1. Overview
People develop and deploy epistemic norms – normative sensibilities in light of which they regulate both their individual and community epistemic practice.Footnote 1 Individuals’ normative epistemic sensibilities lead them to respond to evidence in similar ways, to seek evidence in related (even coordinated) ways, to share what they commonly each will recognize as relevant evidence and considerations, and to join in a kind of cooperative evaluation of the worthiness of beliefs produced. There is a similarity to folk's epistemic normative sensibilities – and it is by virtue of this that folk commonly can rely on each other, and even work jointly to produce systems of true beliefs – a kind of epistemic common good. Agents not only regulate their belief forming practices in light of these sensitivities, but they make clear to others that they approve or disapprove of practices as these accord with their sensibilities – they thus regulate the belief forming practices of others in an interdependent pursuit of a good.
Such general observations led Graham and Henderson (Graham Reference Graham, Henderson and Greco2015; Henderson and Graham Reference Henderson and Graham2017a, Reference Henderson and Graham2017b) to argue that common epistemic norms are social norms. Developing their case, they drew on an analytical framework provided by Cristina Bicchieri's (Reference Bicchieri2006) discussion of various kinds of norms. On this account, norms generally are rules (or normative sensibilities) widely shared within a community by which agents within the relevant community respond to social choice problems – choice situations in which the results to be gotten from one's choices or practice depends on the choices or practices of others in an interdependent community. I continue to use this framework here – together with an important elaboration in Bicchieri (Reference Bicchieri2017) – as it affords an analysis of the various related ways in which normative sensibilities function in communities of interdependent agents. It allows one to distinguish the somewhat differing character of the motivations in play across classes of choice situations confronted in human epistemic life. The framework allows one to probe how these normative sensibilities function in the various associated choice situations.
Bicchieri does not aspire to an explication of the everyday concept of a social norm – and, in any case, it is quite plausible that there are broader and narrower usages of the terms “norm” and “social norm.” Rather, she advances an analysis of several distinguishable ways in which normative sensibilities function in the regulation of behavior. What she terms “norms” involve sensibilities functioning in one range of ways, and what she terms “social norms” involve normative sensibilities functioning in a yet more specific way. Acknowledging the technical character of her analysis, I will write of normsb and of social normsb. I argue that sometimes epistemic normative sensibilities function as social normsb – and sometimes they function in ways that are not characteristic of normsb. Specific normative sensibilities function as normsb to the extent that the concerns of agents (in contexts) do not (saliently) have them confronting a social choice situation (a coordination or cooperation game). In contrast, non-normb normative sensibilities commonly serve as the agent's representation of state-of-the-art ways of pushing forward what may be individual projects – ways that are at least to some degree serviceable independent of what others may do. Insofar as such sensibilities are more or less shared in a community, one might call such normative sensibilities customary state-of-the-art sensibilities, or customs (the latter is Bicchieri's term, and I here write of customsb). Such sensibilities would often be socially acquired and transmitted, but their salient motivational character (in context) would be distinct from that associated with social norms – as conformity to the norms would there not be conditional on expectations concerning the conformity of others.
To the extent that agents’ preferences to conform to a normative sensibility turn on expectations that others likewise conform (on “descriptive expectations”) and/or on expectations that other's normative evaluations and reactions are keyed to such sensibilities (to “normative expectations”), they are functioning as normsb of some stripe – they then yield coordination or cooperation. On Bicchieri's analysis, social normsb are normsb that are responsive to one class of social decision problem – cooperation games. In such contexts, also termed mixed-motive games, agents have some motivation to conform – thereby contributing in producing the results of cooperation, and some motivation to defect so as not to pay some cost that comes of cooperating. They have mixed motives with respect to the rule or sensibility. But normative sensibilities may function to regulate an individual's behavior while functioning as customsb. To take an epistemic illustration, insofar as the agent simply prefers to use what are commonly understood to be high quality, refined and reliable, cognitive processes when forming beliefs no matter what the agent expects the other chumps will do, that agent's normative sensibilities there function as customsb.
I here use Bicchieri's framework, as it provides a powerful analytical tool for thinking about some ways in which epistemic normative sensibilities function. I argue that epistemic normative sensibilities commonly function as social normsb regulating belief fixation in what amount to cooperation games. Yet, I also argue that these normative sensibilities also commonly serve to represent customary state-of-the-art ways of furthering individual epistemic projects. Further, one and the same set of epistemic normative sensibilities can function in both ways within an epistemic community in which agents are somewhat diverse in their motivations and situational contexts – functioning for agents as customsb in some contextual situations, and as social normsb for agents in other contexts. Human epistemic life confronts the typical epistemic agent with contexts of both sorts.
After further discussion of normsb and social normsb (section 2), I provide the central lines of an account of how epistemic normative sensibilities function as social normsb (section 3). I then characterize how epistemic normative sensibilities commonly function as customsb (in section 4). I go on to develop the view that one and the same normative sensibility, one and the same epistemic norm, can serve both as folks’ customary (state-of-the-art) way of pursuing their individual good and as a social norm. In developing this point, I draw on and recast a suggestion in Bicchieri (Reference Bicchieri2017) to the effect that a rule may transform from a customb to become a social norm, and transform back again. My suggestion is that one and the same rule may serve in both ways within a community as folk in that community confront contexts with different salient costs and benefits. Finally (in section 5), I take inspiration from the work of Michael Tomasello and his collaborators treating of the way in which normative sensibilities commonly arise as a socially situated “way in which it is (to be) done.” I suggest that, insofar as agents are then motivated to be a contributing part of a productive epistemic community, to play their part and think of themselves as owing this to others – and insofar as such thinking is a deep part of how such sensibilities function, there is a real sense in which these sensibilities are fundamentally social norms, … and a real sense in which they are fundamentally customsb.
