Science Built Something Into Its Foundation We Can Learn From
There is a word that sits at the centre of science’s self-understanding and is almost universally misunderstood outside it.
That word is “theory.”
Often, you hear someone say, off-handedly, that evolution is only a “theory.” That is, something approaching speculation, and no better than an educated guess.
In science, though, theory means the opposite. Evolution, the big bang, germ theory. These are not educated guesses awaiting confirmation. They are some of the most rigorously tested, most evidentially supported, and most predictive accounts of their explanatory domains available. They are the closest thing to established fact that science is willing to provide.
Even when a scientific law is established, there is a distinction worth noting. Laws describe the “what” of a causal phenomena, based on repeated experiments, observations, and empirical evidence. A theory is relegated to the “why.” And science makes a very clear claim: the “why” of a phenomenon can never move beyond theory. It can never be established as fact and it will always require revision.
This is the most important institutional decision science ever made.
It is also the linguistic embodiment of the first principle this publication has been building across the four previous articles. To quote the first article: “The world is entirely too complicated to make a fully confident, actionable map, plan, and narrative. But we should try to, anyway.”
In other words, the map is not, and can never be, the territory it describes. The account of a thing cannot be the thing. The best available description of anything is a perpetually incomplete one. We must build our incompleteness, our failure to fully map the territory, into the foundation of all our work. And we must admit that this failure is not a temporary condition that will eventually give way to honest investigation. It is, rather, the permanent condition of honest investigation.
Science was the first institution to make this move, building the incompleteness of its “theories” into the foundation of its work, and the progress that followed was not a coincidence. Instead, this ability to recognize its own limitations freed it to begin its honest investigation into our world, ourselves, and the universe.
What science discovered, what science might be called the progenitor of, is what this publication calls a metamodernist posture. It is the recognition that failure is our destiny, but that progress can still be achieved if we hold that failure as a first principle rather than a problem to be solved. What we are attempting to do is extend that move to all other domains of knowledge-seeking.
The Human Side of Science
That is not to say science is without faults. It is, after all, a human endeavour, susceptible to the same greediness, short-sightedness, and misplaced motivations that all human endeavours can get embroiled in when they stop optimising for truth, and opt for qualities like prestige, publication, and career.
This has produced, over the years, many examples of scientific anomalies being suppressed by paradigms that proved overly sticky, producing a zeitgeistic and oppressive narrative that overrides genuine discovery.
The tension between human fallibility and scientific progress was debated directly between Thomas Kuhn and Karl Popper in 1965.
Kuhn, in The Structure of Scientific Revolutions, described scientific advancement as the progression of paradigms and paradigm shifts. Normal scientists work within the accepted model, reinforcing it. Revolutionary scientists break it. His example was the geocentric model of the universe (the scientific zeitgeist at the time, its current paradigm) giving way to the heliocentric model, which shifted the paradigm. It is through the latter, not the former, that major scientific progress happens, Kuhn argues.
Popper disagreed. He argued that Kuhn was committing a form of unfalsifiable historicism where a theoretical law is created out of descriptive past events, then assuming history only moves in one direction. More dangerously, Popper argued, if scientists adopted Kuhn’s model it would become self-fulfilling, turning everyone into normal scientists content to reinforce paradigms rather than challenge them. Instead, Popper believed scientists must hold themselves accountable to the facts, pursue falsifiability, and constantly question the present-day paradigm.
In effect, Popper was making a prescriptive claim about what science should be, and Kuhn was making a descriptive claim about what science has been. They were operating at different levels of abstraction entirely, and framing them as rivals is a category error.
What we get, as with most genuine investigations of complex territory, is the both/neither.
Scientists are human beings. Many are normal scientists working within the paradigm, reinforcing it rather than challenging it. But this is not in opposition to Popper’s ideal, but rather the necessary groundwork for it. Normal science does the inch-work that accumulates the anomalies. At a certain point, enough counter-evidence meets a threshold where the prestige of the current paradigm is dwarfed by the prestige of becoming the revolutionary scientist who replaces it. The human desire for prestige does double work. It both maintains paradigms and defeats them.
