Liberal Science: Fallibilism and Error Correction
Res Extensa #9 :: How we decide what's true through distributed consensus, a look at Kindly Inquisitors and the work of Karl Popper
I'm back after a bit of a writing hiatus around here. In just a couple short months we bought a house, moved, switched kids schools, and had some health stuff going on (all good), all while doing DIY house projects in the background. We've reached something like an equilibrium here over the past week, but it's unlikely to last long. On to the issue!
Our culture has deep-seated disagreements about what goals we should pursue — politically, socially, economically, scientifically. Conflicting visions of society set different people on different paths, where ideas and priorities clash about what should be done.
Disagreement on objectives is natural and inevitable. There are too many ideas and each one of them has rational justifications why it should be pursued. There'll never be total consensus on what's right and true. We’ll never have 100% agreement on what we want to achieve.
But how we go about seeking progress is even more important to maintaining the free, prosperous, and productive society we enjoy. It's critical that we understand the boundaries of how we agree on and implement proposed ideas.
In his book The Open Society and Its Enemies, philosopher Karl Popper wrote about this question at great length: how should we go about the process of defining what's true, and what type of society is most optimized for progress? In his definition, an open society is "one in which individuals are confronted with personal decisions as opposed to a magical or tribal or collectivist society." In order to get there, it's important that we understand the mechanics that foster permissionless advancement and not a regressive return to tribalist roots.
The Open Society
The term "liberal science" in this issue's title has a distinct definition. It seeks to describe how, in a liberal society, we explore the boundaries of our knowledge in an open way. How do we decide what's true?

Earlier this year I read Jonathan Rauch's 1993 book Kindly Inquisitors, a phenomenal extended essay best described as a defense of free inquiry and open scientific discovery. I read it in a few days, and in the process of writing up my notes, read it again. It's excellent! Go get a copy and check it out.
In George Will's introduction, he refers to it as "sharp as a stiletto", which is an excellent way to describe a short book (less than 200 pages) with a higher density of quality than you'll find just about anywhere. In a world where no publisher seems to permit <400 pages on any topic, it's delightful when you find a work of surgical mastery like Rauch's attack on authoritarianism and speech suppression.
The book seeks to define liberal science and what makes it a unique (and uniquely modern in many ways) process for creating knowledge. Rauch isn't prescriptive about what specific ideas, policies, or scientific theories are right or true, his focus is solely on the process through which societies arrive at consensus.
Defining Liberal Science
We already have terms for describing the this sort of open, decentralized network of decision making in other areas.
In the political sphere we call it "liberal democracy", and say we have "rule of law". A citizenry arrives at a body of laws by which all must abide, including rulers themselves. No one is exempt. We arrive at decisions through plurality or majority consensus. And yes, sometimes this is slow, often painful, but it allows for incremental progress without corrupting the decision making apparatus.
In economics we have free market capitalism, where we allow a vast meganetwork of producers and consumers to come to agreement through prices. Few authorities should have their fingers on the scales (though modern incarnations of capitalism are far from completely free, for better or worse).
Rauch's "liberal science" describes such a system applied to the creation of knowledge, our search for truth. It has has roots in Popper and work of other intellectuals like John Stuart Mill, David Hume, and John Locke. His aim is to clearly define how an open, liberal system of ideas should work. Expansion of our breadth of knowledge requires an organic process, designed for robust advancement without taking shortcuts that would destroy the open society itself.
Of course, we live in a society rife with disagreement. Rauch's argument is that a liberal society cannot avoid this discord, and that, in fact, the liberal scientific order feeds off of this agglomeration of ideas bubbling around together. As he defines it, liberal science intends to do is create an ecosystem in which we can arrive at collective sensemaking through skepticism, debate, persuasion, and testing.
There are two central operating principles:
No one gets the final say — Also known as the "skeptical rule". No line of inquiry is ever considered closed. Any existing idea or theory, no matter how widely accepted as truth is immune to being re-exposed to new criticism. Think about something like Newton's laws of motion: they gradually achieved status as defining principles of physics, until challenged three centuries later by Einstein's work on relativity. If the debate was ended after Sir Isaac finished his work, we wouldn't have permitted our search for the underlying structure of the universe to continue! Einstein would've never escaped his patent office gig and we'd still be wrong about how gravity works.
