1 Introduction

This paper follows the catchy idea presented by Hindriks and Guala (2019) that institutions are enablers. For example, the institution of traffic rules enables drivers to reach their destinations, and the institution of private property enables individuals and companies to reserve the exclusive right to benefit from some assets.

The dominant theory of institutions takes them as formal or informal rules constraining individual or collective actions, thereby enabling individuals and collectives to solve the problems they recognize in their surroundings. However, the theory assumes that specific segments of the world are specified or singled out as problems that call for institutional solutions. Meanwhile, this specification is by no means a trivial capacity, and – as I will be arguing – is only possible within a pre-existing, institutionally governed structure of epistemic pursuits. Therefore, problems to be solved are not merely given, as if awaiting someone to address them in a rule-governed manner. Instead, they must be recognized as problems, and this recognition depends on institutions.

So, I argue in this paper that institutions not only enable individuals and collectives to achieve their goals; first and foremost, they enable individuals and collectives to have a goal – generally speaking, to select and recognize certain segments of an otherwise rather convoluted and messy environment as problems to be addressed – and as a result, to have a demand, especially a demand for further institutions.

Therefore, the goal of the paper is to unveil and discuss a key epistemic function of institutions, that is to say enabling people to problematize their surroundings, which paves the way for (institutionalized) problem solving.

What I am going to stress later on is that the epistemic function of institutions is first realized in the form of questions, which are linguistic expressions of acts of problematization. Finally, I will focus less on what an individual subject questions or problematizes, and more on the social dynamic of questioning. This dynamic will be conceptualized in terms of epistemic dependence.

The concept of epistemic dependence refers to situations in which what one knows or believes to be the case hinges not only on one’s own epistemic or cognitive capacities, but first of all on the capacities of others. So, I depend epistemically on my friend or my community when variations in their beliefs would make for variations in a my own beliefs.

Erotetic dependence, which - I propose - is a specific type of epistemic dependence, refers to questions and our effort to problematize our surroundings. In short, it refers to situations where the questions I ask are epistemically dependent on the questions formulated by my peers.

With this in mind, let me reiterate that the paper aims to demonstrate that either all institutions have a built-in “epistemic aspect/ingredient” or that within the broader network of institutions there are specifically “epistemic institutions” whose function is to regulate erotetic dependence within a collective.

To substantiate my case, I shall make use of the standard distinction between elite (experts) vs. people (non-experts) which has also been employed in institutional economics. On this basis I will argue that properly functioning (epistemic) institutions maintain a dynamic in which a problem, first identified by non-experts who are directly acquainted with a specific situation, is then presented to elite expert groups, thereby establishing erotetic dependence of the latter on the former. Subsequently, once the experts come up with answers (and further questions), the direction of epistemic dependence, specifically of erotetic dependence, changes so that the non-experts now benefit from their dependence on the elite. This is because the non-experts are now able to articulate the problems in their surroundings in greater detail and with greater precision. This dynamic underlies the collective’s capacity to formulate a shared goal or problem to be solved.

In contrast, badly functioning institutions either draw no connection between the epistemic pursuits of both groups, or if they do, the connection gets broken when the elite expert group does not “pay the non-experts back,” so to speak, and refuses to give proper recognition to the pursuits of the people so the latter can benefit from the elite answers/solutions.

The claim, therefore, is that we first need epistemic institutions (or epistemic aspects/ingredients within broader institutional networks) responsible for the recognition of problems and the creation of specific demands. Only once this framework is in place can institutions that enable problem-solving be established in the form of rules. Schematically, we have:

Epistemic Inst. → Problem recognition (questions) → Demands → Institutions as problem-solvers (answers)Footnote 1.

The paper is organized as follows. First, I describe the dominant account of institutions called the rule-following view, already flagged above. The focus on institutions-as-rules stems from the fact that economists have produced a substantial body of evidence referring to how the rule-following approach works in real-life economic policies.

Second, I shall roll out a step-by-step proposal for representing institutions with a specific structure of epistemic dependencies. In particular, with reference to the relevant literature in social epistemology.

Third, I roll out the elite vs. people conceptualization and propose that properly functioning institutions are characterized by a division of epistemic labor between these two roughly distinguished groups. This claim will be illustrated by empirical findings referring to the institution of worker participation, or worker voice, within the workspace improvement process.

It should also be noted that there is a widely discussed alternative to to the rule-following approach – the so-called equilibrium approach – championed by Aoki (2001, 2007) and originating within the context of game theory (see Binmore, 2010, Hedoin, 2017; Greif and Kingston, 2011; and Hindriks and Guala, 2015 for an integrative view). I believe that the arguments which shall be developed here would also apply to the equilibrium approach, but this would require a separate paper to discuss properly. I only mention it briefly in the final section, discussing the added value of the proposed approach when compared with existing ones.

2 Institutions: Beyond Rules

2.1 The Rule-Following Approach: Theory and Fact-Checking

The prevailing conceptualization of institutions in the economic literature takes them as formal (e.g., the institution of private property) or informal (culturally-grounded) rules governing behavior (see e.g. Acemoglu, 2006, Acemoglu et al., 2021; Hodgson, 2006; and Ostrom, 1990, 2005 for a slightly different and, in my reading, thorough proposal). As Douglas North famously puts it, institutions are “the rules of the game of society or, more formally, the humanly devised constraints” that “establish a stable structure to human interaction” (North, 1990, p. 6). They are “codes of conduct, norms of behavior, and conventions” (ibid.: 36).

There is widespread debate on whether the rule-following approach can plausibly account for the phenomenon of institutional change as well as for the very origin of institutions (see Greif, 1998; Kingston & Caballero 2009; Hindriks and Guala, 2015; Guala, 2016). But I shall focus on a different concern, beginning with the problem of the implementation of an institution, a topic supported by a significant amount of empirical data. But overhauling the data is not the goal. The point is that paying special attention to implementation issues provides us with a glimpse of the notion I would like to elicit – the epistemic function of institutions (Petracca & Gallagher, 2020 is in a somewhat similar vein).

To begin with, let us note that in the rule-following approach “institutions are ultimately best understood from a functionalist perspective that recognizes that they are responsive to the interests and needs of their creators (although there is no guarantee that the rules selected will be efficient)” (Greif & Kingston, 2011, p. 14). Therefore, two crucial things stand out.

First, there are interests and needs which institutions are supposed to answer. Secondly, they may answer these needs poorly. This creates a challenge, for in order to assess what a given institution aims at we must first be able to identify the needs and interests. Meanwhile, a serious difficulty that we come across when trying to make a fair assessment of the rule-following view, as it works on the ground, is the fact that the realm of genuine needs, and interests in particular, can be somewhat muddled. This is because the rule-following approach has been embroiled in political practice with various troubling consequences. In particular, it began to serve as a rationale for various international institutions, such as the OECD, the World Bank, and the IMF, as they began to put pressure on developing countries so that they would implement “better institutions”, which was, at least insofar as declarations are concerned, was aimed at increasing their economic opportunities (see e.g. Alvarez et al., 2000; Aron, 2000; Chang, 2001; Glaeser et al., 2004). This was not necessarily the case though. Ha-Joon Chang (2011) points out in his critique of institutional economics that “economic growth has fallen rather dramatically in developing countries of Sub-Saharan Africa and Latin America, which have, under enormous external pressures, rather faithfully reformed their institutions in the neo-liberal direction during the last three decades. They were growing much faster in the 1960s and the 1970s, when they lacked those ‘liberalized’ institutions.” (Chang, 2011, p. 483; see also Bronckers 1994).

The main point of Chang’s criticism is that the institutional approach to economic development in a sense became hostage to powerful interest groups. One might argue that this has nothing to do with the plausibility of the rule-following approach as such. The question is then to what extent these practical/political missteps can teach us, as a more general (philosophical) lesson.

2.2 Local Demand for Institutions

The rules recommended to developing countries, thus formal institutions in North’s (1990) sense, had originally arisen in countries that had already built their own wealth, and they were usually dedicated to privatization and strict intellectual property protections as well as the liberalization of financial markets and the real economy. This gave rather mixed results, e.g. in post-socialist countries entering capitalist reality after the collapse of the Soviet Union (see Cavanagh, 2004; Klein, 2007; Kowalik, 2011; Allen, 2016).