Popper describes the foundation that keeps the anomaly alive: the “theory” of incompleteness built into the foundation of science’s ethos. Kuhn describes what happens because it is humans, infinitely fallible, doing the science. And indeed, we need both: the Popperian ideal to aim at, to keep us constantly honest and aware of our short comings, and the honest Kuhnian account of how the work is actually done. Neither alone is sufficient, but neither, in any case, should be dismissed.
The Replication Crisis
We see the human side of science most visibly in what became known as the replication crisis. It was the widespread failure, beginning in the early 2010s, to reproduce the results of published scientific studies. Replication has been called the cornerstone of empirical science, and its failure put the credibility of established scientific knowledge into genuine question.
Psychological research was at the forefront. In August 2015, psychologist Brian Nosek coordinated the Reproducibility Project, which repeated 100 studies published in three high-ranking journals. Of the 97 original studies with significant effects, only 36% were successfully replicated. Of the remainder, 25% contradicted the original findings outright, and 49% were inconclusive.
Similar findings followed. A 2018 study of 21 social and behavioural science papers from Nature and Science found 62% could reproduce original results. Another examined 28 studies, finding 50% failed to replicate. Medical research fared better, with a metastudy of 49 medical studies from 1990 to 2003 found 92% of therapies were replicated, but the picture across a vast array of fields was troubling enough to demand a reckoning.
The causes were recognisable. Science produced at an unprecedented rate, and the pressure to publish could be summed up in the idea that one must “publish or perish.” Thus, prestige became the goal rather than truth, and the Kuhn and Popper argument was made visible in real time.
But here is the both/neither on the replication crisis, too.
It is evidence that science failed, that certainty crept in where it should not have, that humans optimised for the wrong thing, and that the posture was violated under institutional pressure.
It is also evidence that science worked. The crisis exists because scientists were checking each other’s work. Normal scientists were working against normal scientists to correct the faults of their own institution. A problem was recognised and addressed. The foundational commitment to falsifiability, our incompleteness baked into the foundation, meant the failure could not be permanently concealed. The posture reasserted itself as it naturally does when it is part of the foundation, because the foundation that contains it will eventually give out.
Science is both fallible and self-correcting, containing both Kuhn and Popper, yet it is defined by neither alone. It can rise above its failures precisely because it already built the possibility of failure into its foundation.
Science Outside Science
Beyond this foundation, there’s another factor that makes science the most successful knowledge-producing institution in human history. It is not the laboratory, nor the specific methods of any particular discipline, but something broader.
The hypothesis. The testable element. The thing the experiment is checking. Without it, there is nothing to falsify, nothing to build a body of evidence around, and no mechanism for the territory to push back against the map.
This, along with the incompleteness built into our foundation, is what honest investigation in any domain must borrow from science.
This publication has already attempted exactly this. The third article asked what government is for and proposed a hypothesis: the maximum freedom for the maximum number of people. With a desired outcome stated, a hypothesis becomes possible. With a hypothesis, experiments can be designed. Through those experiments we begin to see whether given policies expand freedom or contract it, and whether the territory responds or contradicts the map. Oppositional camps form, each with their own falsifiable hypothesis on how to achieve the most freedom, each containing some genuine observation about the territory. And through that opposition the investigation deepens.
All of it depends on maintaining, like science, a belief in our own fallibility. A recognition that we have only theories and hypotheses and incomplete maps, because the territory will always exceed the account we give of it.
The theory of government. The theory of happiness. The theory of war. The theory of human progress. The theory of meaning.
A word that says: this is our best account at the moment, but it will be revised. And the desire for revision, or much stronger, a commitment to revision in the face of certain failure, which is itself not failure. It is the real condition of any genuine investigation, rather than a temporary problem that sufficient effort will eventually solve.
Like science, we need to agree on a desired outcome as a compass. We need a hypothesis to test against. We need opposition as the mechanism of further inquiry. We need falsifiability. And we need to hold the incompleteness permanently at the first principle, baked into the foundation of all subsequent inquiry.
Because, to put it bluntly, it’s theory all the way down. That’s all we have.
The metamodernist claim and the claim of this publication is not that everything is science. It is that the structure that has made science the most successful knowledge-producing institution humans have built should be extended, then appropriately scaled, to every domain where we are trying to understand something.
Science has been doing this for four hundred years.
The word theory is the proof.
