Most uses of language can be understood in both referential and functional terms. If I tell the policeman “He ran the red light”, in referential terms I am claiming that, in some world external to my language, there was a car driven by a person I refer to as “he” which crossed an intersection while a red lightbulb was lit. But my words have functions as well, quite apart from what they refer to. A person might be fined or go to jail as a consequence of what I say. I might be conveniently exonerated of responsibility for an accident. Those consequences might be independent of the referential accuracy of the remark. Or they might not be. Perhaps there will be other corroborations, and inconvenient penalties if I am deemed to have lied. Regardless, it is simultaneously true that words refer to things and utterances have consequences. Both as speakers and as listeners (or as writers and as readers) we need to consider the “meaning” of a use of language on both levels if we are to communicate effectively.

Often there are tensions between referential accuracy and functional utility. Referential accuracy does not necessarily imply virtue. Whether we agree with the practice or not, we all understand what is meant by a “white lie”. Statements with identical referential meaning can yield profoundly different social consequences depending on how they are said. To “speak diplomatically” does not mean to lie, but rather to pay especial attention to the likely effects of an utterance while trying to retain referential accuracy. To “spin” has a similar meaning but a different connotation, it suggests subordinating referential clarity to functional aspects of speech in a crassly self-interested way. But paying attention to the functional role of language is not in itself self-interested or crass. We all pay attention to how we speak as well as what we say. If we did not, we would needlessly harm people. Even if we are scrupulously truthful, we all make choices about what to say and what to omit, when to speak and when to remain silent. When we discuss our inner lives, often the consequences of our utterances are more clear (even to ourselves) than their referential accuracy, and perhaps we let the desirability of the consequences define what we take to be the truth. Perhaps that is not, or not always, without virtue.

This bifurcation of language into referential and functional strikes me as illuminating of the stereotyped left-right axis in politics. In broad, almost cartoonish, terms, one might describe a “left” view that humans as individuals have limited power over their own lives, so the work of politics is to organize collectively to create circumstances and institutions that yield desirable social outcomes. The “right” view is that, absent interference by collectivities that are inevitably blind to fine-grained circumstances (and that usually are corrupt), individuals have a great deal of power over their own lives, so that differences in outcome mostly amount to “just desserts”. It’s obvious why there might be some conflict between people who hold these different views.

On the key, core, question of whether individuals have a great deal of power or very limited power to control outcomes in their own lives, the stereotyped left view is, in referential terms, more accurate. If you are born in poverty in a war-torn country and fail to achieve a comfortable American-style upper-middle-class life style, it’s hard to say that’s on you, even if some very tiny sliver of your countrymen do manage to survive to adulthood, emigrate, and prosper. In narrower contexts, the question becomes less clear. For those lucky enough to be born in a developed country, are differences in outcome mostly a result of individual agency? For Americans born white, raised in middle-class comfort, and provided an education? For people born with identical genes? The case that differences in outcome result from choices under the control of individuals, for which they might be held responsible, grows stronger as we restrict the sample to people facing more similar circumstances. But even among the most narrow of cohorts, shit happens. People get sick, debilitated even, through no fault of their own. As a general proposition, individual human action is overwhelmed by circumstance and entropy. Policies designed with grit and bootstraps for their engine and individual choice for their steering wheel usually fail to achieve good social outcomes. This is the sense in which it’s true that “the facts have a well-known liberal bias“.

But, before the left-ish side of the world takes a self-satisfied gloat, it should face an uncomfortable hitch. In functional terms, widespread acceptance of the false-ish right-ish claim — that people have a great deal of power over their own lives, and so should be held responsible as individuals for differences in outcome — may be important to the success of the forms of collective organization that people with more accurate, left-ish views strive to implement. This isn’t a hard case to make. A good society, qua left-ish intuitions, might provide a lot of insurance to citizens against vicissitudes of circumstance. A generous welfare state might cushion the experience of joblessness, housing and medical care might be provided as a right, a basic cash income might be provided to all. But a prosperous society with a generous welfare state requires a lot of people to be doing hard work, including lots of work people might prefer not to do. If people are inclined to see their own and others’ affairs as products of circumstance, they might easily forgive themselves accepting the benefits of a welfare state while working little to support it, and even lobbying for more. They might find it difficult to criticize or stigmatize others who do the same. That would lead to welfare-state collapse, the standard right-wing prediction. But if an ethos of agency and personal responsibility prevails, if differences in outcome are attributed to individual choices even in ways that are not descriptively accurate, if as a social matter people discriminate between justifiable and unjustifiable uses of public benefits and stigmatize the latter, the very prevalence of a right-wing view of human affairs might falsify the right-wing prediction and help to sustain the left-wing welfare state. Conversely, the existence of a left-wing social democratic welfare state renders the right-wing view less wrong, because it diminishes disparity of circumstance, increasing the degree to which differences in outcome actually can be attributed to individuals’ choices. Irreconcilable views reinforce one another.