These economic facts lead to a serious philosophical issue. First, note that if it were part of the very notion of an institution that it must emerge locally and cannot be “transplanted” or enforced in a top-down manner, the above-mentioned struggles in developing countries could have been avoided, at least to some degree.

Therefore, local applicability of institutions constitutes a very serious theoretical puzzle. In short, according to Chang (2011), the implementation of institutions-as-rules often fails because these institutions do not correspond accurately to actual needs, and crucially, actual demands. The latter is a key notion. As for any other entity, there needs to be a demand for specific rules, thus also for institutions-as-rules, otherwise they can only remain paper fictions or a great burden on developing economies. Chang argues:

Economic development changes institutions through a number of channels. First, increased wealth due to growth may create higher demands for higher-quality institutions (e.g., demands for political institutions with greater transparency and accountability). Second, greater wealth also makes better institutions more affordable (…). Third, economic development creates new agents of change, demanding new institutions (Chang 2011, p. 476)

The issue of demand is critical. Let us consider a historical example relating to trade and tariffs. Think of Alexander Hamilton’s famous infant industry argument (see Bairoch, 1995). Hamilton articulated a demand coming from growing and thus precarious enterprises in the United States for some sort of barrier against competition from large European firms dominating the market. The new firms needed time, Hamilton argued, like a child needs time to mature, to prepare for competition in an open, free market (for a more general overview of this topic, see e.g. Corden, 1997; Melitz, 2005). Now, when considered as abstracta, such protectionist institutions-as-rules can be implemented anywhere, but what makes them genuine institutions, I argue, is a real local demand for them. However, here is a subtle point worth eliciting: the demand on the side of “infant” entrepreneurs in Hamilton’s case was not for rules themselves. I propose that the demand was rather less determinate – it was a demand for something to happen, as a result of which there would not be ships docking at American ports, loaded with cheaper products from England. Therefore, this demand referred to a desired state of affairs, not to any policy itself. A relevant policy answers the demand if it makes the demanded state of affairs (e.g. European ships out of US ports) a fact.Footnote 2

2.3 What it Takes to Demand an Institution

The key point I would like to highlight should already be clear: all actors involved in the creation and implementation of an institution-as-a-rule recognize certain things or states of affairs (thus potential facts) as beneficial or useful. Otherwise, there would be no demand for that institution whatsoever. Therefore, however trivial it must sound at first, all the actors are in a certain epistemic state linking them to what is the case and to what they need. Note, however, that nothing is said so far about what the invoked “epistemic state” is. Naturally we tend to call it knowledge, but this would be premature. Later on, I shall give it a more precise characterization, though in different terms.

For now, I propose:

(D1) A community C indirectly demands the institution-as-a-rule Inst iff (a) there is a possible state of affairs S that would benefit C; (b) members of C are in an epistemic state referring to S; (c) members of C positively value S; (d) Inst can contribute to S’s becoming a fact.

The mentioned states of affairs are taken here as objective, which means that whether or not some conditions in the world would benefit people is not just a matter of taste and subjective assessment – it is a mind-independent circumstance. That being said, these states of affairs cannot for instance float around in the domain of trivial abstraction and boil down to something like “economic development” or “growing GDP”. Of course, most of us would prefer to be well-off and not worse-off, but here it must be something more concrete. If so, people might not be in the relevant epistemic state linking them to these objective conditions. We can have a situation in which there are objective states of affairs justifying a certain demand, but those who should be interested in this demand are not able to get in touch with the relevant states of affairs. Why? What causes their ignorance? We can also have a situation in which there are no such objective conditions and yet a certain policy is still enforced. Why is that so? Again, what leads to such an unjustified enforcement? None of these questions are rhetorical.

To see first why this epistemological puzzle is significant, let us reconsider Chang’s (2011) argument from implementation failures, which is supposed to show – and it does so quite convincingly – that the rule-following view has a very real difficulty with local applicability. An appealing explanation for why a given rule remained a paper fiction in many developing countries or for why it – when finally implemented – stifled economic development in at least some sectors of their economies (as was the case in post-socialist countries with their ability to generate innovation and produce higher margins), may not be the alleged stubbornness of the local population, as a somewhat Western-centric view argues. It may instead be a more fundamental fact about how the epistemic pursuits of a given society are organized. Below I shall address the latter in terms of epistemic dependence.

All that being said, there is an assumption present in the rule-following approach that specific segments of the world, specific fragments of the acting individual’s surroundings, are recognized appropriately, meaning that the individuals are in the suitable epistemic state mentioned in (D1). Meanwhile, this may not be the case.

All of this is open to two different interpretations. One is that, as Chang (2011) suggests, economic development precedes the creation of an institution and gives rise to certain objective conditions, as well as creating groups of people interested in these conditions, and thus capable of being in the relevant epistemic state referring to them – standardly these are entrepreneurs or capital holders on the one hand, and (well-informed) workers on the other. Economic development must first produce the demand for institutions. Therefore, even if institutions contribute to further development, they cannot be a major factor behind it (see Greif, 1998, Glaeser et al., 2004, and Acemoglu and Robinson, 2013, 2019 for a more general discussion).

The other interpretation is that before institutions emerge, in the form of a system of rules, there are various earlier stages of institution creation or institutional development that do not boil down to rules, but dig deeper into how various actors on the market operate and what epistemic states they are capable of being in. Therefore, it is not the case that the demand for institutions comes first and institutions come second, for this very demand, and the relevant epistemic state, are possible only within some previous institutions. Institutions-as-rules, therefore, are answers to demands, but institutions of some other kind are demand-creators and therefore – crucially for us here – epistemically relevant factors. This paper argues for the latter option.

I therefore argue here that the epistemic state mentioned in (D1), as a necessary condition for a demand to come into being, is a function of some pre-existing institutional setup. And that this underlying institutional setup cannot be extracted in terms of rule-following, given the fact that there must be a demand for rules first.

However, some could question whether a state of affairs, supposing that its beneficial character is indeed objective, needs to be recognized as such in the first place. One could for instance question whether the state of being in good health requires any special act of recognition. I will address this concern below, using epistemic injustice as my point of reference.

2.4 (Social) Epistemology of Acquaintance

I propose to articulate the ‘epistemic state’ required by D1 in terms of acquaintance. The term famously comes from Bertrand Russell who writes:

I say that I am acquainted with an object when I have a direct cognitive relation to that object, i.e. when I am directly aware of the object itself. When I speak of a cognitive relation here, I do not mean the sort of relation which constitutes judgment, but the sort which constitutes presentation (Russell 1910, p. 108)

Speaking of acquaintance was supposed to postulate the most fundamental link rooting the whole construct of propositional knowledge in objective reality. In line with the latter, it must be emphasized that being acquainted may not always lead to full-fledged propositional knowledge. Think of the following example: when you come across a thing laying on the street and you don’t know at all what it could be, your discovery can be spelled out in the form of a question like “What is this?” or, as Wittgenstein (1953) might insist, even in the form of some less articulated utterance such as “Wow!”. Either way, you set up a certain perceptual link, and on this basis, some structure of acquaintance-based thoughts (Dickie’s, 2010 term) begins to emerge in your mind. The point is therefore that acquaintance develops, it has a history, and thus begins earlier than the layer of fully-fledged propositional knowledge. I shall come back to this issue, since setting up acquaintance by acts of questioning, or problematizing, proves crucial later on.

Now, all this applies also to social interactions and social structures. They also determine the way we acquaint ourselves with our surroundings. Therefore acquaintance, I presume, can also be targeted as a social phenomenon, in step with what social epistemology teaches (see e.g. Goldman, 1999; Goldberg, 2010; Lackey, 2014).

That being said, the crux of the proposal rolled out in this paper is that we can represent institutions as active factors in this social/epistemic dynamic of acquaintance. Namely, institutions are social enablers of acquaintance; they embody a certain allocation of epistemic capacities within a society.

One further note should be made though. I have just stated that institutions are social enablers of acquaintance, as if following the somewhat optimistic characterization given by Hindriks and Guala (2019). Meanwhile, institutions may also disable acquaintance, and in fact it is precisely a short analysis of these latter cases that should provide us with a sense of how important the problem of acquaintance at the social level is. Broadly speaking, these are examples coming from studies of various kinds of epistemic injustice, as it is called in the relevant literature (see Fricker, 2007; Kidd et al., 2019).

