Institutions and the Artworld – A Critical Note

1 Is the Artworld a Social Institution?

Institutional accounts of art, although less popular now than in the 1970’s that were the heydays of social constructionism, seem increasingly less plausible once one begins to apply insights from contemporary accounts of social institutions as they are studied in the emerging field of social ontology (Searle 1995, 2010; Tuomela 2013; Epstein 2015; Guala 2016) and in economics (Pratten 2015). The artworld, on the account we propose, does not differ from other widespread and distinctively human spheres of activity like science (or “the world of science”, to sharpen the analogy with art), education, the sphere of travel, or even cleanliness. All involve, and in many cases arguably require, social institutions to flourish. The institutional similarities among these spheres will become clear once the independently motivated coordination function of institutions operative in those spheres of activity comes into view.

Let us first briefly remind the reader of what was at stake. George Dickie’s original definition of art, in the first round of what would be a long and hefty debate (see Graves 2010 and Fokt 2013 for a useful overview) reads as follows:

A work of art in the classificatory sense is (1) an artifact (2) a set of the aspects of which has had conferred upon it the status of candidate for appreciation by some person or persons acting on behalf of a certain social institution (the artworld). (Dickie 1974, p. 431)

Key in this definition was the concept of an artworld (sometimes capitalized as Artworld), a neologism borrowed from Arthur Danto’s eponymous paper (1964). But merely calling the artworld an institution does not make it one.^[1] The concept of social institution adds an important qualification to the designatum of “artworld” and its applicability must be motivated. Genuine institutions have their characteristic properties independently of how they are named or labeled (Guala and Hindriks 2014; Guala 2016) or how participants in the institution think of the rules they follow (Searle 1995). Whether a human practice involves formal or informal institutions (and not, for example, a natural order, individual preferences or a moral code) need not be fully transparent to those who act within the institution, who follow the rules unreflectively (Buekens 2013) or who have no conception of there being alternatives to the way they solve problems of interaction among participants in the practice. We are not always self-conscious about the fact that a practice in which we participate has an institutional dimension to it, and becoming reflective about the institutional character of a practice can bring about major shifts in our attitudes towards it.

It is therefore not a trivial question whether the artworld does refer to a social institution as Dickie’s definition uncritically assumes. If it does, then an account of institutions – their function, the way the rules of the institution work, their status-creating powers and how a distinctive social ontology within emerges – should straightforwardly apply to it. However, if the concept merely referred to a human practice or a sphere of activity, the institutional definition of art loses much of its provocative power, for the idea that art arises in a culture and constitutes a typically human sphere of activity is utterly trivial and was never disputed.^[2] That is the dilemma we will work with in this paper.

Much of what was thought to be controversial in Dickie’s original definition was neutralized by two caveats: Dickie spoke about art in the classificatory and not in the normative sense, and, secondly, the social institution that figured in the definiens – the artworld – confers upon an artifact the status of a candidate for aesthetic appreciation. The intended upshot of the argument was, as Danto pointed out in “The Artworld”, that it is not an exhibited or intrinsic or natural property that is responsible for turning an object into a candidate for appreciation. Danto’s intuition pump that was supposed to support institutional accounts made use of Andy Warhol’s famous Brillo Boxes, which were exact copies of ordinary brillo boxes to be found in the store around the corner. The decisive difference was that the Stable Gallery in Manhattan, where Warhol’s Boxes were first shown to the public, was an established art gallery. And the art gallery was itself an ingredient of the artworld – it functioned as an institution that contributed to the Boxes’ aura qua work of art.

Like social constructionism in general, the institutional theory of art came with a distinctive provocative flavor: if phenomenon X turns out to be a social construction (a phenomenon not found but created by us), its existence was contingent upon ways of thinking of or conceptualizing the world. The idea was that what counts as art depends on something external to it, viz. “a theory of art (…) the theory that takes it up into the world of art, and keeps it from collapsing into a real object”, as Danto put it (Danto 1964, p. 581). The second key statement in Danto’s famous paper came at the very end of the paper, and it has a distinctively constructionist flavor: “It is the role of artistic theories, these days as always, to make the artworld, and art, possible. It would, I should say, never have occurred to the painters of Lascaux that they were producing art on those walls. Not unless there were Neolithic aestheticians.” (Danto 1964, p. 581)^[3],[4] “Museums, connoisseurs and others are makeweights in the Artworld”, Danto concluded (1964, p. 584).

