Clarifying science’s conceptual framework
Clarifying science’s conceptual framework
A new framework for understanding reality—and why science isn’t always the right tool
By Adam Davies
For over 400 years, science has been humanity’s most trusted tool for discovering truth. With its emphasis on transparency, replication, and controlled observation, it has built our modern world—from vaccines to space telescopes. But what if this venerable method breaks down under certain conditions—not because of human error, but because it’s being applied in the wrong context?
That’s the central insight behind Polytactics, a conceptual framework that shows that science systematically fails with “tricky problems”–those where strategic adversaries can undermine investigations or intentionally manipulate evidence. These aren’t fringe cases. They’re everywhere—from espionage and fraud to criminal trials and even personal relationships. And understanding why science falters in these settings doesn’t diminish its power. It clarifies its boundaries.
Science doesn’t just fail randomly with tricky problems. It fails in consistent, predictable ways. And those patterns of failure reveal something deeper: reality isn’t a single puzzle. It’s a landscape made up of structurally different kinds of problems—each demanding a different way of being solved. So when science clashes with intelligence analysis or criminal investigation, it’s not a failure of logic. It’s a category error.
In saying this, Polytactics breaks with the dominant view—the idea that science is the ultimate filter for truth, and that it can, more or less, be applied to all real world problems. That view treats reality as virtually monolithic: made up mostly of problems that science is built to handle. But Polytactics, paints a different picture – where reality is split into three distinct types of problems. And each problem type, by its very nature, calls for a different approach. Science is still exceptionally powerful in its own domain—but it must defer to other rational methods elsewhere.
The helps clarify how different established and rational disciplines—science, history, and intelligence—are connected. They aren’t rival ideologies, or arbitrary solutions, but necessary adaptations that emerge naturally from different structural constraints of the same underlying system. It also reveals why attempts to apply the scientific method universally can backfire, leading to blind spots, denial, or misplaced skepticism.
Before Edward Snowden’s 2013 revelations, many scientists and skeptics dismissed the idea of global mass surveillance. Why? Because by scientific standards, the evidence was weak: anecdotal, circumstantial, and tied to what sounded like an implausible conspiracy. Following the usual heuristics—Occam’s razor, distrust of secrecy, demand for replicable data—science concluded the claims were nonsense.
But others, applying an intelligence-style mindset, came to the opposite view. In that world, conspiracies are expected. Anecdotal clues are often all that remain after strategic adversaries have hidden or destroyed direct evidence. What looks like noise to science can look like signal to intelligence.
When Snowden released the documents, he didn’t just blow the whistle. He retroactively validated a different way of interpreting the problem. The scientific approach had confidently predicted that such a program couldn’t exist. The intelligence mindset had predicted that it could—and likely did.
This wasn’t just a disagreement. It was a clash between frameworks – a deeper epistemological divide. Polytactics explains why: this was a classic tricky problem, defined by secrecy, deception, and adversarial interference. Science was misapplied—and so key assumptions, usually so powerful in their correct domain, became liabilities. Intelligence-style reasoning, by contrast, was built for exactly this kind of terrain.
The patten that emerges when science fails with tricky problem shows that reality is not monolithic. It falls into three structurally distinct categories:
Honest Problems — No risk of being strategically misled, because evidence is so abundant that no one is realistically capable of doing it, or because the strategic impact is too low for anyone capable to bother
E.g. Measuring the speed of light, planetary motions, jellyfish migrations
Passive Tricky Problems — Real risk of being strategically misled, specifically because of adversary action in the past that is no longer ongoing.
E.g. Was Julius Caesar murdered? Did my great-great-great grandmother have a child out of wedlock?
Evidence is on a manageable enough scale, and strategic impact was (at some point in the past) enough to motive capable adversaries to manipulate evidence enough to mislead future investigations.
Active Tricky Problems — Real risk of being strategically misled by active adversaries who can adapt to your investigations in real time.
E.g. Do Iran have nuclear weapons? Who robbed the bank? Is my spouse having an affair?
Evidence is on a manageable enough scale, and strategic impact today is enough to motivate capable adversaries to actively mislead investigations in an ongoing, adaptive manner.
Crucially, Polytactics shows that just by separating reality into these problem types, the fundamental constraints of each problems themselves causes major established investigatory methods – science, historical methods, and criminal / intelligence methods, and along with each of their specific characteristics – to arise naturally as a dictated and necessary response to the intrinsic characteristics of these basic problem types. The tenets of the Scientific method arise naturally as a response to honest problems. Historical methods arise naturally as a response to passive tricky problems. And Intelligence / criminal investigative methods arise in the same way as a response to active tricky problems. This formally connects science, history, and intelligence, showing they are not arbitrary silos, or competing methods, but three sides of the same coin – deeply connected necessary solutions to a shared problem space.
Polytactics also resolves a striking disconnect between theory and practice in how we apply methods of inquiry. The dominant conceptual paradigm holds that science is authoritative and therefore should be applied to all problems. But this is clearly not how we operate in practice. In real-world settings, science is primarily used for a narrow class of problems—those involving natural phenomena without strategic adversaries. When we face adversarial problems, society reliably turns elsewhere. Courts rely on criminal investigation and jury deliberation—not statistical significance or experimental replication—to determine guilt. Intelligence agencies detect espionage through secrecy, deception, tradecraft and inference, not transparency or controlled experiment. These methods rest on entirely different assumptions about evidence, truth, and justification. And yet, they are not fringe alternatives—they are institutionally and culturally embedded as the appropriate responses to certain problem types.
Why, if science is authoritative? Polytactics answers this question. It not only aligns our conceptual framework with our actual practices, but also explains why different methods—with fundamentally different epistemic assumptions—are systematically preferred for different kinds of problems. It shows that these preferences are not cultural quirks or historical accidents, but necessary adaptations to the structural constraints of the problems themselves.
Just by making the one shift at the heart of Polytactics—replacing a monolithic view of problems with a triadic one—unlocks a cascade of deep insights. Without needing new tools or extra assumptions, it instantly resolves some of the biggest puzzles in science and philosophy, including:
Resolves the demarcation problem–the centuries old question of defining what is, and is not science–by showing the boundaries of science (and its characteristics) are dictated by the problems it is adapted for, instead of a checklist of method steps.
Explains the longstanding divide between science, history, and intelligence methods, showing all three are simply expressions of the structure of the underlying problem space
Explains the source of science’s extraordinary power compared to other frameworks–as a result of honest problems being uniquely conducive to collaborative, cumulative progress, whereas tricky problems are not.
It explains why preemptively dismissing conspiracies is a common and successful scientific heuristic, despite major conspiracies being documented– because in the domain of science conspiracies are intrinsically impossible–not .
It accounts for the odd pattern of why reproducibility is critical in chemistry, but not in paleontology or astronomy: It is the scale of evidence that matters, not reproducibility, in determining if science can be applied to a field, because this renders it honest or not.
Most importantly, it gives us a principled way to decide what method to use, when, based not on institutional habit but structural fit.
In an age of disinformation, geopolitical tension, and institutional mistrust, knowing how to investigate reality isn’t just academic. It’s potentially existential. If we continue applying scientific tools to adversarial problems, we risk confusion, paralysis, and misplaced certainty, and render ourselves highly vulnerable to strategic manipulation.
Polytactics doesn’t reject science. It restores it—by showing where it truly shines, and where it must make room for its epistemic siblings.
Because when the problem changes, so must the method.