Fricker (2007) distinguishes two types of epistemic injustice: testimonial and hermeneutical (as for the issue of testimony see e.g. Pritchard 2004; Fumerton, 2007, Goldberg, 2010). Both are built upon prejudice against certain speakers because of their identity. First, we can speak of testimonial injustice when a group refuses to admit that someone’s testimony is credible due to prejudice against that person’s social identity. For example, someone’s testimony in court can be dismissed as not credible if that person belongs to a poor racial minority group, while all members of the jury belong to a better-off majority (on average). If the jury was selected that way, this can point to either the prejudice of the specific people who performed the selection, or to more structural, institutional factors, Fricker (2007) argues. In the latter case, the prejudiced selection is performed, but no specific person can be held responsible. In such cases, some procedures are so fully embedded in social practice and cognitive functioning that nobody even reflects upon their existence.

Although concentrating on the victims of such structural injustice is morally understandable, let us pay attention to those who unintentionally perpetuate the system (arguably, there are people who do this intentionally and purposefully, but they are not philosophically interesting; their motivation in the form of a specific ideology is usually apparent). The point being – and this is the reason why Fricker’s notion of testimonial injustice illustrates my point on acquaintance – that even if each member of the majority group may be morally capable of adopting a different, more equitable jury selection process, insofar as the example still applies, they do not recognize such a need, and therefore do not see it as something worth pursuing. We can therefore speak of a specific kind of acquaintance, namely acquaintance with some object or state of affairs as something morally desirable, and the example just given shows us that there may be social factors making it impossible for the individual to be acquainted with that morally desirable states of affairs. Note the subtle difference between institutional structures relating to epistemic factors such as acquaintance and structures of power. The point is that the former does not boil down to the latter. The majority group in the example is most likely the more powerful, or in many respects the privileged one. And yet, due to the institutional order they live in, they lack the privilege to be acquainted with the perspective of the less privileged (however paradoxical that must seem).

Further on, Fricker discusses hermeneutical injustice which is not so much about the credibility of this or that particular testimony, but rather about having or lacking the capacity to articulate something, especially some experience. Let us pay close attention to the following controversial point: one can suffer from something and at the same time – only adding to the suffering – be unable to spell out where exactly the problem is, so that the people who could do something about it, solve the problem, do not understand what is going on, or where the problem is.

Here is a nice summary from Anderson (2012):

Hermeneutical injustice occurs when a society lacks the interpretive resources to make sense of important features of a speaker’s experience, because she or members of her social group have been prejudicially marginalized in meaning-making activities (…). For example, prior to the introduction of the concept of sexual harassment into public discourse, people tended to interpret women’s discomfort, offense, and trauma at unwanted sexual advances at work and school as hysterical reactions to innocent flirtation, or as reflecting frigidity or humorlessness. Sexually harassed women suffered hermeneutical injustice because they lacked the interpretive resources to make sense of the injustice they were suffering, due to their prejudicial epistemic marginalization (… ) (Anderson 2012, p. 166)

Per the example just given, the credibility of the testimony of the harassed women is one thing, but an even more fundamental issue here is the victims’ lack of appropriate vocabulary to express their own experience.

That being said, I propose to distinguish two cases: one when the appropriate conceptual framework is just not there yet and where it is a structural and institutional issue as to what the division of epistemic labor should look like in order to create such a vocabulary; the second case is when the appropriate vocabulary is already there, but is not distributed equally in society. For the sake of illustration, we could think of members of the elite where the vocabulary is already in use and contributes to prevention, but due to the fact that that elite does not maintain any epistemic bond (the kind of bond it could be shall be examined later on) that would transfer these linguistic frameworks down the social ladder, the vocabulary remains exclusively within their circles. Therefore, the harm being done to the harassed women not part of the elite may not be articulated clearly even by these very women, even if they do obviously have an unarticulated feeling that something is profoundly wrong. And this means precisely that a certain beneficial state of affairs – in this case a state of affairs in which they are treated in a morally desirable way – is not perceived as such, i.e. is not problematized, singled out as an issue and therefore not recognized as something beneficial, or at the very least not clearly. For in order to see something as a beneficial state of affairs and then to demand it, one has to possess “interpretative resources”, as Anderson puts it, to articulate the relevant problem. In my vocabulary, this is an example of the inability (or limited ability) to acquaint oneself with a certain objectively beneficial state of affairs as a potential target of action and demand. This phenomenon illustrates my point that social structures not only enable people to get access to specific things and achieve specific goals; at a more fundamental level of social dynamics, they decide on acquaintance, or to put it simply, on the distribution of epistemic resources such as specific conceptual frameworks and other “interpretative” tools.

2.5 Problematization

These two examples of epistemic injustice provide the crucial concept I would like to build my proposal upon. Note that what Anderson (2012) calls “interpretative resources” is in other words a capacity to see something, some existing state of affairs as an issue, as something which needs to be addressed, thus as a problem. This is therefore the capacity to problematize certain parts or properties of our surroundings, including the social order we find ourselves in. We must realize that for someone who was raised in a society where women were not regarded as credible sources of knowledge, e.g. testimony, the very realization that this is a certain fact about that society, an attitude which could perhaps be altered, is quite a significant epistemic feat in itself. The case of harassment is similar – some kinds of behavior strike us as inappropriate once we realize that things could have been different, thus when we are capable of an appropriate act of problematization.

This last case provides a chance to answer the concern flagged earlier that there may be beneficial states of affairs, in which this beneficial nature does not rely in any way on whether or not someone recognizes them as such. For example, being in good health is a state of affairs we pursue without the mediation of any epistemic factors. First, recall my insistence that speaking of a relevant state of affairs must refer to something more specific; it should not be something like “growing GDP” and equally in this case, this should not just be “being in good health”, which stands for a loosely specified set of conditions. After all, what does “health” specifically refer to? But if we move towards better delineated conditions, but which still accommodate a wide range of cases, such as “avoiding excess cholesterol”, it seems clear that having such a demand requires at least some level of acquaintance with the objective fact that excess cholesterol is bad, and that this epistemic state can clearly be distributed unequally within society, or even be blocked from being common within society, e.g. by the actions of the fast food industry. Therefore, I would argue, this epistemic state also depends on social factors, including the crucial role played by institutions. Generally speaking, it may seem that since the fact that some general objective conditions are beneficial is transparent, it requires no extra epistemic or social enablers. Yet if we decompose these general conditions into parts, the steps in the realization of the general idea of “being in good health”, when we move down to more specific conditions, an extra epistemic and social enabler becomes a significant factor in deciding on who gets access to which conditions.

All that being said, in the third part of the paper, I shall develop a more systematic account of institutions as social enablers of epistemic acquaintance of a particular kind – as enablers of acts of problematization. This shall be done in terms of epistemic dependency and – more appropriately – in terms of a certain balance of epistemic dependencies.

3 Institutions and Epistemic Dependence

3.1 The Social Template of Epistemic Dependence

All the cases brought up earlier illustrate the existence of a certain social dynamic structure, which decides how and under what conditions it is possible (or impossible) for specific individuals or groups to take certain ingredients of their environments – be it things, information resources, states of affairs, but also capacities (e.g. being able to vote or to run a marathon), etc. – as potentially reachable targets of action and allocation of assets (money, energy, time, etc.). In other words, the puzzle we face is how it is possible for someone to be acquainted with something, as a potential denotation of a demand, which then translates into a demand for certain institutions-as-rules. This part of the paper elaborates on this problem in greater detail.

Going back to epistemic injustice, the question that comes to mind is what the main cause of it is, and a quite natural answer is “power”. There is a certain structure of power and influence in society, which determines who knows and does not know what, as well as who benefits from this setup. More generally, if we look through the empirical literature on institutional change and economic history, it seems that institutions are in large part exactly what makes these power structures what they are: institutions embedded in a given society, in its practices, habits, history, and in written laws, constitute and enforce certain power structures – that is their main job.

Formal economic models of institutional change, spelled out in terms of power structures, have been developed in recent years by Daron Acemoglu and his colleagues. For example, Acemoglu (2006) investigates the mechanisms behind inefficient institutions, and describes a power struggle between the people and the elite which may or – more frequently – does not have an interest in changing these institutions. What decides on whether institutional change takes place is first of all how much influence on the political process each of these two groups has, and that distribution can be decomposed into specific measures (see also Alesina et al., 1996; Alzarez et al. 2000; Acemoglu and Robinson, 2013; Acemoglu et al., 2021).