What is intended by saying that epistemic sensibilities are fundamentally social norms, or fundamentally customsb – or fundamentally both? One way of thinking about what is fundamental here is to think about what motivational profiles associated with these sensibilities are not derivable from others. If one profile is significant, and not derivable from the other, then it counts as fundamental in the relevant sense. On this picture, normative epistemic sensibilities might be fundamentally both. The idea of something being fundamental is of course related to the idea of a foundation – and this suggests the idea of an arch, which has two foundations. I suggest that our epistemic norms have two foundations – the veritistically-centered concern of the individual agent to employ state-of-the-art practices of belief formation for themselves, and the concern of the agent for a veritistic community good together with the desire to contribute to that good and to not have others undermining that project by not doing their part.
My project here is to use a powerful framework for thinking about normsb and customsb to sort out the fundamental ways – plural – that our commonly shared epistemic normative sensibilities function in human epistemic life.
2. Norms, and social norms in particular
In the framework developed by Cristina Bicchieri (Reference Bicchieri2006), a normb is a “behavioral rule” – something on the order of a representation in (many of) the minds of the agents within the relevant population. It represents a way of acting or reacting.Footnote 2 Social normsb are held and applied in a way that is conditioned by expectations for conformity on the part of others in the relevant community. Such normsb (together with expectations had by members of the relevant population) afford an explanation of some regularities in behavior or practice found in the group or population.Footnote 3 Bicchieri's framework encourages one to think about how and why communities of interdependent agents would develop shared normative epistemic sensibilities, facilitating epistemic cooperation, issuing in behavior furthering individual and community goods.
Bicchieri's approach is rooted in economic thinking regarding cooperation and coordination in social games – situations in which the consequences of an individual's choice will turn on the choices made by others in the choice situation. Such an approach provides a fruitful framework for thinking about just what various kinds of norms “do for folk” – and plausibly, what motivates people in conforming to such sensibilities. Bicchieri (Reference Bicchieri2006) distinguishes three kinds of normsb, which she labels “social norms”, “descriptive norms,” and “conventions.” In each case, the normb in question is said to be a kind of “behavioral rule” for which enough agents have an understandable preference for conformity. The preferences for following the normb are conditional – they are keyed to expectations that the agents may have: to expectations concerning whether sufficiently many others in the relevant population follow it, and commonly to expectations concerning whether enough others normatively evaluate others in terms of the rule. When individuals follow the rule, they do so in view of these conditional preferences and associated expectations.
Specific kinds of normsb arise in response to a specific kind of social choice situation. Kinds of choice situations are understood in terms of classes of decision-theoretic social games. For example, what counts as social normsb are normsb dealing with mixed-motive games.
Cooperation games are mixed-motive games. Here there is a good to be attained by cooperating – a good in which all members of the groups can partake to the extent that it is produced – and an individual cost to be paid in coordinating. In these games, the marginal benefit gotten by the individual from the individual's own contribution (cooperating) is less than the cost to the individual of the individual's cooperation. Examples include prisoner's dilemmas and public goods games.Footnote 4 In a public goods game, each player is issued a stake and can choose to contribute all or part of it to a public pool. For each unit contributed to the public good pool, the individual (and all other players) get back something less than a unit. At the same time, this return on each contribution is sufficient to ensure that if players generally do contribute, they all come out better off than they would were they not to have generally contributed. Thus, in the relevant games, there is a cooperation payoff and a temptation payoff. As a result of the temptation payoff, individuals are tempted to “free-ride” – to partake of the public good that the others provide, while not contributing themselves.
There are social games in which there are not mixed-motives – for example, coordination games. Here one's expected payoff is determined simply by whether folk happen to coordinate. If all drive on the right, or if all drive on the left, everyone's expected payoff is higher. If they do not so coordinate, everyone's expected payoff is lower. Suppose that folk generally drive on the right side. There is no temptation payoff to be gotten by driving on the left.
Bicchieri's taxonomy is as follows: Normsb is the genus – and these are the rules regulating social choice situations of various stripes. Social normsb are rules regulating mixed-motive games, aka cooperation games. Descriptive normsb have to do with situations that can be characterized as coordination games (rather than mixed motive games). Conventionsb are a special class of descriptive normb (Bicchieri Reference Bicchieri2006: 38).
Bicchieri's taxonomy will prove a useful springboard, but considerable care is needed when categorizing epistemic norms using Bicchieri's taxonomy. Application requires us to get clear on the choice situation confronting epistemic agents. To what extent do epistemic agents have mixed motives – and what are the character of such motives? To what extent are the costs and benefits dependent on what others in their epistemic community do – and thus to what extent do they face either a coordination or cooperation game? To what extent can epistemic agents be motivated largely by something like veritistic ends in connection with some inquiry – so that they there might plausibly be motivated more in the way characteristic of customs. Further, and this will be very important, it should be obvious that agents within an epistemic community may be somewhat variously motivated. Veritistic (or related epistemically central) concerns may dominate for some – and certainly for some in some contexts. Competing concerns may loom larger for others – and certainly for some in some contexts. One and the same agent may be variously motivated across questions and contexts. Thus, within one and the same community in a given period, different agents may confront choice situations differing in kind. Some may be motivated in the manner characteristic of customs, others in the manner of social normsb.
3. Why think epistemic norms function as social normsb?
3.1. Interdependence and normatively keyed partnering – these are diagnostic, but not definitory marks, of social normsb
Human communities are remarkably epistemically interdependent. We humans are informatively helpful and thereby come to share a stock of beliefs.Footnote 5 We leverage our common information as we coordinate in ongoing inquiries and as we weigh out alternative courses of action. Within a given community of epistemically interdependent agents, it is not surprising that individuals come to be somewhat alike in their personal epistemic normative sensibilities. Indeed, they would need to be – as epistemic communities would tend to fragment to the extent that there were within them deep and pervasive differences in epistemic sensibilities. Agents would tend to selectively rely on those others who had epistemic normative sensibilities, and thus epistemic practices, that were largely in keeping with their own. Minimally, such agents then do for one what one would do for oneself were one able to be more places at once. Only those who have evinced epistemic normative sensibilities on the same order as one's own could reasonably be treated as fully trusted sources of belief. To do otherwise would compromise the quality of one's own epistemic situation. If there are agents whose sensibilities are better than one's own, they do for one something epistemically better than what one would have done. In such cases, one expects the community of relevant experts to police their epistemic house to insure fitting quality control themselves – at least in that domain in which they are expert.
