Why are we so polarized, and how can we move forward? A perspective from social epistemology.

Contemporary academic philosophy has recently taken a “social turn” regarding the way it thinks about knowledge and related issues.  Put differently, philosophy has turned away from the traditional ideal of a self-sufficient inquirer, doing his or her best to figure out what is reasonable to believe, in favor of a conception of knowledge that emphasizes the importance of “social-epistemic dependence” on others.  The dependence is “social” in that it is a dependence on other persons, as well as other aspects of a knower’s social environment, including facts about how that environment is socially organized.  The dependence is “epistemic” in the sense that it is a dependence on others for one’s knowledge, as well as for other intellectual goods, such as reasonable belief or opinion.

Very much related to all this, a number of contemporary philosophers have sought to emphasize the role of trust in shaping our intellectual lives.  That is, they have been concerned with how a moral—or at least practical—relation to other persons influences such things as what knowledge we have and which of our opinions count as reasonable or rational.  At first glance, this can seem to stray too far from traditional ways of thinking about knowledge and belief.  Intellectual standings such as knowledge and reasonable judgment seemingly have more to do with evidence and good reasoning than with our moral and practical relationships to other persons.  And even if personal relationships and other social factors can provide a psychological explanation for why a person holds the beliefs that they do, such factors seem irrelevant to explaining how or why a person’s beliefs are reasonable, rational, or otherwise intellectually respectable.

Nevertheless, it seems to me that the reasons behind the new “social epistemology” are ultimately compelling.  In particular, this new way of thinking helps to explain quite a bit about our current social and political landscape, including its echo chambers, “fake news,” “alternative facts,” and increasingly insulated world views.  One salient feature of that landscape is the way in which opposing groups increasingly characterize each other as either intellectually or morally flawed.  That is, the default explanation for opposing opinions is that those holding them must be either stupid or evil, or perhaps both.  This is a disturbing situation, not least because it invites us to characterize other people as “lost causes.”  How can you reason with someone who is too stupid to see the plain facts, or too immoral to care?  What is disturbing, then, is not only the sad state that we are in, but also the poor prospects for advancing out of it.

I said above that the new social epistemology helps to diagnose what has gone on here.  But just as importantly, it seems to me, it gives us resources for hope rather than pessimism regarding how to move forward.  Specifically, it gives us resources for seeing other people—even those with whom we most vehemently disagree— as reasonable rather than morally and/or intellectually flawed.  And if we can see people as reasonable rather than just stupid or immoral, we can see them as reasonable to talk with and otherwise engage. This oversimplifies, of course, because reasonableness is not an all-or-nothing thing.  The more careful point is that social epistemology gives us resources for seeing people and their opinions as more reasonable, less flawed, than we otherwise might.  And this means resources for addressing a social and political situation that might otherwise seem intractable.

But how is all of this supposed to go?  To see it, we need to get some ideas on the table.  Specifically, I draw attention to an important category and then three plausible points.

The important category is that of an epistemic community, defined as a group of people who are cooperating with respect to some set of information-dependent tasks.  That is, an epistemic community is a group of persons who are cooperating to get something done, and whose cooperation depends on their sharing relevant information.  For example, a business corporation constitutes an epistemic community in this sense, insofar as it involves people cooperating so as to conduct appropriate business, and insofar as that involves sharing relevant information (about product quality, supply chains, sales, etc.) in the process.  So is a family an epistemic community, insofar as family members cooperate in the course of “family business,” and must share relevant information in doing so.  In fact, epistemic communities are all over the place, they overlap and intersect, and they go in and out of existence all the time.  And everyone of us, at any given time, is a member of several such communities.

The first point to make is that epistemic communities, as such, have two important concerns: a) the acquisition of quality information for the purpose of carrying out their relevant work; and b) the distribution of that information to those who need it.  The second concern is closely related to the idea that epistemic communities take advantage of a division of epistemic labor.  That is, the person who acquires relevant information (about recent sales, for example) is not always the person, or not always the only person, who needs it.  The possibility of sharing relevant information, so that not everyone has to do everything for themselves, is one of the strongest motivations for having epistemic communities in the first place.