While institutional accounts of art may be less popular today (but see Fokt 2013, 2017 for an interesting overview of recent developments and the many meanings of the concept of artworld), some important lessons can be drawn from the way it connected a human practice, the objects produced and the process whereby the products were created. According to Danto’s model, it was theory (or Theory!) that was the hidden, driving force. According to Dickie, it were members of the artworld. We will focus on the latter approach that has in various versions inspired the institutional theory of art.

Let us first mention, just to put them later aside, some less important issues. A superficial worry with Danto’s and Dickie’s institutional account was that at a theoretical level it was seen to be reflecting a disturbing feature of Conceptual Art, epitomized in Marcel Duchamp’s Fountain: there are no limits on what can count as a candidate for aesthetic appreciation. The only constraint was, it suggested, that an object be publicly recognized, or accepted, as a candidate for aesthetic appreciation. The intuitive appeal of Dickie’s definition was, as Dutton, nicely put it, “the presupposition that Dada is a central form of artistic practice” (2006, p. 367). Another objection was that those who doubted the artistic integrity of conceptual art – who thought (or think) that modern art is somehow fake – could dismiss Dickie’s definition and the creative role assigned to theory by Danto as a fancy post hoc justification of what is in the critic’s eyes a highly dubious practice (a justification generated by the emerging constructivist atmosphere and postmodernist attitudes of the day). But that reaction rests on a misunderstanding. The institutional account (“the artworld creates arts”, “theories pick out objects and place them in the artworld”) does allow, as Robert Kraut pointed out, critical dialogue, negative evaluation and attributions of fraudulence – all of this being consistent with the idea that the status of being a candidate for aesthetic evaluation is constituted by the artworld and not discovered by engaging in a practice (Kraut 2007, p. 41).

2 The Artworld, Art and Solutions to Recurrent Coordination Problems

The artworld is (according to the institutional approach) constituted by a community of agents, who are engaged in conferring a certain status to objects, which is that of being a candidate for aesthetic appreciation. (This fits well with the well-known Searlean approach to institutional entities, for which the idea of collectively imposing a status function is key to understanding the nature of institutional facts). But whether and to what extent this justifies an institutional definition of art depends on what you take to be the core features of social institutions, and to what extent the artworld and works of art enjoy many or most paradigmatic features of bona fide institutions c.q. institutional objects, properties and events “generated” or “produced” by the artworld.^[5] Our contention is that the artworld cannot be a social institution, although the artworld as we know it and as it has historically developed surely has created numerous types of institutions. This is because art, like any other activity that attracts interaction among humans, requires that coordination problems be solved and stable equilibria be found and maintained. The emergence of coordination problems and finding (or stumbling upon) stable solutions, however, is extrinsic to the phenomenon of making and appreciating art that form arguably the core of the artistic sphere of activity. The artistic sphere of activity is in that sense fundamentally different from economic spheres of activity, where institutions are an essential ingredient of the practice because economic activity is defined by human interactions, and more comparable to science and scientific activity (the practice of producing knowledge about the world), or education (the practice of transmitting knowledge to new generations), or, if you think of it, the practice of keeping things clean – practices supported by but nevertheless not grounded in, social institutions.