The similarity of epistemic normative sensibilities within communities does not result simply from patterns of selective partnering. Epistemic partners are as much made as found. Agents not only acquire information from others – beliefs – agents learn from others how to inquire.Footnote 6 One learns how to evaluate one's own and others’ learning – and one readily falls into so evaluating. One thereby regulates both one's own belief formation and one's reception of other's beliefs accordingly. Further, one regulates others as one makes clear to them when one thinks that they have not formed beliefs acceptably. Of course, one also sometimes expresses one's epistemic sensibilities in systematic instruction.
It will be important to inculcate the normative sensibilities in initiates, to be somewhat attentive to which folk fail to conform to these, to marginalize those who do not conform, to mark those who are particularly adept in their conforming practice. In so doing, agents are not merely insisting that others are “like us” – following the same fashions – rather, they are insisting that others coordinate and cooperate in the production of an individual and public epistemic good.
Thus, there are at least three strands in the dynamic by which agent's epistemic sensibilities come to be more or less shared.
The first is a matter of how normative sensibilities can be transmitted. Folk learn from others how to learn. This can be done by explicit instruction – one initiates new community members (we school children, instruct university courses, and oversee labs). One articulates general practices, and caution folk concerning common pitfalls. But, much significant transmission of epistemic sensibilities also may be managed by example rather than by explicit instruction. One evaluates one's own and others’ practices – and often enough one does so in a public fashion. One critiques one's own practices, the practices of others, and that of one's initiates (think, for example, of conference presentations, commentaries, and class discussions).Footnote 7 In all this one hopes to impart (and to learn) practices that have benefited from ongoing refinement – practices that are informed by past successes and failures within the community. Both in early epistemic training and in advanced contexts, people take the lessons and the more or less gentle nudges to heart – as I will say, they learn “how it is (to be) done.” The training that is associated with becoming a member of a scientific community or sub-community that Kitcher (Reference Kitcher1993) discusses is itself just an advanced stage of the training that is a part of becoming a member of more general kinds of everyday epistemic community.
The second component of the dynamic by which epistemic norms come to be shared turns on feedback from the world: some practices may produce more or fewer successes (or frustrations) than others. This can condition or inform the normative sensibilities within a community. There will be a tendency for epistemic practices to be informed by directly encountered, and by communicated, epistemic successes and failures – making for an imperfect tendency for sharing of the relatively successful practices and associated norms. We do not merely “get on the same page” with others in our epistemic community, we tend to get on the same informed page. The result is a kind of “cultural ratchet” given a general characterization by Tomasello:
Human artifacts and behavioral practices commonly become more complex over time (they have a “history”). An individual invents an artifact or way of doing things, and others quickly learn it. [The artifacts and the associated practices spread through the group.] But if another individual makes some improvement [in such artifacts and the associated productive practices], everyone tends to learn the new improved practices. [A different class of artifacts come to be common.] This produces a kind of cultural ratchet, as each version [each class of artifacts and practices] stays solidly in the group's repertoire until someone comes up with something even newer and more improved. (Tomasello Reference Tomasello2009: x–xi)
The third component of the dynamic by which epistemic norms come to be shared turns on selective association. If one has reason to believe that some agents within one's community do not form beliefs as one thinks fitting – in a way at least as fitting as one would seek for oneself – one would presumably hesitate to draw on the epistemic results gotten by those agents. This would amount to a graded cutting of epistemic ties with those agents. As a result of selective association, folk in a given epistemic community would come to have personal epistemic sensibilities that are similar.Footnote 8
While these observations carry some weight, there is reason to wonder whether they really call for treatment in terms of social normsb. To what extent do human epistemic agents in communities confront a social choice game – as would be characteristic of normsb generally? And to what extent are epistemic agents confronted with mixed-motives – as would be characteristic of social normsb? How should one understand the epistemic choice situations – the “epistemic game” – prompting the development of epistemic norms?
3.2. The epistemic choice situation, the narrow epistemic game
To get started, think of the epistemic game in highly simplified and restricted terms, as revolving around the traditionally understood epistemic good of having true beliefs. On this veritistic accounting, when an agent produces a true belief, that is a gain, and when the agent shares that truth with others, they gain (while the producing agent, still possessing the truth, is not veritistically diminished thereby). The production of false beliefs is a veritistic loss – and others lose when the false belief is shared. A failure to produce a true belief is a lost opportunity, at least when the agent could have produced that belief. Call the choice situation understood in terms of such veritistic accounting the narrow epistemic game. Footnote 9
Note that the above accounting supposes that another's production of a true belief is automatically a veritistic gain to others – that the transmission of beliefs is full and frictionless. This, in turn would obtain only insofar as communicating one's beliefs to others did not compete with one's own production of further true beliefs. Obviously, all this is pretty unrealistic.