This brings us to a second important point.  Namely, that epistemic communities are also characterized by norms or standards for acquiring and distributing relevant information.  This is a different way of saying that epistemic communities have an interest in quality information, or information that meets relevant standards for reliability and accuracy.  Moreover, the norms or standards governing the acquisition of information will be different from those governing the distribution of information.  That is because our dominant concern when acquiring information is quality control—we want only accurate and reliable information to get into the system in the first place.  But our dominant concern when distributing information is to get it to the people who need it.  That does not mean that quality control is irrelevant to information distribution, but that it plays a less important role than in the acquisition stage.  Consider the analogy of a military base.  There are high standards for getting in through the gates— identifications must be checked, perhaps vehicles will be searched, or other tests will have to be passed.  But once in, a person is not then allowed to go wherever they please—there are rules for who is allowed to go where and when even within the base.  Nevertheless, the rules for getting through the gates in the first place are not the same as those for travelling around once in, nor should they be.

And now for the third and final point that is needed.  Namely, that informational transactions with other persons occur both in the service of information acquisition (in service of getting quality information into the community) and in the service of information distribution (in service of sharing quality information with others in the community).  And that means that different informational transactions will be evaluated differently, depending on the context in which they occur.  Put differently, information coming from other persons will be evaluated according to different norms and standards, and appropriately so, depending on whether that person is a member of one’s own epistemic community.

To illustrate, consider a conversation that takes place between a job applicant and a personnel director.  Clearly, it would be inappropriate for the personnel director to simply believe whatever the job applicant tells her—for example, that he has some relevant expertise, or that the people at his former job loved him.  On the contrary, it is the job of the personnel director to evaluate that information as believable or not, and she would be failing at her job if she did not do so with proper diligence.  Contrast that exchange with another:  After due diligence, the personnel director reports to her CEO on the applicant’s qualifications.  Relative to this context, it is appropriate for the CEO to trust what the personnel director tells her.  More carefully, it is appropriate for the CEO to treat her conversation with the personnel director differently than the personnel director treated her conversation with the job applicant. And that is because the second conversation takes place in a context of cooperation that the first conversation does not.

The same point can be made in terms of an appropriate division of labor.  A corporation would be dysfunctional if its CEO had to apply the same standards of care in evaluating her personnel director’s reports as the personal director should apply in evaluating a job applicant’s reports.  On the contrary, a well-functioning corporation is defined by the kinds of cooperation that allow for efficient divisions of labor, and in this case an efficient division of epistemic labor.  That point holds for epistemic communities more generally. By their very nature, well-functioning epistemic communities are defined by relationships of cooperation and trust that allow for an efficient division of epistemic labor.  And that normative structure entails that information coming from different persons will be evaluated differently, and appropriately so, depending on those persons’ relationships to the relevant epistemic community.

In sum, the very nature of epistemic communities entails that different people should be afforded different levels of trust, that information coming from different people should be evaluated according to different standards.  And that explains a lot about how things can go bad.  For suppose that a community initially gets something wrong, for whatever reason.  That mistake, once made, will be more easily distributed within the community.  Moreover, that mistake will be more difficult to correct, insofar as opposing information from outside the community will be evaluated by a higher standard.  This is because any information, mistaken or not, will be evaluated by higher standards if coming from outside a community than if coming from within it.  What is worse, this is a necessary consequence of a well-functioning epistemic community.  In other words, people within an epistemic community are right to assess information according to these varying standards. This explains why, once an opinion gains a foothold within a community, it can be hard to root out, mistaken or not.  No stupidity or immorality is necessary to drive the explanation.  That is not to say, of course, that epistemic communities and the people within them are never stupid or immoral!  It is only to say that those factors are not needed, in principle anyway, to explain entrenched disagreement between different communities.  To frame the point differently, such factors need not exhaust our explanations regarding why people disagree with us.  The picture can be more complicated.

Accordingly, resources in social epistemology can help us to explain entrenched disagreements and even polarization among different communities.  But what about resources for moving away from polarization and toward agreement?  Here things can seem grim.  In fact, our explanation seems to predict increasing polarization, especially between communities that are already characterized by mutual distrust.  I want to suggest that there is a way forward, however, and one suggested by our social epistemology model.  It involves the hard work of building better, or at least different, epistemic communities.  And that involves the hard work of building relationships of trust.  One reason for hope in this regard is that the boundaries of epistemic communities are neither rigid nor singular.  As noted above, epistemic communities intersect and overlap, and they come in and out of existence all the time.  Moreover, relationships of cooperation and trust come in any number and variety, including family relations and friendships.  Finally, each of us is already a member of multiple epistemic communities.  All of this makes for opportunities to reshape our existing communities and to build new ones, and to thereby interrupt and redirect existing flows of information.  That, admittedly, is not easy to do.  But it is the hopeful alternative to treating each other as lost causes.

This post is based on some ideas from the author’s recent book The Transmission of Knowledge available from Cambridge University Press.