The distinction between the elite and the people seems rather rough and simplistic, but the fact is that from the times of Karl Marx (and Vilfredo Pareto), whether one was a Marxist or not, the distinction has been used in multiple variations by economists and social scientists. Of course, Marx’s own framework was spelled out in terms of the division between capital holders and workers, and it is still in use. In their most recent book, Acemoglu and Robinson (2019) come up with a similar distinction between the state and the people, where the former, in my reading, actually also means the elite, insofar as the “state” can be represented as a cluster of institutions manifesting, so to speak, the will of influential individuals (or in a dictatorship, primarily that of a single individual) and groups (wealthy people, sometimes aristocrats, higher members of bureaucracy, people from the media, etc.).

I shall also employ this basic framework, yet my key point is that it will not be a framework for power struggles. Recall the examples of epistemic injustice, but also the examples given earlier to illustrate the idea of a demand for institutions-as-rules. In all cases, there were certain power structures involved, quite obviously, yet what made these cases unique were their epistemic ingredients. Testimonial injustice is certainly caused by power structures, but the inability to recognize the credibility of the testimony made by an underprivileged person is not about power itself. This applies even better to hermeneutical injustice: sexual abuse is a pathology of power structures. And yet, the lack of conceptual tools on the side of the abused to articulate the abuse is not itself about power. In a different vein, going back to the Hamilton case, the demand formulated by the American “infant” entrepreneurs and directed to the government, was clearly an exercise of power, yet the content of that demand was an exercise of the entrepreneur’s epistemic capacities: they were able to recognize a certain state of affairs as beneficial and demanded institutions to be implemented so that the desired state of affairs became fact.

More generally, whether or not a certain demand for institutions-as-rules is really the case causally depends on the structure of power in a given society, but my point is that it also epistemically depends on what various groups within that society can or cannot acquaint themselves with. The key phrase here being “epistemically depends”. It refers to a certain social epistemic dynamic described by Lackey (1999); Goldberg (2010, 2011), Pritchard (2004, 2015), more recently Broncano-Berrocal (2020), among others.

I therefore use the elite vs. the people distinction, standard in economic analyses, as a template for a specific kind of epistemic dependence which shall be rolled out in the next section (see e.g., Jeffrey 2017 for a different use of the distinction related to knowledge and expertise). Then I shall define a properly functioning institution in these terms.

Some comments are in order before we move further. First, note that I did not say “I shall define an institution,” but instead “a properly functioning institution”. This means that the definition to be put forward establishes a certain theoretical fiction (for “properly” refers to a certain ideal state) which is in principle operationalized. Therefore, the definition does not say “Look, an institution is this or that,” but rather “Look, institutions, when working properly, do this or that” or in a de dicto manner “Look, speaking of institutions refers to such and such a function.”

Secondly, when I say I shall “define institutions” as social enablers of epistemic acquaintance, i.e., enablers of acts of problematization, a natural question comes to mind as to whether all institutions have this epistemic function built in. To answer this concern we would have to define the conditions of identity of institutions, which is beyond the scope of this investigation. The cautious answer is: for any institution, either it has the built-in epistemic aspect or ingredient which I am about to describe, or it depends on another institution that fulfills the said epistemic function. Dependence here means that the demand for the former was created by the latter. Perhaps it would make sense to distinguish a specific class of “epistemic institutions” underlying the whole institutional network, as already suggested, but I will not discuss this option here. To a degree I take it as a conventional issue. Let us consider an example: the institution of private property and all of its derivatives. If we consider it a formal institution, e.g. as enshrined in the Trade-Related Intellectual Property Rights Agreement (TRIPS), the epistemic aspect is not there. Instead, it is assumed that the awareness of such a thing as intellectual property already exists in the minds and the practices of agents on the market and the political stage. But if we take private property as a cultural artifact, crucial in some but not all cultures, thus as an informal institution, I’d say that it is a very complex network of norms and behavioral patterns some of which do fulfill the epistemic function of making it an issue whether a thing (and if so, what thing, under what conditions, etc.) or the right to use it can be attributed exclusively to one person. To take a different example, think of the above-mentioned cases of epistemic injustice. Now, we can take the judiciary system that does not recognize women as credible to be a single complex institution, and under this kind of description it certainly does have (as I already discussed following the literature) the said epistemic aspect built in. But if we divide this complex entity into a variety of simpler, more task-specific institutions, formal and informal alike, then it makes more sense to distinguish a specifically “epistemic institution” existing within that variety and doing the epistemic service to all the others, so to speak,

Eventually, the way we prefer to identify institutions may be contextual, thus dependent on what we need the given identification for. One way or another, the key point is that the institutional network as a whole does act not only as a social enabler of problem-solving but also as a social enabler of problematization.

Thirdly, we could doubt whether a distinction between the elite vs. the people is necessary as such for an institution. I would answer that, as mentioned previously, this is just one instance of a more general and fundamental distinction which seems to be present in any society; a distinction appearing under many names (e.g. employer – employee, state – people), and while neither of these instances seem necessary on their own, I would contend that the fundamental social division of epistemic labor which they all exemplify, is indeed necessary for institutions.

Finally, if both groups, the elite and the people, are conceptualized in terms of epistemic dependence, this brings to mind the distinction between ‘experts’ and ‘non-experts’, already flagged in the first paragraph, therefore also bringing up the debate on the role of experts, especially scientists in society (see e.g. Christiano, 2012; Douglas, 2021). It is important to keep these links in mind, since the expert vs. non-expert division is yet another rendering of the assumed, and more fundamental, division just mentioned. That being said, there is no room here to address this literature properly.

3.2 Epistemic Dependence and Questions

In this section, I shall first define a new kind of epistemic dependence and then portray it as a major factor determining institutions. To begin with, here is a general characterization of epistemic dependence from Duncan Pritchard:

Whether or not an agents’ undefeated epistemic support for her true belief that p counts as knowledge that p can significantly depend upon factors outwith her cognitive agency” (Pritchard 2015, p. 306).

This is a very general indication of situations in which what we know hinges on something other than our own epistemic or cognitive capacities. Sandy Goldberg provides a more fine-grained account and defines epistemic dependence of two kinds – direct and diffuse:

(ED1) “A subject S1directly (epistemically) depends on another subject S2 with respect to S1’s doxastic attitude D when the following condition holds: there are variations in S2’s epistemic perspective that would make for variations in the epistemic properties of D.”

(ED2) “A subject S1diffusely (epistemically) depends on her community C with respect to her doxastic attitude D when the following condition holds: there are variations in the practices of the members of C, and variations in the states and dispositions of those members, that would make for variations in the epistemic properties of D; and this effect remains even after we subtract the effects of any direct epistemic dependence S1 exhibits with respect to D.” (Goldberg 2011, p. 113).

Diffuse dependence is crucial from an institutional perspective. However, I would like to move away somewhat from the doxastic approach maintained by Goldberg, Pritchard, and the relevant literature. Instead, I adopt an erotetic perspective, thus one that puts emphasis less on beliefs, but more on the questions which these beliefs are answers to. It is not supposed to mean that beliefs are irrelevant; the claim is that focusing on questions goes deeper into the structure of our epistemic efforts. Some benefits of this theoretical decision shall be made explicit in due course.

To begin with, here is a classic notion put forward by Henry Hiż which frames the issue more generally:

Knowledge can be classified according to what questions it answers. (...) That scientific knowledge is systematic may be viewed as a result of systematically arranged answers to systematically arranged questions (Hiż 1962, p. 253).

There is a whole segment of philosophical logic devoted to questions called erotetics, but that is not too important for us here (see Hamblin, 1958; Belnap and Steel, 1976; Harrah, 1985; Wisniewski, 1995). As for epistemological matters, more recently this erotetic route has been taken up by Jonathan Shaffer. He gives an illustrative example:

Knowing whether there is a goldfinch in the garden or a raven is a relatively easy task of bird-identification. Virtually anyone (with decent vision and minimal expertise) can know whether there is a goldfinch in the garden or a raven. In contrast, knowing whether there is a goldfinch in the garden or a canary is a harder task of bird-identification. Perhaps only an expert birder will be able to tell the difference. (...) So one might well know whether there is a goldfinch in the garden or a raven, but fail to know whether there is a goldfinch in the garden or a canary (...) (Schaffer 2007, p. 388.)