The narrow epistemic game is, undoubtedly a drastically simplified accounting of the epistemic choice situation faced by agents individually and in communities, and we will soon need to move beyond it.Footnote 10 But, before doing so, I want to focus on ways in which, in its main outlines, the narrow epistemic game looks pretty different from a paradigmatic mixed-motive game such as the public goods game. In a public goods game, one contributes at an initial loss to oneself, and one's marginal return on one's own contribution is less than what one has contributed. One can come out ahead, provided enough others contribute similarly, but for any set of contributions from others, one does best by not contributing oneself. One has a motivation to contribute, and a motivation to free-ride on the contributions of others. But, let us suppose, as it is plausible, that epistemic norms call for processes that are effectively state-of-the-art ways of reliably producing true beliefs. When one then conforms to such epistemic norms, one produces beliefs that are likely true (modulo one's state of the epistemic art). Then, in contrast with what obtains in a mixed-motive game, when one contributes by conforming to these norms and by sharing that belief, it is not as though one does so at a marginal epistemic loss to oneself. One has the epistemic gains of one's production, gotten via conformity to epistemic norms – and one does not suffer a marginal epistemic loss by sharing. Unlike sharing food or money, an instance of epistemic sharing does not leave one with fewer truths for oneself. If the norm is indeed significantly truth conducive, conformity to it yields a personal veritistic gain. From the narrowly epistemic point of view – thinking solely in terms of veritistic value – one has no incentive to defect from norms for the production or sharing of beliefs.
Notice that I qualified the above points by saying that we were highlighting how, in its main outlines, the narrow epistemic game looks pretty different from a paradigmatic mixed-motive game. We will soon want to note some more fine-grained ways in which there might well be a kind of mixed-motive to be confronted even in the narrow epistemic game. Before doing so, let us draw this lesson from the above: if the epistemic choice situation were characterized along the lines of the narrow epistemic game, then there would be few mixed motives confronting epistemic agents individually or in community, and there would be comparatively little need for epistemic normative sensibilities to function as social normsb. Epistemic normative sensibilities would then seem to represent agents’ state-of-the-art customary way of reliably attaining true beliefs – and the motivation of attaining such would be fully motivating. There would be little call for according different community standing to agents based on their differing conformity to these sensibilities.Footnote 11 Once the epistemic sensibilities were learned, others could be presumed to conform to them whole hog. There would be no need to enforce the norms – by gossiping about the sloppy, thereby diminishing their standing in the epistemic community. No need to be sensitive to variations in conformity to norms across one's epistemic community. Now, the author of this paper cares deeply about veritistic gains and losses, and believes many others do as well. Still, all should recognize these consequences of thinking of the epistemic choice situation solely in terms of the narrow epistemic game to be unrealistic – the epistemic game is not the narrow epistemic game. Henderson and Graham (Reference Henderson and Graham2017a, Reference Henderson and Graham2017b) explore these issues at some length.
Now for a qualification that Kevin Zollman has helped me see: even were agents playing something very like the narrow epistemic game, they might confront something like a social choice situation insofar as the state-of-the-art best practices (rules) for individual epistemic agents might be somewhat different from what practices (rules) would produce the most shared veritistic goods within the community. For example, at the social level, it might be desirable that agents individually devote some moderate portion of their inquiry to exploring risky lines of thinking – theories that are not likely to pan out. Agents might desire that such alternatives be explored so as to ensure that they as a community do not come to be satisfied with a mere local theoretical maxima (see also Kitcher Reference Kitcher1993). At the same time, agents might individually prefer to devote their inquiry fully to only the most promising lines of thought. Thus, agents would seem to face a social choice situation with conflicting pulls.Footnote 12 Zollman discusses several further examples of what may be social choice situations confronted in even the narrow epistemic game. Of course, once one appreciates that communication of one's results can compete with one's own production of true beliefs, and that there is much friction and noise in the transmission of results, beginning to leave behind the narrow epistemic game, one finds yet more bases for mixed motives even when costs and benefits are veritistically accounted. But, human agents do not live by veritistic goods alone.
3.3. Temptations confronted in our epistemic lives – and our normative response – why epistemic norms need be social norms
It is plausible that one's epistemic norms are social norms responsive to a mixed-motive game. At the same time, it is plausible that the veritistic gains and losses that defined the narrow epistemic game loom large in this mixed-motive game. Indeed, it is our hypothesis that epistemic norms reflect state-of-the-art standards for how to further projects understood in largely those veritistic terms – but that, in those choice situations, in those projects, epistemic agents yet face what count as temptations that arise from a wider set of concerns. Effectively, the epistemic choice situation understood in terms of veritistic gains and losses is to be understood as embedded in a choice situation where the epistemic agent can be tempted to cut corners in various ways in light of competing considerations. There is nothing radical here. It seems common enough for folk to develop norms for the pursuit of certain kinds of projects or practices – ones in which those agents face concerns that are understood as extraneous to the central ends of those practices. Such wider concerns – such wider prospective gains and losses – may lead them to degraded performances in the domain on which the norms are focused.
So, the practice of forming (and sharing) beliefs – epistemic practice – might revolve characteristically around getting (and sharing) truths about the world. It might be essentially concerned with veritistic gains and losses. At the same time, agents involved in that project may on occasion and in situations find themselves pulled in conflicting directions. In addition to the to the competing motivations noted above (motivations that found their home with a veritistically accounted project that could be pursued with marginally different effects at an individual level and at a social level) there are two prominent classes of non-veritistic values that may compete with veritistic motivations:
Competing pursuits attractors. These are temptations centered on non-veritistic values and projects the pursuit of which would lead agents to at least episodically compromise in best state-of-the-art belief fixation. Desiring certain forms of entertainment, feeling the pull of certain social obligations or gains, spreading their genes, acquiring certain foods, or getting extra sleep, may lead agents to curtail, limit, and cut corners in what would be their epistemic state-of-the-art.
Content attractors. These are temptations to believe specific contents – content that is attractive for various reasons other than veritistically/evidentially probative reasons. Wishful thinking may lead one to believe that certain events are likely (or unlikely). Humans are attracted to denying the finality of death – commonly supposing an afterlife or a subsequent life in which justice is served. There are a range of self-affirming beliefs (I myself am much wiser, handsomer, and nicer than most others). Similar attractions apparently obtain when thinking about the groups with which one identifies. And apparently one finds it attractive to think that there is a just class of gods that exemplify moral good not fully attained by us humans.