No one has personal authority — Known as the "empirical rule". No specific individual has the right to decide, regardless of credential. No one has special permission to be "more right". Ideas must be exposed to a decentralized community of "checkers". Knowledge is the rolling critical consensus of this network. The experience of no one in particular is what's most important — if a result can't be replicated by the network at-large, it isn't truth.
There's a great quote on this from philosopher Charles Sanders Peirce, who Rauch calls one of the “patron saints” of the book:
“One man’s experience is nothing if it stands alone.”
Liberalism as a philosophy permits a system to evolve along its own path. Applying Darwin's theory to ideas, we can think of liberal science as a "survival of the fittest" competition between the most tested, validated consensus belief.
The fuel of natural selection is variety — trying millions of variations and testing them through exposure to the environment. In the same way, we want millions of diverse ideas to swirl around together, with the most checked ones gaining traction and validity through testing. A mixture of ideas fed into a liberal scientific ecosystem leads to better outcomes long term. Here's Rauch:
As we check and criticize and find common ground, as we propose ideas and they fall apart and we try again, our knowledge advances. The picture of the world so produced is not a neat set of consistent statements, all lined up in rows under the proper headings; it is a seething ecology of hypotheses in constant flux and competition and contradiction.
"A seething ecology of hypotheses" is a beautiful description.
Error Correction
The effect of any natural selection-like process is "error correction", the gradual identification and removal of error in a system. In biology, a maladaptive trait is an error that will be selected against in future generations.
Last year I found this great discussion with physicist David Deutsch where he discusses error correction in the context of Brexit:
The entire video is worth a watch, but particularly important is how critical an error-correctable system is to the maintenance of the open society. He compares the British "first-past-the-post" electoral system to that of the EU community: in Britain the government has to perform on a knife's edge and satisfy the electorate, whereas the supranational EU government is layers removed from individual citizens. This structure isolates decisions from rapid testing.
If an electorate's decision drives policy that is found to be bad in practice, we want systems that embrace removal and change, preferably before policies become entrenched. Says Deutsch:
The more a system is conducive to removing rulers and policies without violence, the more it is democratic. And that is the most important criterion in politics, more important than choosing right rulers in the first place.
Any system that purports to know in advance what the consequences of a ruler or a policy are going to be is based on a fundamental error. Systems which try to do that clamp down on progress because they entrench the ruler or the policy.
The philosophical term for error correction is "fallibilism", a term coined by Peirce and widely written about by Popper.
In evolution, species develop environmental adaptations based on trial and error deployed on a massive scale. Genetic adaptations begin as random mutations. Through exposure to the environment, well-adapted mutations stick, while poorly-fitting ones disappear.
In the world of ideas, progress comes from the ability to freely experiment, to try thousands of concepts on for size, test them, and ditch or ignore those that don't work. People should be motivated and enabled to try out ideas in lower-risk settings when we can check them for validity. When a species develops a mutation that's a poor fit for an environment, it's invalidated and gone before it has time to take deep root. If it gets protected and entrenched in a population artificially (animals raised in captivity, or genetically-modified plants, for example), it's exposed to catastrophic tail risk once back in its natural environment.
Ideas work the same way: insulated from being tested, bad ideas will cause disaster if allowed to become entrenched.
Knowledge is in constant flux; we're never finished determining the truth on any specific dimension. "No final say" tells us that all ideas, theories, or systems are open to challenge. Well-tested and best-fit systems are simply resilient to these challenges and take phase change-level shifts to successfully unseat.
All of this closely pattern-matches with other ideas in the Res Extensa milieu — decentralization, gradual change, evolution.
Other Principles
Rauch contrasts the Liberal principle from several other common ones in how we sort truth from fiction. Each one has merit — these principles often start out with noble intent, but each is prone to corruption and misuse.
Here's how he defines 4 other principles:
Fundamentalist principle — Those who know the truth will decide who's right. So... who should decide the truth?
Simple egalitarian principle — All sincere persons' beliefs have equal claims to respect. How do we then disambiguate the true from false? If everything is in bounds, how do we make progress?
Radical egalitarian principle — Like the simple, but the beliefs of persons in historically oppressed classes or groups get special consideration. Similar to the first two, who gets to decide those group definitions? Somewhere along the way it has to be one and not another who gets to pick.