Then, he proposes:

Here the knowledge relation is relativized to the question: KspQ rather than Ksp. KspQ may be read as: s knows that p, as the true answer to Q. It is not enough to know that p—one must know p as the true answer. In other words, it is not enough to know the proposition that just so happens to be the answer—one must know the answer as such (Ibid., p. 392).

Therefore, I propose to focus our attention on questions instead of beliefs. Here is the general reason for that decision (a more specific rationale shall be elicited later on). As Schaffer points out quite convincingly, two apparently identical beliefs may differ either in content or in what we might call epistemic significance. Someone’s true belief about a goldfinch contrasted with a canary requires more epistemic effort than someone else’s true belief about a goldfinch in the garden contrasted with a raven. The former contains more information than the latter. The difference between the two hinges on what question the person answered in a given situation. In other words, what was problematized by that person in their environment: a likely goldfinch contrasted with ravens, or a likely goldfinch contrasted with canaries.

Therefore, if one is to form a belief, one needs to perform an act of problematization of some fragment or aspect of one’s surroundings, and this problematization may differ when it relates to its acuteness, or resolution. This can be defined in terms of how close the concepts applicable to a given situation are in the relevant taxonomy – e.g. the concepts of “goldfinch” and “canary” are closer to each other than “goldfinch” and “raven” (another possible characterization of resolution shall be given shortly).

3.3 Erotetic Dependence, Acquaintance and Demand

A question can remain open or, in a sense, partly open when one does not know the answer, when one takes into consideration a certain group of possible answers, but has not decided on one yet and, finally, when in fact there are many true or satisfactory answers, depending on context and certain additional factors. In all of these cases we do not have a formed belief at a given moment. But by questioning or by problematizing a given state of affairs, the questioner is already in some epistemically significant bond with a part of reality; this bond already exists in the domain of that person’s epistemic efforts, which means that the person is in some way already acquainted with the state of affairs in question. This is the key point, going back to our previous considerations. For being acquainted with a certain worldly phenomenon is necessary to form a demand, including demands for institutions-as-rules.

As already noted earlier, I propose to define acquaintance, for the sake of this study, in terms of questions, not in terms of beliefs and knowledge. Knowledge is a solution to a problem, yet one needs to know the problem first. Therefore, one must be acquainted with some object or state of affairs in the world as constituting a problem before anyone can know anything about it.

Here is a very simple formulation of this postulate:

(ACQ1) A subject S becomes initially acquainted with a real entity x when S problematizes x, thus forms a question referring to x.

Here is an equivalent formulation, more in the spirit of Schaffer:

(ACQ2) A subject S becomes initially acquainted with a real entity x when S problematizes x, thus contrasts the actual scenario including x with a group of possible scenarios having x as a part.

Note that having a belief comes with consequences. Beliefs must lend themselves to examination, meaning that they are either true or false, and – quite crucially – a belief may seem true at one moment and turn out to be false a moment later. My point is that defining institutions in terms of epistemic dependency and extracting the latter in terms of beliefs would be unrealistic and I shall clarify the reasons for that later on.

Therefore, restricting my interest to diffuse epistemic dependence, I propose the following definition of diffuse erotetic dependence, based on Goldberg (2011):

(ED3) A subject S diffusely erotetically depends on their community C, with respect to their questions Q, when the following condition holds: there are variations in the practices of the members of C, and variations in the states and dispositions of those members, that would make for variations in the epistemic properties of Q (in the first place, their content); and this effect remains even after we subtract the effects of any direct epistemic dependence S exhibits with respect to Q.

Consider an example of family relations approached from this erotetic angle – suppose that you come home with a serious problem that occurred that day at work. You consider options a, b and c and thus ask something like “Should I do a, b or perhaps c” and the definite meaning of each of the options is determined by what they would imply. You are thinking that if you do a, then some further w will likely occur, but also some v, which is not what you want. And likewise for the other options. Now, as a result of the conversation you have at dinner with your relatives that afternoon, you come to the conclusion that v, which was not a scenario you welcomed, will lead to some t, and that t, might make v worth reconsideration. Therefore, your initial question “Should I do a, b or c” has a slightly different content now; it means something slightly different to you now, due to the fact that one of the options you took into consideration from the start is now seen differently through the lens of its ramifications, especially the one you did not take into account previously. If the family conversation draws your attention to this formerly unfamiliar scenario, this means that you now erotetically (to a degree, of course) depend on your family as a small communityFootnote 3. This community asked the same question, thus performed a relevant collective act of problematization, considered all the options, altered their collective epistemic state somewhat, by virtue of noticing the new scenario, and now you have changed your epistemic state somewhat as well, as a consequence. This dynamic satisfies (ED3).

Having the concept of erotetic dependence, we can do two things. First, we can give another characterization of what it means for a question to have a resolution:

(R) The resolution of a given question Q can be measured as the function of the number of possible contrastive alternatives in Schaffer’s (2007) sense, together with their consequences, attributable to Q.

Now, to tie together parts one and two of this study, we can finally come back to the central notion of demand:

(D2) A subject S demands w iff (a) there is a condition x in the world such that S is acquainted with x at least in the sense of (AQ1) or (AQ2), thus S is capable of problematizing x by means of a question Q; (b) w is part of at least one of the possible alternative scenarios involving x, i.e., w in a sense contributes to the resolution of Q; (c) S prefers (with varying strength) only the scenarios containing w.

The latter point requires some comment as to the relationship between questions, their resolution, and demands. For while the concept of a question belongs to epistemology, demands seem to belong to pragmatics. These are two different realms, although one can – and does – contribute to the other. The idea is that the epistemological approach focused on questions rather than beliefs contributes to the puzzle of demand. In other words, if we want to gain a deeper understanding of how demands come about, we need to consider the properly epistemological problem of questions. In short: to formulate a demand, one must be able to formulate a question or to problematize a specific segment of one’s surroundings first. Demanding, as proposed in D2(a), is grounded in the capacity to question/problematize. Specifically, the said capacity unveils possible scenarios; the number of these possible scenarios stands for the resolution of relevant questions; finally, this resolution, thus the number of possible scenarios, decides on how fine-tuned one’s demands are. This is because the realm of possible scenarios is the reference of one’s preferences (as D2(c) says). This, however, does not mean that one’s demand is to be identified with the resolution of the questions one asks. Instead, the former is possible thanks to the latter. Therefore, what we have here is a specific combination of epistemic and pragmatic factors.

Given all the consideration in part one above, we can now see the whole dynamic postulated here and express it in the form of a hypothesis binding institutions with erotetic dependencies in a given society, which means that the implementation of rules ultimately depends on how a society comes to ask the questions it is able to ask (what its problematizing capacity is). This hypothesis consists of: first, an outline of a certain stream of causal links between observables and, second, a simple application of modus tollens:

(H1) If (a) a fixed institution-as-a-rule Inst is successfully implemented (which is a matter of degree) in community C, (b) there must be a demand for Inst in C. If there is to be a demand for Inst, (c) there must have been a certain condition x in the world that the demanders in C got acquainted with, meaning that they were able to problematize x in the form of a question Q and recognize a certain setup of possible scenarios involving x. Finally, (d) the demanders’ capacity to problematize x is determined by the structure of erotetic dependencies within S.

(H2) Specific alterations in (d) are a predictor of specific alterations on the part of (a), meaning that when we manipulate the ways in which demanders diffusely depend on their community C in their pursuit of articulating Q, we shall obtain predictable changes in the degree of success of the implementation of Inst in C.

This hypothesis lends itself to empirical testing, and for that reason I lay it out here as an option, but not a strong conviction to be defended from my armchair. Though it would require a huge amount of work in economic history, broadly understood, to trace the correlations postulated in (H1) and either verify or falsify (H2).

This last caveat is important, since one could question whether the structure of erotetic dependencies I have just described is constitutive of institutions as such, or if it perhaps relates only to certain kinds of institutions. I do not want to determine the solution to this puzzle a priori, since I believe it to be an empirical, and not a purely conceptual question. Thus, H1 – H2 do articulate the claim that some exemplification of erotetic dependence is an ingredient of any rule-involving institution. This is because, as H1 stresses, for each rule there must be a certain cognitive act of recognition, which creates the demand for that rule. Otherwise, the rule remains a paper fiction (we then have a case of implementation failure) but not a de facto constraint on behavior. Yet given what H2 says, in principle, this claim is testable. Even if actual testing would be an extremely complex research project.