Occasionally one may observe that your students show evidence of such non-veritistic motivations. Some may become distracted with things on their computer, and thus fail to pick up on the truths you are laying before them. Sometimes one finds evidence that they may not have exercised state-of-the-art due diligence in putting together their research projects.
4. One rule or normative sensibility, functioning multiple ways with a community
In light of the above, it seems plausible to think that the commonly acquired and more or less shared normative epistemic sensibilities – the epistemic norms – to be found among a community of interdependent inquirers commonly function in a continuum of ways – depending on what mix of projects and considerations are salient or prominent for the agents at the time. Simplifying, one can say that epistemic normative sensibilities have at least two prominent faces:
As Customs. They serve as widely shared models for ways of satisfying a select set of concerns – as readily deployed ways, understood to be conducive in the pursuit of a project understood as affording a good. It is plausible that the veritistic gains and losses (or related gains and losses taken to be central in the narrow epistemic game) loom large here. Focusing on these central veritistic gains and losses, epistemic norms would seem to function as normative models for state-of-art practice by which agents can effectively pursue the generation of true beliefs (individually, and in groups). To the extent that agents in classes of cases can be understood as responsive solely or predominately to such veritistic concerns, their epistemic normative sensibilities serve as models of customary (state-of-the-art) ways of attaining satisfaction on that score.
As Social Norms. They also serve to regulate an individual and joint practice in which interdependent agents contend with mixed motives. As noted earlier, individual epistemic practice is conditioned not solely by the veritistic motivation associated with the narrow epistemic game, but also by yet wider, competing, concerns – it is influenced by wider ranges of gains and losses. Thus, epistemic agents can face mixed motives, and insofar as such mixed motives are salient to oneself and to those on which one is interdependent, the relevant normative sensibilities can function as social norms.
Here we can pick up a suggestion found in Bicchieri (Reference Bicchieri2017). She notes that one and the same “rule” or normative sensibility with respect to a practice might function as a model for agents of the customary way of regulating their practice in a way conducive to satisfying individual wants/needs in a domain, and as a model by which agents regulate community practices in a way conducive to community goods in the face of mixed motives (and thus as a social norm). She notes that her definitions of various kinds of norms suppose that the relevant rules are responsive to an unambiguous and static choice situation. But she also recognizes that marginal cases are possible:
Here I shall offer a few static definitions. They are static because in real life, the social constructs I talk about may morph into each other and often do. A custom may become a social norm in time, and a social norm may revert to a custom. (Bicchieri Reference Bicchieri2017: 3)
So, Bicchieri envisions one rule, or normative sensibility, having different roles over time – as the predominate choice situation in the relevant community changes. At one time, the rule regulates practice in a domain with few mixed-motives and it then and there functions as a customary way of meeting the concerns characteristic of that domain. Then, at some later period, there have come more significant mixed motives from wider concerns – so that the rule functions as a social norm.
This seems right. But, one can also envision one and the same rule responding to what amount to various contextually salient choice situations that may confront agents across a community during one reasonable stretch of time – a period. One can envision one period in which the contextually situated choices of agents are somewhat diverse – with some agents at times contextually facing few mixed motives, while other contextually situated agents face more mixed motives. Epistemic life seems to have this character. Thus, epistemic agents within the community find themselves sometimes in the one kind of case in which their epistemic rules serve them as customary solutions by which they can further their own and community epistemic projects, and sometimes in the other kind of case in which they are contextually subject to mixed motives to be resisted in pursuing their own and community epistemic projects. Epistemically “external concerns” can grow and recede. These observations suggest that within communities there can be, and commonly are, an array of choice situations in which, rather than social norms and customs morphing into one another across time, one and the same rule might function as a custom for some contextually situated agents and as a social norm for others – and agents may find themselves at one time situated in the one way and another time situated in the other. Overall, a practice in a group might then conform to a rule that functions both as a custom and as a social norm for variously contextually situated agents within the same period. In such a case, the distributed practices in a group at one and the same time would be partially explained as customary ways of individually satisfying some subset of ends associated with an understood practice, and partially explained as a result of a social norm coordinating agents in the face of mixed-motives.
This may seem pretty abstract and it may be helpful to consider an example of what is intended. Norms of hygiene provide an illustration, and a strong analogy with epistemic norms. One gets the benefits of one's own good personal hygiene, and one gets some benefit from the personal hygiene of others. The agent's own hygiene thus yields a personal good (the increased likelihood of health for the individual) and a public good (as those in a group that is coordinated in such good personal hygiene enjoy a lower risk of disease). Hygienic sensibilities, which may be widely shared at a time, can represent a customary state-of-the-art way of pursuing the individual good. Compatible with this, there is an important element of coordination in the production of a public good (disease avoidance) to be discerned here – this is evident when one thinks about the extent to which folk in a community are hygienically interdependent. The effectiveness of a given agent's hygiene practices is conditioned by the practices of others in the group. As a result, those hygienic sensibilities, which may be widely shared at a time, can then serve as models for the regulation of practices in one's community – and agents may enforce those sensibilities.