Institutions (the adjective “social” is redundant) are in Douglass North’s now classic formulation

the rules of the game of society, or, more formally, the humanly devised constraints that shape human interactions (…) They are a guide to human interaction so that when we wish to greet our friends on the street, drive an automobile, buy oranges, borrow money, form a business, bury our dead, or whatever, we know (or can learn easily) how to perform these tasks. (North 1990, p. 4)

Just as there are physical limits to individual choices, choices based on coordination are regulated by systems of rules that provide information about and incentives for appropriate ways of interacting in different social situations (Smit et al. 2011). An equilibrium approach to institutions (the approach favored by the authors of this paper) would stress that institutions reflect regularities in behavior, which are agreed to by all or most members of society, where the clustered and often complex regularities are best seen as solutions to coordination problems that can be described in game-theoretical terms – more specifically, the theory of coordination games. One key insight developed by game-theoretical approaches to emerging norms and equilibrium-selection is that such solutions can emerge spontaneously, that choices of equilibria are often based on what is salient, and that they regulate and channel expectations of anonymous participants (Lewis 1969; Ullmann-Margalit 1977). The way a particular coordination problem is solved in the past can assume the status of rules or norms when it can be taught to others, when it is codified so as to avoid ambiguities and can enhance the salience of the solution.

Key to this approach is that institutions are clusters of rules that tend to solve coordination problems and collective action dilemmas that exist independently of the institution and that they provide collectively beneficial (though not always optimal) solutions to these problems. Art and science, insofar as they are practices that create coordination problems among agents (due to division of labor and interpersonal hierarchical dependency relations or other more natural social facts; see Fiske 1992), are practices in which institutions emerge that shape stable and easily learnable ways its goals can be achieved.

On this account, institutions are systems of interconnected rules that direct and shape preferences of participating agents in achieving shared goals – goals that exist independently of the kind of coordination required to achieve the goals. If the goal of science is to find out significant and useful truths about the world, it does not follow that just because science involves the deployment of institutions (universities, research agencies, funding organizations, libraries) those institutions are some way or another constitutive in shaping science’s overall goal. “Finding out significant truths about X” is not a goal that requires coordination, or that for conceptual reasons requires successful coordination among agents. (That explains why Wittgensteinian analogies with ludic games – games as prototypical examples of institutions – are misleading in this context because playing and winning a game are often taken to be ends in themselves).

From the point of view of the participants (the “internal” point of view, to borrow a useful term from philosopher of law H. Hart here [1961]), institutional realities manifest themselves in a wide variety of verbal and non-verbal status markers: that is, in signals, dedicated places, texts, documents and inscriptions, ways of classifying items, ornaments, etc., which channel expectations of participants, tell participants what to expect from others, and how to interact with others. The vast realm of public coordination devices that surround us as a realm of signs can be said to represent, stand for or indicate ingredients of a social ontology insofar as those objects, persons, places and times begin or continue to regulate mutual expectations of participants. The devices themselves function as declarative and imperative signals that tell participants how to choose or what to do in a given situation. They are, if you like, part of a second-order institution: the institution that makes a first order stable equilibrium public via signals, markers, indicators and labels, and that invites newcomers to conform with the established practice. On this account, having an institutional status (or having a status within an institution) is an abstract relational property of agents, objects, events, places or times they have obtained in virtue of being involved in stable solutions to recurrent coordination problems. Public markers create common knowledge among the participants who are supposed to understand how the practice works and that objects or persons have a status within the institution. They also offer publicly available reasons to act in certain ways. If you do not understand the second order coordination devices (the public markers with instructions attached to them), you do not see institutional objects, actions, places or persons as reason-giving entities, as things that give you a reason to make this or that strategic choice, but (nonetheless) the underlying equilibria selected by the practice do not require labeling to become equilibria. And finally, when we become reflective about this process, we can bring the process of creating stable solutions to coordination problems under intentional control, which requires language and, more specifically, a self-conscious practice of rule-making and labeling (Smit et al. 2011; Buekens 2013; Guala 2016).