There is however more “armchair” work here, which could potentially lead to further developments of H1 – H2. The crucial puzzle is of course the “How” in “How does society come to ask the questions it is able to ask.” And here I would argue that this too is decided by institutions. The idea I put forward is that an institution, when approached from the epistemic perspective as described above, can be represented as a relatively stable structure of erotetic dependencies between the elite and the people.

3.4 Institutions in Terms of Erotetic Dependencies

To begin with, my proposal is that by knowing how a certain institution (in the proposed sense) functions, we should be able to describe how and when the questions attributable to the elite epistemically depend on the questions attributable to the people, as well as how and when this functions the other way around. In simple terms, the issue here is to what extent the elite erotetically depend on the people and how the people erotetically depend on their elite – put in yet a another way, when and to what extent the concerns of the elite are determined by the concerns of the people, as well as when and to what extent the concerns of the people are determined by the concerns of the elite.

To that end, I propose to think of a Cartesian space with two variables: time and a fixed measure of resolution of questions, as provisionally characterized in (R). This stands for how fine-grained the questions are in delineating different options or scenarios of action. In other words, resolution stands for how much a given society is able to problematize in its surroundings, thus how much information the questions convey.

Strictly speaking, there is one such system of coordinates for a fixed community and for a fixed set of logically correlated questions devoted to unpacking roughly the same problem. For more on how questions are tied together, especially about erotetic implication, see Belnap and Steel (1976) or Wiśniewski (1995). I assume here – thus take it for granted due to a lack of space to discuss it – that we can think of a certain stream of correlated questions as representing a development of a given pursuit of problematization.

Once the Cartesian coordinates are set, I propose that the functioning of any (epistemic) institution, considered in the manner sketched above, can be represented as a combination of two functions binding the timing of the pursuit of problematization with resolution: one line represents the erotetic dynamic of the elite (the red line in Fig. 1), thus the changing resolution of a given stream of questions articulated by the elite, and the other one stands for the erotetic dynamic of the people (the blue line), understood accordingly. This simple representation is to a certain degree inspired by Acemoglu and Robinson’s (2019) idea of a “narrow corridor” (of freedom), although their notion refers to a certain dynamic of power, in fact a power struggle and power balance between the state and the people.

Fig. 1
figure 1

The ideal structure of epistemic dependencies

Full size image

The crucial point is what these two lines have to do with one another. I want to portray something like an ideal institutional scenario. This is in step with the remark I made much earlier, that the key to the question of what institutions are is the question of what they do, and if we are to answer the latter, pointing to ideal functioning, even if clearly fictional, does the job.

Here is the key take away from Fig. 1: The “corridor” or “window” between points t1 and t2 stands for the period at which the elite is erotetically dependent on the people. Before that, between t0 and t1 we have a situation in which the people figured out that there is a particular problem they are faced with; they came to some preliminary articulation of that problem (thus got initially acquainted with the problem in step with AQ1 or AQ2), and then at a certain level of resolution, their line goes flat – they are not able to further increase the resolution of the questions on their own. At the same time, the elite is not interested at all: it is there and yet its line is down to zero. Thus, the balance actually begins at point t1. The space between t1 and t2 can be called the “corridor of trust,” for – I propose – this is the period where the elite depends epistemically on the people. Yet let us keep in mind that this dependence is of a very specific kind, namely it concerns the questions the elite is capable of formulating. The questions that the elite is capable of asking are determined by the concerns of the people. Note that here the resolution of the questions posed by the people is bigger than those posed by the elite. How are we supposed to understand that?

Before I answer this puzzle, let me first note the utility of the erotetic approach as distinguished from the doxastic one. This is the main reason why I prefer to focus on questions. It is not claimed that the beliefs of the elite are less fine-grained and dependent on the beliefs of the people, since this would undermine the very core of the former’s epistemic status – when it comes to experts especially, this would undermine their very expertise. After all, the key attribute of being a member of the epistemic elite, if I can put it this way, is to know more, simply put. The elite’s very being elite, when it is defined epistemically (this is different from financial elite or political elite, etc.), hinges on being skilled in highly demanding cognitive endeavors. On the other side, demanding that the beliefs of the people should be epistemically dependent on the claims made by the elite, including experts in the sciences, would be unrealistic. People invest their resources, especially their time, into different endeavors and therefore cannot learn particle physics or plunge into a careful study of Macbeth’s moral decline. The whole point of having a division of epistemic labor is to make the society effective as a whole without the ideal of everybody knowing and understanding everything. However – and this is my crucial point – the epistemic dependence of questions or problematizations is different. If we choose to speak of questions instead of beliefs, there seems to be nothing wrong with the idea that the issues that the elite end up knowing more about have been problematized by the people and, moreover, that the whole epistemic machinery attributable to the elite needs some time to warm-up and regroup, as it were, before it can attack the new problem head on. During this warm-up period, the elite’s questions are not sufficiently fine-tuned. Meanwhile, the people might have faced the problem for quite a long time before the elite took notice of it, and therefore their conceptualization of that problem, even if spelled out in everyday terms, perhaps by means of metaphors and various stories, may be able to grasp its more salient elements. Let me stress that the resolution of the question, thus at the end of the day, the resolution of the cognitive effort of problematization, is different from the resolution of beliefs, theses, claims, etc. In the final analysis, the idea that the problems that the people are concerned with should also draw the attention of the elite means nothing other than the quite commonsensical requirement that the elite should not be decoupled from the rest of society; on the contrary – they should answer some of the real needs of that society, even if their further development goes far beyond what the society at large can cognitively grasp.

On the other side, the erotetic dependence of the people on the elite, does not impose on the former the unrealistic and counterproductive demand of acquiring expert knowledge or exceptional sensitivity; it is a much more imaginable scenario in which the society at large, thanks to some reasonable effort put into becoming familiar with what (and how) the elite is questioning (an effort that does not disrupt other functioning), can have a better grasp of the complexity of the world, which is a valuable thing in itself. First and foremost, it is something that provides a better overview of possible scenarios of various decisions they may make in their lives. For example, being aware of how economists these days problematize the problem of inflation, does not require non-economists to master the mathematics behind the relevant models, but nevertheless can influence the latter’s investment decisions regarding things such as buying a home, stocks, cryptocurrency or bonds.

A good example comes from the literature on what is called “epistemic de-colonization”. In Hall and Tandon (2017), Rajesh Tandon brings up a personal story, harking back to the 1970s when, after he had obtained a science diploma from a British university, he went back to India; he would meet people in Rajasthan who, despite lacking any kind of formal education were, in his assessment, more knowledgeable regarding local agricultural techniques and methods than he was. Inspired by this, he would go on to create an organization called Society for Participatory Research in Asia (PRIA), whose goal was to preserve and implement the indigenous people’s practical knowledge and enable them to contribute to addressing local concerns such as agriculture, water supply, healthcare, etc. PRIA still exists. My reading of this is that the indigenous people initially had a much better recognition of their own local concerns than even the best educated university graduates. This means that they asked better questions – problematized their local environment much more acutely. Now, while it would be controversial to hold that the proper scientific grasp of the situation and then the solutions proposed in scientific terms will not exceed the indigenous ones at some point, there should be no doubt, to my mind, that the initial state of the art in terms of questions favors the local perspective.

Going back to the proposed scheme, we can say that the corridor or window – the “corridor of trust” – between t1 and t2 is the time when the elite is in a sense nourished, so that their epistemic grasp of things can at some point go up in terms of resolution and outperform the people’s epistemic efforts; they are nourished quite literally, for this is also the time when developed societies fund research, via public agencies offering grants, without knowing whether it will produce any benefits (which is why it is called the corridor of trust).

