, she focuses on the interplay between organizations and the media.