God is an ironist. If left-ish views are referentially accurate while right-ish views are functionally useful, then a wise polity will require an awkward superposition of left-ish perspectives to inform policy design and right-ish perspectives as public ethos. Singapore is ostentatiously capitalist, is widely perceived as a kind of protolibertarian paradise, yet it builds a rich welfare state out of mandatory, government-controlled “savings” and extensive intervention in health care and housing markets. The Scandinavian countries are left-wing social democracies, built on a politics of trade union solidarity, yet the right-wing Heritage Foundation ranks them about as “economically free” as the United States despite governments that spend much larger shares of GDP. Nordic politicians bristle at being called “socialist”, and they maintain higher levels of labor-force participation than the welfare-stingy US.

Like Yin and Yang, black and white, right and left might stand perpetually in opposition even as they require one another to form a coherently incoherent whole.

Lyman Stone is a wonderful writer I have only recently discovered. (Thanks Will Wilkinson!) You should read everything he writes. This, unfortunately, is not his best work. But it presents a useful occasion to say some things about “institutions”.

“Institution” is one of those words much more frequently used than understood. It has multiple, quite different meanings, both in common usage and as a term-of-art in social affairs. Let’s go through a few of them. (I’m sure there are more. The word itself is an institution.)

In its most common usage, “institution” just means formal organization. We might say “Bluegrass Community and Technical College is an educational institution”. The pompous president of a corporation might regale assembled shareholders with an account of how things are proceeding “in this institution”.

A second meaning in common usage is simply something unusually prominent or important to some community. “That crazy hot-dog vendor is an institution in this town.” Or, um, “The word itself is an institution.”

In discussions of social affairs, the word “institution” does not mean either of those things, exactly. Because language, even technical language, is beautiful, and because language, in order to form itself, feeds upon associations and connections in the minds of those who affect to wield it, you will find echoes and overlaps between the word’s common-use meanings and its meanings in social theory. But they are not the same.

A narrow definition of “institution” in social theory is something like “a pattern of social behavior in which participants occupy stereotyped roles which are independent of the identities of the individuals who perform them”. Motherhood is an institution, while your habit of babysitting your friend’s nephew is not. The former is a general social pattern independent of the people who, multiply and at different times, perform the roles of mother and child. The latter is just something you happen to do. If you become famous and “friendnephewsitting” becomes a widely adopted practice then, yes, then it would become an institution. But, for now, no. On the bright side, babysitting itself is most certainly an institution. You’ll always have that.

Finally, there is a broader use of “institution” in social affairs. Economists, God help us, frequently refer to things like “the quality of a nation’s institutions” in spinning just-so stories about why some countries thrive and other falter. They don’t mean to say that one country does patterns-of-social-behavior-with-stereotyped-roles better than any other. All human agglomerations form institutions in that sense just fine. Motherhood, for example, is close to universal. Bribery is most certainly an institution, prominent to various degrees in many polities. Most economists (but not always or all, this is interestingly contested) would deem prominence of bribery as an institution in society to be reflective of “low quality institutions”, however fervently and frequently the practice is enacted. All societies have “strong institutions”, in the sense of having entrenched patterns of social behavior. However, different societies enact different institutions. When an economist makes comparisons between the quality of a country’s institutions, she is not referring to their resilience or ubiquity. She is making judgments about the adaptiveness of the bundle of patterns-of-social-behavior-with-stereotyped-roles prominent in one country relative to the mix of institutions found elsewhere.