Imagine that at some point within a group, some scientist announces an understanding of some set of highly troubling diseases – one that implies a mode of their transmission, and a set of practices by which agents could minimize the risks of such transmission – say periodic handwashing, not touching one's eyes with unwashed hands, using some manner of latrines, and the like. Suppose also that the advice is here transmitted as a recommendation for individuals, “here is how one can minimize one's risk.” The diseases being fearsome and salient to many in the group, the great bulk of folk are readily so motivated. Thus, the procedures function as customary matters of state-of-the-art hygiene. Now, as the threat of the disease comes to be less salient to at least some in the group, adherence to the practices begins to decline. Also, perhaps, some of the practices come to be more burdensome – wells dry up, and some folk must walk farther to find facilities. Individuals in such situations may be tempted to not undertake the practices on occasions. Now, the motivations associated with the customary state-of-the-art practices are, in such increasingly common contexts at least, less decisive. Again, the practices become less common. Now, this is noticed, and it may be noticed that when others fail to undertake the hygienic practices they thereby add to the risk faced by others, even those who are using good hygiene. Plausibly forgoing the indicated hygienic practices will draw an evaluative response from those who do conform – and the representation of these practices comes to then function as a social normb – it will become a rule that is enforced in familiar ways within the group (by shaming, gossip, ostracism, and the like). Now, there will likely be a diversity of motivations to be found operative in the population at any given time. At least in many situations, many will be motivated personally to deploy the state-of-the-art hygienic practices for themselves. For them, the practices are customs. Some may not find the risks of defection salient, or may face greater than costs in conforming to those best practice (damn well), and they may be tempted to cut corners. Yet, it may be common knowledge that folk expect others to conform to the hygienic practices and think less of those known to defect from such practices. To the extent that folk conform for such reasons, the rules of hygiene function as social normsb. Now, there may be subgroups and perhaps contexts in the wider group where the different kinds of motivation dominate. Perhaps those who have seen the ravages of the diseases in question, readily undertake the practices for their individual good and would do so even (or especially) were others not to. Perhaps those with few obstacles to such practices also are motivated individually to conform. So, a fair bit of the conformity in the group will be explicable as the use of customary best practices. At the same time, others will prefer to conform to such practices in some significant measure because they expect that others expect them to, and will sanction them if they do not. Notably, the proportion of the population who are motivated in the one way or the other may shift over time. So that the understanding of state-of-the-art practice may function predominantly one way at one time, and predominately another way at another. Further, as indicated above, the same individual may be differently motivated to conform across time. (Likely many of those whose own use of the state-of-the-art practices is motivated in ways characteristic of customs are also among those whose practice of sanctioning non-compliance is characteristic of social norms.)
Applied to epistemic practice, this suggests a picture in which agents’ shared epistemic normative sensibilities function both as a model for the state-of-the-art customary way of satisfying their characteristically individual epistemic ends and as a social normb regulating individual and joint practice. Thus, the distributed practices in the epistemic community at one and the same period would be partially explained as customary ways of satisfying some subset of ends characteristic of epistemic practice, and partially explicable as social normsb by which agents manage their own and others’ practice with a view of community epistemic cooperation.
5. From the playroom to the seminar room humans socially acquire and enforce a normative sense for “how it is done” epistemically
The reflections above (and in Henderson and Graham Reference Henderson and Graham2017a, Reference Henderson and Graham2017b) make a strong abductive case that epistemic norms can be understood as functioning as social norms in some significant degree. To function as a social norm is to be a social norm. This is one face of our epistemic norms – they are community norms that should be responsive to, and serve to regulate, a sprawling set of joint epistemic projects. Still, I have also acknowledged another face of epistemic norms – and it is one that will be familiar from the perspective of traditional, largely individualistic, epistemology. The traditional approach would have us frame matters in terms of the individual epistemic agent normatively regulating his or her own belief-forming practice in the pursuit of attaining true beliefs and avoiding false belief. Of course, noticing epistemic interdependencies within communities, such an individual agent would also come to be concerned with the practice of others on whom the agent might be dependent.
Which, if either, of these two facets of epistemic norms is the more fundamental?
5.1. A reassertion of individualist epistemology? Thinking one facet is the most fundamental
One might suppose that epistemic normative sensibilities are most fundamentally a matter of individualistic epistemology – while admitting that they are secondarily made to do double duty as social normsb. With dialog and dispersion, certain belief fixing practices become customary state-of-the-art practices, and, in reaction to interdependencies and to variations in actual practices, normative sensibilities representing those state-of-the-art practices are derivatively deployed as social normsb.
Understood generically, such an individualistic picture of normative sensibilities is not restricted to epistemic sensibilities. Think of an individual practice concerning which one has attained some sense that certain ways of pursuing the practice yield better success than others. Suppose that one came by this sensibility as the result of individually experienced results.Footnote 13 On this basis, an agent forms an individual normative sensibility that serves in the first instance to discipline that agent's own ongoing practice. Such would be individual normative sensibilities.
Now supposedly the practice yields something of value – and, we may now add that this is a good in which others can share. So that one can come to have an interest in other's successes – as these may be at least indirect gains for oneself. Others have a corresponding interest in one's own successes and practices. Accordingly, agents would deploy their individual normative sensibilities when evaluating and responding to the practice to others in their community. The idea is that what is rooted in individual practice, namely normative models for that practice based in experienced rates of success, now come to be applied to others with whom one is to some degree interdependent.
This framework has room for forms of social transmission. Perhaps another individual's variant practice was discernibly effective, so that folk noticed the relative degree of success in attaining the relevant value. That individual's innovations in practice then spread by what is commonly termed success-biased transmission (for a discussion of success, prestige, and conformist transmission and the evolutionary bases, see Henrich and Henrich Reference Henrich and Henrich2007).
To emphasize, on this rather traditional way of thinking about epistemic norms, they can be derivatively, while not fundamentally, social norms.