Now as already alluded to in the remarks about PRIA, beginning with t2 up to t3 the resolution of the questions articulated by the elite rapidly goes up, exceeding those attributable to the people, which means that experts are finally warmed-up and have supplied their epistemic “armed forces” with some new equipment and are finally ready to attack the problem with all the sophisticated capacities at their disposal. From that point on, the erotetic dependence shifts, and now the people depend erotetically on the elite. In practice this means, in an optimal scenario, that there are various means by which the high-resolution problematizations of the issue that the elite come up with are transmitted to the people, so that the latter can inform their acts of problematization with some pieces of expert knowledge and more precise concepts (this does not have to mean scientific knowledge; our conceptual capacities owe much to writers, artists, thinkers, also belonging to the elite), as a result of which the people can also experience a period of a more or less rapid, noticeable and potentially consequential increase in terms of the resolution of their problematizations. Thus the space between t2 and t3 may be called the corridor of “Enlightenment”, for it is something like what the Enlightenment ideologues imagined, yet rearticulated in more realistic terms. Finally, beginning with t3, both lines, each on its own level (some could argue that ideally the people’s line would reach the height of the elite’s, which is perhaps doable in some cases, but rather unlikely in the general ideal scenario) go flat, meaning that nothing more at a given moment can be done epistemically with the problem at stake (the dynamic just described can be repeated, with some changes, later on; it is an open issue which I do not address here).

Finally, someone might ask how all this relates to institutions. The answer is simple: this whole story is about a certain balance of erotetic dependencies, and this balance is kept in place and determined by (epistemic) institutions. Put strongly, it is one of the features of an institution that it is responsible for creating and maintaining a certain balance of epistemic, especially erotetic, dependencies. Assuming provisionally that (H2) is true, institutions are enablers of epistemic acquaintance with certain segments of the world, therefore they make it possible to recognize these selected states of affairs as beneficial and therefore, in step with (D1) and (D2), to formulate demands – including demands for institutions-as-rules.

Therefore, the functional representation depicted in Fig. 1 can also be understood as a structure of demands: it shows how, in an ideal scenario, certain demands resulting from the initial problematization of a given issue by the people are answered by the highly-skilled epistemic efforts of the elite, and this in turn makes the people capable of problematizing the issue in greater depth, and thus articulating even more demands. This is at least a significant part of what we mean by a well-informed public.

The schema portrays, as I mentioned, the ideal scenario of the balance of erotetic dependencies. It is clear that in reality we meet different scenarios of these two lines going in different directions; thus, various deviations from the ideal setup. We can mention some of them briefly, for paying attention to these deviations should tell us that being close to the ideal balance is not something that comes about effortlessly – there need to be some factors that “constrain and enable”, as Guala & Hindriks (2019) nicely put it, the society to keep as close to the ideal as possible. Economic history teaches us that this closeness is a relatively rare condition in the real world (e.g. in addition to Hall & Tandon 2017 see also Pitts 2017 or Haleta 2018 referring to epistemic (de)colonization). Therefore, thanks to studying deviations, we should acknowledge the institutional aspect of the story being told here: institutions are balance-keepers (or imbalance-keepers, when it comes to poorly functioning institutions).

In this context, someone may argue that the transmission of high-resolution questions from the elite to the people is possible only when the former recognizes an obligation towards the latter, which means that the elite is trustworthy. If so, then the “corridor of Enlightenment” is also, in some sense, based on trust. Perhaps the whole structure needs trust as the most basic social “glue,” as proposed by Gallagher and Petracca (2022; see also Hawley, 2014). This is because “in institutions, trust can be externalized to organizational arrangements that are reliable indicators of behavior” (Gallagher & Petracca, 2022, p. 11). This, however, merits a separate discussion.

So, one deviant scenario, perhaps when there is not enough social “glue” to make the system function properly, is when the elite’s line goes up while the people’s line stays flat or even declines. This is the basic schema of epistemic injustice as described above. The injustice consists of the fact that the elite do not pay back for the phase where they were nourished and enabled to develop their skills. The elite now refuse to raise the epistemic condition of the people. The latter, I propose, should not be conceptualized in terms of the elite’s maliciousness, for such a conceptualization (even if to a degree accurate, provided that we can “measure” this moral condition at all) does not help in making the situation better. Instead, it is much more practical to conceptualize this pathological state in institutional terms. This means that there must be some local, embedded institutional factors that determine such a far-from-ideal dynamic of erotetic dependencies.

Another possible scenario of deviation from the optimal setup would be a situation in which the elite are incapable of producing high-resolution problematizations, and therefore their line goes flat before it reaches t2 and perhaps even goes back to zero. This portrays a society in a state of epistemic-institutional despair, either perpetuated generation after generation due to insufficient institutions, or resulting from a particular catastrophe, such as war or natural disasters.

Finally, a very interesting scenario is when the people’s line goes up while the elite’s line remains flat below – how is this possible? Remember that each such functional depiction is limited to a particular society, therefore some deviations may also point to external factors. In the situation just described we could have something like the importation of expertise, thus a situation in which the people are quite knowledgeable, but this does not result from the effort of the local elite, but from purchasing expertise from external sources present on the market. At least in principle, this can happen in the case of epistemic colonization. Since the colonizers may need a competent workforce, for example, while keeping these quite knowledgeable workers barred from sources of actual expertise and the possibility of developing expertise on their own.

3.5 Supplementing the Integrative Approach

In this final section of the paper, I shall deal with a possible concern as to what additional value the question-based conception presented above brings to the table when compared with existing amendments of the rule-following view. It could be argued that the fact that this focus on rules is somewhat too abstract and needs to be made more applicable to local instances of human behavior is nothing new.

Hindriks and Guala’s (2019) proposal can be seen in this light as well They put forward an integrative account of institutions, combining the rule-following view with the equilibrium approach. According to the latter, championed by Aoki (2001, 2007) and originating in the context of game theory (see Binmore, 2010, Hedoin, 2017), institutions emerge as solutions to coordination problems within groups of actors or players, such that each individual seeks to achieve their own goal. Each player is aware of other players and has a certain expectation as to what they will do, what their next move will be. In each of these situations, which may be represented as a game, there is a possible state of equilibrium of expectations (formally represented as Nash equilibrium; see Binmore, 2010; Greif and Kingston, 2011; Hindriks and Guala, 2015): for example, in car traffic each driver wants to get to their destination as quickly as possible, and each of them has certain expectations as to what other drivers do. A state of equilibrium of expectations results in a specific organization of their movements – e.g. all drivers going in one direction stick to one side of the road.

Hindriks and Guala rightly note that whether we take institutions to be inscribed in the form of rules governing behavior, or realized directly in the form of a certain pattern of behavior, the objective is – or at least may be thought of as – the same. It is the improvement of a society’s efforts to achieve its goals through cooperation. This last point actually explains the very emergence of institutions and their persistence, in other words their etiology. We read:

The claim that institutions constrain and enable can be used to shed light on the function of institutions. As is widely accepted, institutions enable societies to cooperate in ways that benefit all or at least many of their members. We argue that the function of an institution is to promote cooperation and thereby to generate cooperative benefits. (Hindriks & Guala 2019, p. 2, preprint)

With the integrative view already available, one might ask what the advantage of the erotetic approach proposed here is. To begin with, note one subtle aspect: the concept of cooperative benefits assumes that the members of a society, thus the players in the cooperation game, have already defined their preferences. In other words, the benefits that cooperation is supposed to achieve, in one way or another (either spontaneous equilibrium or enforced rules), are already familiar to the players. In simple terms, this implies that the players are somehow already acquainted with what they want. Yet I would argue that we need to be able to say more about how institutions create conditions for the very emergence of preferences and other relevant cognitive artifacts, such as beliefs.

That being said, when someone asks why the erotetic approach is better, the short answer is: it is not supposed to be better. Instead, assuming that Hindriks and Guala (2019), or any similar integrative conception which sees institutions as solutions to coordination problems determined by preferences, is the correct approach, my goal here is to supplement it with an account of where these preferences and problems to be solved originate. In other words, while it is perfectly reasonable to see institutions as cooperative solutions to shared concerns, they should also be seen as problem-selectors or enablers of acts of problematization. The latter means, as stressed above, creating conditions in which members of a society can problematize certain aspects or properties of their surroundings – select states of affairs that become a concern. For example, it takes a cognitive effort to even notice the issue of testimonial injustice which women suffer from. Only once this problematization has been achieved can an institution emerge to address the problem.

In other words, I argue that the problems that can be tackled, as well as goals which can be pursued by means of institutional compounds composed of coordination + rules are not merely there, waiting for us to address them; in many cases, it requires significant effort to “distill” them from the immense vortex of states of affairs which we are embedded in. Accounting for institutions in terms of erotetic dependencies is therefore a step towards doing justice to this fact. Since, as I argued throughout this paper, the very capacity to select or “distill” the problems to be tackled can also be a function of institutions.