3 Artworld Institutions Solve Coordination Problems

Institutions, as ingredients or constituents of the artworld, solve obvious or subtle coordination problems. The museum, the gallery, educational programs about art, a department in a humanities faculty and art schools are ways of organizing art-involving practices (exhibiting art, collecting art, studying art, developing creative skills required for the production of art forms, etc.). But, we contend, it does not follow that those practices are constitutively linked to the making of art; that is, those practices do not and cannot quasi-magically turn objects into art or (as Dickie would have it) candidates for aesthetic appreciation. The activities that produce such candidates have their origins in pre-artistic psychological tendencies (a certain awe for special achievements, the felt need to get in contact with objects contaminated by the aura of their producers, the need to transmit skills needed to create them) that are not specific to art and serve countless other purposes (Bloom 2010). In this respect, compare the artworld to the world of science. Science, as a sphere of distinctively human activity, gradually created various types of social institutions in which scientific insights were found, taught and publicly evaluated and appreciated, and those institutions are studied in the sociology of science. But science itself and its products (theories and models of the world) are not social institutions – not things that exist in virtue of the social institutions that the practice of investigating the world around us has created. The institutions are only in a metaphorical sense “in charge of” finding out important truths. Moreover (and going back to art again), it is utterly mysterious how galleries, museums and departments of arts could be “in charge of” what George Dickie described as creating candidates for appreciation, and to what extent participants in those institutions can act “on behalf of” the artworld. It is at these crucial points that the institutional theory must draw on unhelpful metaphors, picturesque analogies and suggestive language that create illusions of insight. We know very well what the role of a museum (qua type of institution) is – how it coordinates interaction between visitors, curators, artists, sponsors, and philanthropists; but it is not the role of the museum to create objects of artistic appreciation (compare: a toy museum does not create toys). In genuine institutions it is often very clear who assigns a status, which status is assigned, and when and how the status can be changed or lifted.

We pointed out that institutions are instrumental in bringing about pre-institutional shared and complementary social interests in an orderly and economic fashion. The best way to see the connection between evolution and culture as offering similar solutions to social interactive problems is to appreciate how game theory can be applied in both areas (Gintis 2014). Borders are the institutional successors of territory behavior, contracts and exchange structures developed out of tit-for tat interactions, hierarchical power relations reflect pecking orders, etc. But It is unclear how the artworld as it functions in Dickie’s original definition could be connected to pre-existing human needs or how the rules of the artworld (if any) codify pre-art social behavior. The real connection between artistic activities and institutions seems to be this: art-creating activities are often meshed with shared interests and joint goals; and insofar as institutions are networks of stable equilibria that solve recurrent coordination problems a theory of institutions can explain what are orderly and economical practices – in the artworld, the world of science, or the world of education. What a museum, an art gallery or music academy does for us is, from that perspective, perfectly clear and not very spectacular. How the artworld itself (over and above the wide range of formal and informal institutions that coordinate art-involving intentions and actions) performs the additional job of conferring candidacy for aesthetic appreciation upon physical items remains utterly mysterious.^[6] Agents with an institutional office (curators, for example) can act on behalf of a museum or a gallery, but never on behalf of the artworld, just as one can act on behalf of a scientific academy or as conference organizer, but not on behalf of science – if one acts “on behalf of” science, one is simply pursuing scientific goals. Acting “on behalf of science” is at best an empty (and pretentious) metaphor. Acting on behalf of one’s position as (say) a lecturer or a director of a scientific institution is not.

It has often been remarked that institutions make new kinds of behavior conceptually possible (“winning a set”, “signing a leasing contract”, “buying a beer”, “becoming a full professor”) relative to pre-institutional behavior. This point is particularly stressed in John Searle’s declaration-based approach to social ontology and the ontology of institutions (Searle 1995, 2010). Searle holds that institutions are structures of what he calls constitutive rules, i.e. rules of the format “X counts as Y in context C”, where the Y-term denotes or describes a status function of a physical item. When a person obtains an institutional status via a “status function declaration” (some sort of collectively accepted performative act), a profile of deontic powers (rights, duties, privileges) applies to her and agents with whom she interacts (under a description that mentions the institutional properties). The structure of normative positions can be studied independently of the institutions that create them (Lindahl 1977). Moreover, mature, well-codified institutions behave, according to Searle, in many ways like ludic games: the rulings introduce terms and other public markers that help us publicly represent the newly created status functions, which are themselves “invisible”. (This is consistent with the fact that institutions solve pre-institutional coordination problems: the public representations mark for the participants, which rules apply, and the rules assign a deontic status to agents. The innovative power of institutions is often to create common knowledge of solutions to interaction problems).