5.2. Two equi-fundamental facets
The forgoing rather traditional line of thought has a certain familiar plausibility to it. However, I want to explore an alternative picture – one in which from the earliest moments in the biography of agents their epistemic normative sensibilities are as much social as individual – one in which epistemic norms are social norms, and deeply social, from the start. The alternative understanding turns on the idea that human agents – which includes human epistemic agents of course – acquire much of their normatively laden understanding from others. The epistemic practices of others, and their epistemic interdependencies upon others, are salient for human epistemic agents, including emerging epistemic agents. From very early and continuing throughout their lives (from the playroom to the seminar room) epistemic agents are confronted with an ongoing epistemic practice. They observe these practices – getting a sense for how they are done. They seek to take their place in this community by having a hand in the relevant practice modeled about them – the practices of those on whom they rely, the practice that would allow others to rely on them. What is taken up or internalized is internalized as normatively laced models/sensibilities – which I will represent using the construction: the way it is (to be) done.Footnote 14 Thus, the practice is taken up normatively on the basis of a kind of assimilation of a practice presumed to have a kind of prima facie fittingness. Certainly, for the infant and small child, normative sensibilities are commonly the result of watching as care-givers and other (apparently successful or powerful) agents do things – so that what is learned is a normatively laced understanding of the way it is (to be) done. While more mature agents may have resources for greater critical reflection on practices and norms, learning normative models in more advanced contexts continues to reveal a similar phenomenon: in graduate training and lab practice one continues to learn how it is (to be) done.
The fundamentally social picture: The epistemic agent acquires much of its epistemic normative sensibilities in ways deeply conditioned by interaction with others – where this often involves, not just observing others, but commonly involves joining with others in more or less joint practice modeled by those others.Footnote 15 In joining in a practice, taking a place in an epistemic community, the agent seeks to measure up to “how it is (to be) done,” and understands that others look for no less from them. Commonly, the taking of an evaluative/normative stance may be modeled for them – but it need not be, and agents are capable of adding in the evaluative stance largely unbidden. In this process, the extant sensibilities of others regarding fitting practice are read into the modeled practice and internalized in a way that comports with the doings, expectations, and evaluations of others.Footnote 16
Let us emphasize the myriad ways in which there are social dimensions to what is learned from others. This can be seen in the way that from very early on even observation and attention allocation come to be socially informed by way of a kind of cognitive modeling or mimicry that is easy to overlook. It is clear, for example, that very young children are highly attentive to the eye movements of conspecifics. So, just watching the way in which others scan their environment may afford significant cues to children concerning the scanning practices and attention of salient others. There is apparently a significant tendency to mimic the direction of gaze of others, and thus a tendency to mimic the salient others’ practices of allotting attention. As the child allots attention in roughly parallel ways, the child also learns what is gotten from the allotment of attention. The child internalizes these practices and sensibilities.Footnote 17
Further, as the child learns how it is (to be) done – for practices at various levels – the child comes to be qualified to take more and more of a place within various de facto groups (or communities). Notably, the child does not merely track and emulate the attention and perspective of others – rather the child participates in joint attention with those others. The child seeks to become a functioning member of these groups – to help as well as to be helped. As Michael Tomasello notes, compared to our nearest primate relatives, humans are distinctive in the degree to which we are epistemically helpful – informatively helpful. Indeed, a kind of unqualified informative helpfulness seems particularly characteristic of young children – who become somewhat more guarded as they mature. To be informatively helpful, children must (again) track the perspectives of others so as to provide information that those others seem to lack. In so doing, children are exhibiting one form of sensitivity to and sensibility about distributed responsibility for helping others in epistemically connected ways.Footnote 18 In seeking to become a member of the epistemic community, the child is learning “how it is (to be) done,” and seeking to conform to this as a member – and recognizing that others are noting the child's own progress here.
These remarks on some very early and rudimentary processes are intended as a down payment on the idea that even very rudimentary forms of epistemic practice and sensibility – some associated with observation – can be deeply social from very early on.
This learning of practices and internalization of the sensibilities of others generalizes beyond the rudimentary stages. As things proceed, the epistemic child will observe others investigating various kinds of matters – both individually and in groups. They note what information is deemed relevant and thus shared in cooperative epistemic practices, and note the treatment of various pieces of information as more or less relevant by others in these cooperative practices. Think of the responses to others encountered in the classroom, or over the course of common everyday conversations. From the patterns of engagement presented here children can acquire sensibilities regarding the relevance and significance of various pieces or kinds of information, and they do so by both conformist transmission and success-biased or prestige-biased transmission. Conformist transmission is involved to the extent that children seek to acquire and provide to others the kind of information commonly adduced among their models. Prestige-biased and success-biased transmission is involved to the extent that children seek to provide information of sorts found among the apparently most successful or most deferred-to models among their acquaintance. It seems plausible that this kind of learning is found at the ground floor in much learning of sensibilities regarding deduction, induction, and abduction – and it continues to develop throughout one's ongoing epistemic development.
Much of epistemic practice may be learned from others, being internalized or taken up as the way it is (to be) done. Again, what is internalized here are normatively laced models/sensibilities. There is reason to think that the human cognitive critter is set up (evolutionarily, thus psychologically) to learn how it is (to be) done by observing others. This includes generalizing observed practice into normative sensibilities. It is on the basis of this that one can be confident that in the period in which one might suppose that children are generating their earliest epistemic normative sensibilities, they are enmeshed in a set of dependencies in which relatively knowledgeable (or at least relatively powerful/successful) others are highly salient models taken up normatively. Thus, the practice is taken up normatively on the basis of a kind of assimilation of a community practice presumed to have a kind of prima facie fittingness.
This presumption and normative assimilation seen among young children seems both evolutionarily distinctive of humans and in some sense reasonable.
Reasonable? Certainly the child must be, in some inarticulate sense, struck by the apparent epistemic (and other) power of salient others in the community and of their joint mutually helpful practice on which the infant or young child is strikingly dependent. Notably, caregivers and cooperative practice in the family will strike the child as notably powerful relative to what the child itself can at first manage. Later, their teachers and their teachers’ sources will similarly be salient. It seems reasonable for the child to do as they do – to progressively join them. This community affords much on which the child is always already reliant. As things develop, it is a community in which the child naturally wants a place – wants to have a hand and to be relied on to lend a hand. This is not to say that the child just seeks to conform – far from it: the child seeks to be a part of projects – including epistemic projects – which obviously have payoffs.