Humanitarian principle — Any of the above, but with the condition that the first priority be to cause no hurt. Along some version of the definition, every new theory or idea ever offended or hurt someone when first proposed. What now?
The only one compatible with an open society is the liberal principle indexed on public criticism.
Against Historicism
Popper's The Open Society, and another extended essay of his (aptly titled), tear into the philosophy of "historicism", a philosophy he considers so bankrupt that it gave rise to all flavors of modern day totalitarianism. What is historicism and why is it so poisonous to free and open societies?
If you read the Wikipedia page, it's not immediately apparent to a non-philosopher like me what the heck it means. I had to read quite a bit on it to get it straight in my head.
Here's how Popper defines it (emphasis mine):
Karl Popper used the term historicism in his influential books The Poverty of Historicism and The Open Society and Its Enemies, to mean: "an approach to the social sciences which assumes that historical prediction is their primary aim, and which assumes that this aim is attainable by discovering the 'rhythms' or the 'patterns', the 'laws' or the 'trends' that underlie the evolution of history".
…Popper attacks "historicism" and its proponents, among whom (as well as Hegel) he identifies and singles out Plato and Karl Marx—calling them all "enemies of the open society". The objection he makes is that historicist positions, by claiming that there is an inevitable and deterministic pattern to history, abrogate the democratic responsibility of each one of us to make our own free contributions to the evolution of society, and hence lead to totalitarianism.
Rauch adds one more detail, which helps you see where historicism leads inevitably to totalitarian decline:
It assigns to an elite, intellectual few the job of purging society of false consciousness
Historicism has desirable goals like the pursuit of freedom (obviously attractive), but goes sideways in how it gets there: through creation of a perfected "state".
In Plato's historicist view, an elite few with special knowledge should lead and guide the state. Because historicism claims the "march of history" is unavoidable, we must grant authority to those who know. Truth is whatever the philosopher king says it is. An awfully convenient perspective for the philosopher himself!
In his theory of forms, Plato argued that we should strive to find "essences", the unchanging and absolute ideas, leading to his aversion to uncontrolled progress through an evolving society. In his utopian theory, the philosophers would determine truth and the state would resist change, hold onto the status quo, and bring us back to an essentialist, pure, tribal society. You can see where this worldview was corrupted and used to build modern totalitarian movements, even if Plato himself would've been benevolent.
The true open society distributes the authority to (in Rauch's words) "no one in particular". Decentralizing authority is one of liberal science's core principles.
From the Archives
I pulled a few choice morsels from my Readwise archives relevant to this topic.
First, couple of great ones from that founding fountain of wisdom, The Federalist. The Framers understood the importance of protecting against tyranny, and co-locating power in many independent units (states, branches of government) to avoid dominance by majorities. Here's James Madison in Federalist #10:
It is in vain to say that enlightened statesmen will be able to adjust these clashing interests, and render them all subservient to the public good. Enlightened statesmen will not always be at the helm.
Design systems to be robust to enough to withstand bad rulers! The selection of a bad ruler is an "error" that we want to correct for.
And from Madison and Hamilton, Federalist #52:
Let us consult experience, the guide that ought always to be followed whenever it can be found.
Experience = trial and error.
Nassim Taleb always drops knowledge on this theme. From Antifragile:
Urban planning, incidentally, demonstrates the central property of the so-called top-down effect: top-down is usually irreversible, so mistakes tend to stick, whereas bottom-up is gradual and incremental, with creation and destruction along the way, though presumably with a positive slope.
And one last one from Hayek, Mr. Decentralization himself, in The Fatal Conceit:
The market is the only known method of providing information enabling individuals to judge comparative advantages of different uses of resources of which they have immediate knowledge and through whose use, whether they so intend or not, they serve the needs of distant unknown individuals. This dispersed knowledge is essentially dispersed, and cannot possibly be gathered together and conveyed to an authority charged with the task of deliberately creating order.
As always, thank you for reading. If you’ve enjoyed this, forward it onto your friends and colleagues or subscribe yourself if you haven’t. Let me know what you think of the newsletter so far! Also, check out my blog for more of my writing, or follow me on Twitter.