To illustrate the erotetic approach from yet another angle, this time without invoking entire societies, but using better delineated micro-conditions, so as not to leave the reader with the impression that this whole theoretical construct is somewhat too abstract, let me conclude with another empirical case – the functioning of the workplace. The latter, insofar as it boils down to coordination and rule-following, can be intuitively considered an institution.

More specifically, I believe that a particularly apt illustration of the erotetic dynamic behind institutions, indeed a kind of laboratory of epistemic institutional design, is provided by the phenomenon called “employee voice”, which is “the latest in a long line of terms used to describe employment practices designed to allow workers some ‘say’ in how their organizations are run; previous variants include worker participation, industrial democracy, employee involvement, and empowerment” (Marchington, 2008, p. 231). It is widely acknowledged that this concept has a very broad meaning and lacks a single commonly-accepted definition (see Lewin 2010; King et al., 2021). It can refer to “any form of delegation to or consultation with employees” (Wilkinson & Dundon, 2010, p. 168) or much more specific entities and practices such as quality circles or problem-solving groups (see e.g. Handel and Levine, 2004).

Marchington (2008) distinguishes three types of direct worker voice (as distinguished from a more traditional indirect form most exemplified by unions):

These are task-based participation, such as redesigned work operations, teamworking, and self-managed teams; upward problem-solving techniques such as off-line teams, quality circles, suggestion schemes, and worker input into briefing groups; and complaints about fair treatment, such as grievance procedures, speak-up programs, and whistle-blowing. The first two of these (…) are explicitly aimed at ‘adding value’ within the context of organizational goals. They are designed to give workers a chance to contribute to managerial decision-making, either in their day-to-day work or through formal and managerially instigated processes that tap into employees’ skills and ideas (Marchington 2008, pp. 232–233)

There has been debate in the relevant economic literature on whether worker voice does indeed increase productivity (see Appelbaum et al., 2000; Handel and Levine, 2004; Levin, 2006), but this issue is beyond the scope of this paper. Based on how extensively analyzed this phenomenon is (for an overview, see e.g. Wilkinson et al. 2004; King et al., 2021), and how well-established and historically deep-seated its various forms are (see e.g. Hirschman, 1970), I have assumed that the overall concept of worker voice, whether referred to by this term or others, is a solid example to build upon.

I propose that upward problem-solving techniques (as distinguished by Marchington, 2008) in particular provide a good illustration of the epistemic, and specifically erotetic, dynamic described above. The core idea behind this concept is that the people on the ground, dealing with specific processes on a daily basis, are best positioned to spot the specific components of a process which could be improved. This usually takes the form of problem-solving groups which “comprise small groups of workers who meet on a regular basis to identify, analyze, and solve quality and work-related problems” (Marchington, 2008, p. 237). Once a problem is identified and solutions for it are devised, the next, critical step is for them to be sent up the ladder to the managers. Interestingly, we read that “critics view these practices as problematic precisely because they encourage employees to collaborate with management in helping resolve work-related problems” (ibid., p. 236), which would suggest that this collaboration is constitutive of the process itself. In my vocabulary, this would translate to the epistemic dependence of the managers on the workers.

Note one subtlety: referring to these groups as problem-solving, emphasizes the solution aspect, which is understandable from a practical point of view. Yet if we were to emphasize the identification aspect, they could also be called problem-selecting groups, and from this perspective the managers are not only epistemically dependent – in the most generic sense – they are erotetically dependent. The latter implies that thanks to the workers, the managers are able to problematize certain aspects of the production process, service, or indeed any other type of work.

A classic example of worker voice, and especially of the process of upward problem-solving, comes from the Japanese technique known as Kaizen, developed as part of the Toyota Production System (see e.g. Liker, 2020; Wada, 2020). Here is a particularly telling fragment of testimony from a manager working in the famous NUMMI (New United Motor Manufacturing Inc.) plant in Fremont, CA, which was a joint venture between Toyota and General Motors (operating from 1984 to 2010):

In most other plants I’ve worked in, the issue is not coming up with employee suggestions – workers always have lots of interesting ideas on how things can be done better. The real problem is that the maintenance and engineering people never give worker’s suggestions a high enough priority (…).

In stamping, we’ve had some fantastic kaizen efforts. We’ve got a team working right now on reducing our downtime ratio on the presses. It’s amazing to see these workers doing statistical analysis od two hundred or three hundred downtime incidence. Some of these guys didn’t finish high school. (…) Now they’re putting together Pareto charts (Adler 1993, p. 138).

Again, someone could object that there is no need for specific epistemic efforts, since if there is a broken part in a machine, or rain gets into the building through a hole in the roof, no special cognitive act is required to spot it. However, the Point Kaizen strategy, which boils down to rapid improvement, addresses such obvious cases. While more systemic and complex exemplifications of this technique (see Wada, 2020), are not designed to deal with obvious cases. On the contrary, if I understand correctly, the very purpose of engaging workers in the process of continuous improvement, which is otherwise fully determined by engineers and designers in a top-down manner, is that those who are closest to the production work itself, are more likely to notice something which is less obviously incorrect, but somehow still derails or slows down the overall process. This means that Kaizen, in my vocabulary, begins with an act of problematization performed by employees, who correspond to “the people” within my schema. This act of problematization selects a particular state of affairs which requires improvement, and therefore constitutes a certain demand. This problematization is then brought to the managers, who in this context, stand for “the elite”.

However, let me underline again that what is delivered to the managers is not a belief, a portion of ready-made knowledge, simply because there may be no solution to the problem spotted at this stage. In a relatively simple case, what is delivered can be represented as a question, or set of questions. Or in a more general manner, as content, which refers to the existing, as well as to the demanded, states of affairs. It can have any grammatical or logical form imaginable, the point being that it is the product of an act of problematization, and not a set of beliefs. Therefore, the information provided is of a genuinely erotetic nature, and as already emphasized, this makes the managers erotetically dependent on the workers.

Given this, another important element in the cited testimony is the fact that the managers do not simply address the problem themselves. Of course, they may do so on occasion, but that is not the point. Instead, the testimony refers to a situation in which “the elite” gives something back, providing relatively sophisticated tools to deal with the problem, so that the workers can do it themselves. As I claim, this signifies that we have a two-directional epistemic dependence where, as just shown, at least the upstream transfer of epistemically-relevant content is of an erotetic nature. But the downstream epistemic dependence can also be erotetic, for what the workers are given may not be a ready-made solution to the problem, but rather, per the quote, a more sophisticated toolbox enabling them to re-articulate the problem – which means that their erotetic pursuits are improved – and then to address it. This is perfectly illustrated in the quoted fragment by the anecdote about the workers using relatively advanced mathematics.

All that being said, I believe that a well-functioning two-directional structure of upward problem-solving, which is a sort of worker voice or worker participation institution, exemplifies the dynamic of erotetic dependence, schematically depicted in Fig. 1. We can clearly distinguish a “window” in which workers problematize a certain issue and deal with it on their own, while also transferring it to their managers – making the managers erotetically dependent on the workers – and we then see a “window” through which managers then give back, providing more advanced tools to tackle the issue.

This case in particular deserves further discussion, since while it functions well as an exemplification of the erotetic approach to institutions, it may potentially also be useful in carrying out empirical research dedicated to this conception, assuming that experimental philosophers would be interested in testing the hypothesis H1 – H2 provided above or other, more detailed properties of the mechanism of erotetic dependence, perhaps also with the participation of researchers studying management. I believe this could provide new and potentially fruitful fields of inquiry.

4 Conclusion

The aim of this paper was less to propose an alternative to the two existing theories of institutions – the rule-following view and the equilibrium view – but rather to elicit an alternative perspective, from which institutions can be tackled and conceptualized. The critical remarks made in the first part of the paper with respect to the rule-following theory were intended to reveal a tacit and oft-neglected factor – the fact that each demand for an institution requires a very specific epistemic effort, and that this effort can also be a function of institutional order in a given society. I believe that this epistemic, erotetic approach can be particularly beneficial when we attempt to address, and solve, problems which are of a more systemic nature and which relate to the allocation of knowledge, competence in problem-solving, or credibility. These problems are addressed in the philosophical literature, e.g. in terms of epistemic injustice, but also are also discussed, from an entirely different angle, in the literature on management and the workforce.