Evolutionarily distinctive of humans? Again, the human cognitive critter is set up (evolutionarily, thus psychologically) to learn how it is (to be) done by observing others – to generalize observed practice into normative sensibilities. Thus, in the period in which one might suppose that children are generating their earliest epistemic normative sensibilities, they are enmeshed in a set of dependencies in which relatively knowledgeable (or at least relatively powerful/successful) others are highly salient. In a remarkable series of studies, Michael Tomasello and various collaborators have looked at how readily human children generate normative lessons regarding how it is (to be) done from observing the demonstrated actions and interactions of others. Several of these studies have to do primarily with young children readily acquiring and then enforcing constitutive rules – such as rules regulating a single player game or activity with no significant payoff (Rakoczy et al. Reference Rakoczy, Wameken and Tomasello2008; Schmidt et al. Reference Schmidt, Rakoczy and Tomasello2011; Schmidt and Tomasello Reference Schmidt and Tomasello2012). But, not surprisingly, children latch upon the way it is (to be) done when encountering activity that can yield a reward – notably one requiring coordination across multiple agents. Here, children proved quite capable of spontaneously developing and enforcing norms when none were modeled for them – and they then modeled these norms, transmitting them to new initiates (Goeckeritz et al. Reference Goeckeritz, Schmidt and Tomasello2014).
Importantly, there is evidence that, when children latch onto norms governing projects with rewards, they seem to be motivated in part by the desire to take their place in the relevant community or group – they seek to join and contribute to joint projects. They seek to be the sort of agent on which these others can rely in such joint projects – and they clearly seem to find that the joint activities can be rewarding of themselves independent of their more obvious instrumental payoffs. Thus, Tomasello notes that, in contrast with chimpanzees,
human children collaborated in the social games as well as the instrumental tasks. Indeed, they sometimes turned the instrumental tasks into social games by placing the obtained reward back into the apparatus to start the activity again; the collaborative activity itself was more rewarding than the instrumental goal. (Tomasello Reference Tomasello2009: 65; see also Warneken and Tomasello Reference Warneken and Tomasello2006)
It is fitting to see applications in epistemic contexts. As things develop, the child wants to become increasingly accepted in the epistemic community as a full member – to be thereby respected. Further, the child wants to be respected and accepted because the child has become capable of doing what is fitting for a full member – of having a hand in forming belief of the way it is (to be) done. This requires that the child conform to state-of-the-art norms – norms conditioned by the full set of extant information afforded within the community.
On this picture much of the normative epistemic sensitivity learned from early on is of the character to be learned and applied as a member of an epistemic community. It is learned socially – in interaction with others as models and teachers – and learned as a “how it is (to be) done” that is fitting to epistemic practices in a community of interdependent agents. As children learn how it is (to be) done, others as relatively knowledgeable models, and evaluating judges, and as subjects of each other's evaluations, are highly salient to them. They readily seek to take their place in their community and to be approved in light of the standards for “how it is (to be) done.” In all this one sees connections with the motivations suggested in Bicchieri's account of social norms – that the salient relevant others do things a certain way, with associated shared benefits, that others expect that others will do so, and judge how much and how well others do so, and that others will be marginal in the salient community until one becomes adept and consistent in conforming to how it is (to be) done.
I am not advancing this approach as some deep a priori truth about epistemological (or general) normativity that supposedly holds for all possible cognitive critters. I am not committed to the idea that there could not be creatures all of whose epistemic normative sensibilities are of the sort posited in the traditional individualistic conception. For all I say here there might be aliens that regulate themselves and others in the fundamentally individualistic fashion of the traditional view – and only secondarily deploy their normative sensibilities as social norms. Without prejudice to such matters, the view explored here is that it is a fact about humans that much of their epistemic normative sensibilities are deeply social from the start. It is in interaction with other human epistemic agents that human epistemic agents come to have, to refine, and to transmit, much of their epistemic normative sensibilities as shared understandings of how one forms beliefs. This is done as deeply epistemically interdependent creatures always already enmeshed in, and learning from, interdependent practice. For the child of one or two years of age – being informatively helped and quickly coming to informatively help, being corrected and quickly coming to correct others – there is “the way to think” (both what to think and how to inquire or learn more) that is learned from others, then readily practiced and modeled for others. Of course, there is no suggestion that the resulting epistemic sensibilities – which come to be significantly shared across initiates and adepts – are simply conventional. Rather they are the result of a cultural ratchet. Still, from early to late, the relevant sensibilities are keyed to a joint practice and a matter of normative transmission, correction, and coordination.
Here one might also acknowledge that, in extraordinary circumstances, there might be something on the order of a human epistemic wild-child that somehow survives with minimal succor, and without any appreciable epistemic models. Perhaps such a child could develop epistemic normative sensibilities – something like individualist epistemic norms. My claim is merely that, for humans as we find them, many of their actual epistemic norms are fundamentally social in their etiological dynamics and motivational components. Their norms reflect the mix of facets discussed above (they are both customary and social norms). However, commonly these norms are not merely derivatively social norms – but are such fundamentally.
It can be acknowledged here that observing others requires some rudimentary epistemic practice on the agent's part, but it is not at all clear just how much of a normative sensibility this rudimentary practice requires in its simplest cases. Doubtless the infant confronts frustrations in observational expectations as well as successes – and somehow learns from this. This shapes their perceptual system. Plausibly what is learned so as to inform the infant's perceptual system involves, not just content concerning the world, but also sensibility concerning, for example, needed and fitting perceptual caution. (Much of this may come to be possessed inarticulately and in the form of what Henderson and Horgan (Reference Henderson and Horgan2011) term morphological content.) The latter would amount to a form of normative sensibility that I allow is involved in observing others – and thus is involved in learning practices by way of observing others’ practices. But notice that my claim is not that all epistemic norms are fundamentally social norms, but that much of our epistemic normative sensibility is.