On Searle’s view it is central to institutions that they require collective agreements that assign deontic profiles to interacting agents. The paradigm example of such a collective agreement is that of a promise (a public speech act), where agents, by promising things, create expectations and desires that cannot be defined in terms of purely physical or psychological dispositions. Promises, according to Searle, create rights and duties, and they are paradigmatic for what goes on when and if one acts under the aegis of a specific institution. There could not be “artists in residence” or “curators” or “professors in the history of 18^th century art” without there being certain rights, duties, privileges and other commitments assigned to them and persons they interact with under that institutional assignment. This in turn requires a form of collective agreement with respect to the obtainment of those assignments: we collectively accept that X is going to count as Y. On the other hand, nothing excludes that, even if their actions are intentional under descriptions derived from their institutional status and the deontic profiles that comes with it (“he opened the conference”, “she refereed a paper”) there will always be further descriptions under which their actions are based on reasons not created by the institution. We recognize rights and duties because that is what museums, galleries, and art departments come with.

But on this view too it is problematic how the artworld itself could be an institution, i.e. a collectively accepted practice. It is typical for institutions as conceived of by Searle that they can be created by decree, that those decrees can be issued and promulgated by some appropriately endowed authority, and that sanctions they involve might be organized or even themselves institutionalized. But there is no sense in which the artworld itself was created by decree. Still, there is an obvious sense in which we can, given our knowledge of past practices, create a museum or any other special office to deal with bits of art.

Other considerations, now based on equilibrium-based approaches sketched earlier, point in the same direction. Social institutions tend to make certain joint activities psychologically and computationally undemanding. They stabilize behavior and make us – our actions, beliefs and intentions – more predicable by others. You know what to expect from the visitor of a gallery or a potential buyer at an art auction, and all parties involved usually know what the costs are in cases of non-compliance. A key concept at work here is that of common knowledge (Lewis 1969). Common knowledge (I know that p, you know that p, I know that you know that I know that p, etc.) of the rules of the game is required when making choices dependent on what the agent herself expects others to do in this or that situation. What is common knowledge (or supposed to be common knowledge) in particular settings is created by salient public signs, symbols, documents, announcements, certificates, but also architectural layouts and other symbolic techniques that create focal points for participants in the institution to coordinate on (Schelling 1960). Public signs and symbols, when understood by the participants, often act as “choreographers” (Gintis 2014) of players who are on the receiving end of the public representations. The generation and distribution of common knowledge through documents and certificates is key to understanding cultural practices that require coordination of strategic choices of people who would not otherwise interact with each other.

These observations uncontroversially apply to art-involving institutions. Understanding the function of a museum or a gallery means that you understand your status when you “enter the game”. Their specific lay-out, the architecture, the signs and documents you read or acquire when you enter them (literally!) publicize your deontic profile in the game you are playing: you buy a ticket, become a visitor, see pictures at the exhibition, and perhaps talk to the curator at the opening reception.Searle (1995) suggests that it is “obvious” that these concepts reveal institutional statuses. We have some doubts about that, but what is obvious is that public coordination devices (we called them second order coordination devices) indicate that you are a visitor (your ticket), enter the exhibition (entrance indicators), speak to the curator, and so on. A further and more contentious issue is how and to what extent the institutional setting in which an object appears has impact on the way we value that object – how the setting impacts on the “good reasons” we have for valuing this or that object (see Bloom 2010).

Agents involved in institutional interactions are supposed to act in ways consistent with the deontic profile derived from the rules of the institution as they are commonly understood.^[7] Expectations (Searle’s deontic powers) are key to understanding what acting under an institutional description means, but neither Dickie nor Danto refer to them. And yet, is it not obvious that the art gallery creates formal and informal rights, duties, authority lines and privileges for the agents involved (museum directors, art critics, trendsetting collectors)? Bits of art do not as such have deontic powers; owners, collectors, museum directors, artists, visitors, etc., have them. The similarities with science (or the “science world”, if you like) are obvious: science as such is not a practice that depends for its existence on the possibility that certain (physical) objects count as something else; but it is impossible to think of scientific institutions without deontic profiles assigned to the ingredients of and participants in the institution (students, teachers, deans, etc.).