This agglomerative usage of institutions is the most common one in policy discussions. When we talk about some entity’s “institutions”, we are using that as a shorthand for the patterns of social behavior we observe within it. I have quietly dropped the qualification -with-stereotyped-roles, which we so emphasized in the first, technical definition. That is on purpose. In practice, outside of specific technical contexts, the word institution has come to mean simply any pattern of social behavior that might form part of a portfolio of such patterns which together constitute the je ne sais quoi of a group. Institutions become the stuff that gives different collections of humans distinct character even under identical circumstances. You might think of them as atoms of “culture”, another vague and grandly abused term. But then you might object — you should object! — that purely individual propensities, when aggregated, also contribute to “culture”. True, true! Fortunately, the agglomerative usage of “institutions” is so teleological and imprecise that it amoebically comes to encompass even purely individual propensities, to the degree that they are claimed to affect the quality of the aggregated whole. Individual thrift, for example, becomes an “institution”, because people claim that such thriftiness is material to the character and adaptedness of the whole. This usage creep can be misleading, sometimes dangerously so. For example, it’s not particularly plausible to suggest that elaborate patterns-of-social-behavior-with-stereotyped-roles, like representative democracy, are heritable genetic traits. But it may be plausible that individual propensities are heritable. Once we let individual propensities sneak into our definition of “institution”, then it becomes plausible that at least some institutions are genetic. But if “institutions” can be genetic, so can representative democracy, right? Things go downhill very quickly from there.

This usage of “institutions” becomes infuriatingly vague, but that may be why it is so prevalent. We need labels for all the things that we can’t quite pin down, because we know there are things we can’t pin down that are important in their effects. Social scientists frequently try to identify or operationalize “institutions”, to come up with comparative datasets of institutional character or quality. But these empirical exercises necessarily impose very contestable models of which particular institutions are relevant to the outcomes they measure — “An independent judiciary!”, “Fast business formation!”. Measurement is a huge problem. Societies may have the formal trappings of institutions (reminiscent of our first common usage) without the behavioral substance researchers intend to measure. The formal constitution of an autocracy might be idealistic and democratic, for example, but that obscures rather than reveals the actual institutions of the state. Characterizing institutions is hard. Casual attempts often reveal more about the prejudices of the attempter than they do about the society ostensibly described.

The way I like to think of “institutions” is this: Institutions are to groups what habits are to individuals. Attributing social differences to institutions is similar to attributing individual differences to varations of habit. It’s not wrong or entirely meaningless, but it’s also unsatisfying, incomplete, a bit tautological. How and why did the different habits, individual or collective, emerge? What sustains or might undo them? Changing social institutions is a process similar and similarly fraught to changing habits. Just as it is not enough to simply resolve to stop smoking, if you mean to break the habit of smoking, the mere passage of a law or enactment of a policy may not succeed at changing social conditions embedded in our collective habits. But that does not mean we are helpless, that (in Stone’s mistaken assertion) “the point of institutions is that they are not usually malleable to directed change”. If we wish to stop smoking, we do resolve to stop smoking, but we also work to change our circumstances and incentives in ways that support the desired change. We might avoid smoky bars (or we might have, back when there were still smoky bars). We might wear a nicotine patch, or take drugs that block the effect of inhaled nicotine. That is, to change habits, we do not merely resolve, but we act strategically upon ourselves as though we were an object as well as a subject. We change institutions all the time, and in precisely the same way. There is nothing magical or mysterious or necessarily even hard about creating, modifying, or undoing social institutions. In my other life, I do a lot of work related to blockchains, which interest me precisely because they are a technology for reifying the kinds of circumstances and incentives that contribute to institution formation, and they are amenable to intentional construction and directed change.

It is a trick of conservatives — who, when they live up to their name, wish by definition to prevent some kind of change — to claim that the status quo is immutable, change is impossible, unnatural. Attempts would be destructive. Conservatives try to ground the way things are in mysteries of culture, in ancient hatreds or deep currents of history, in arcana of genetics or race, or, now that they are vague and fashionable, in “institutions”[*]. But institutional change happens every day, all the time, at a micro level within families and businesses, at a political level sometimes quite abruptly, often though not always in directions that were explicitly conceived and intended. As individuals we sometimes find we have fallen into unintended habits. At a social level we sometimes find ourselves in institutions that are harmful or dysfunctional. In either case, we try to change. If we are serious and strategic, sometimes we can.

Meaningful social change occurs at three different levels, at a policy level, an ideological level, and an institutional level. The art of politics is to engender resonance between all three, between the technocratic formalities of policy, and the worldview we take for granted, and the interactions we find ourselves performing. No one says that it is easy. When you think about all the moving parts involved, it is a miracle that any one of us even breathes. But we do breathe. And societies do change, frequently, often as a result of intentional work on the part of people within them.