Science produces theories, the artworld produces art. But neither art nor scientific theories require for their proper appreciation that scientific or artistic institutions be evaluated. Institutions can be evaluated along various dimensions. Do they perform well? Can some of their unintended side effects be avoided? To what extent do they promote (our) moral or social values? Institutions have good and bad, successful and unsuccessful, honest and corrupt versions. There are good and bad museums, well-curated and old fashioned galleries, just as there are corrupt gallery owners and mala fide artists who seek to exploit the rules of the game. But the artworld as such cannot be evaluated – there are no good and bad versions of the artworld. Similarly, it is one thing to evaluate scientific institutions and another one to evaluate the intended products of science – theories and models. We evaluate scientific institutions as stable equilibria that emerge to help scientists do what they are supposed to do – to solve problems that are not themselves intrinsically institutional. The promotion of sound epistemic practices, open discussion and independent research are qualities that can be promoted or endangered by institutions. But the intended canonical product of scientific practice – well-verified theories about the world at large – are not in any sense created by scientific institutions.

And finally, it cannot be true that if institutions arise within an activity or practice, then the full nature of that activity or practice’s ingredients – the objects, properties and actions it allows – are finally coming to light, for this would make almost any activity “institutional”. Numerous institutions emerged in the context of education, but it does not follow that learning to solve a mathematical equation or reading a book thereby becomes an intrinsically institutional act; just as it would be absurd to hold that because we have institutions that deal with travel and leisure, the fact that I traveled 1000 miles yesterday thereby becomes an institutional fact. The institution of a restaurant does not turn eating into an institutional fact.

The institutional theory of art does not fit well with current work on institutions and social ontology; and for good reasons. “Artworld” is best taken to be a generic, non-theoretical cluster concept that designates a sphere of activity that involves activities that promote goals that have led to the emergence of formal and informal institutions built around various forms of activity that involve art (the objects). But those institutions, rather than “creating” objects for aesthetic appreciation, merely solve familiar coordination problems created by artistic activity in ways that other institutions in other areas solve similar coordination problems. There is no sense in which institutions or the institutional components of the artworld “create” art. Robert Kraut holds that “(t)he property of being art does not explain the existence of an artworld; it is the artworld that explains the existence of art” (Kraut 2007, p. 47). Our deflationary account of the relation between art and the artworld questions both claims. There would not be art, unless there were a broad sphere of human activity that we might denote with the concept of artworld (that is why the painters of Lascaux were not artists) – but this trivial observation neither entails nor supports an institutional account of art. It merely predicts that institutions would emerge once art-involving joint goals and commitments emerge.

Our account has the further advantage that it takes the sting out of the debunking allure of the original institutional theory of art: there is, just on the basis of what the original theory says, no sense in which institutions enter in explanations of how we like what we like – why we value art. That art could be defined in terms of what those institutions achieve for us (“creating objects of artistic appreciation”) is unwarranted on the account presented here. If institutional settings affect our aesthetic appreciation, we need to turn to social psychology (Bloom 2010). There is now evidence that the institutional setting in which an object is placed, knowledge of how it was produced and who produced it, can permeate the aesthetic experiences people have when they contemplate particular objects – when they know or believe that they have a candidate for aesthetic appreciation before them (Newman and Bloom 2011). These are important psychological findings and perhaps show that there was, after all, a grain of truth in Danto’s statement back in 1964 that we need theory to appreciate art.

The argument we have put forward generalizes. The world of art shares, with the world of science and the many worlds of religion, that a practice through cultural evolution gave rise to the emergence of (sometimes quite complex) social institutions. The best explanation for this phenomenon is that we are a highly cooperative species that solves coordination problems by creating public markers that create common knowledge of how to coordinate our actions when engaging in a social practice. Many endeavors in art, science, the passing on of knowledge, etc. have goals that can only be realized via coordination, which requires that we – the participants – want each other to know when, were and how to adjust our choices and actions so that those shared goals can be realized. It remains an open question whether the concept of artworld was more than a fancy signifier, which obscured rather than illuminated the important and genuine institutional dimension of art and artistic activity.