Why you cannot run an experiment on society
A physicist can heat the same gas a hundred times and get the same result. A historian cannot rerun the French Revolution to test a theory about it. This is not a technical inconvenience that better instruments will eventually fix. It is a difference in the nature of the subject itself.
The reason is simple once stated plainly: human events do not repeat. Every revolution, every crisis, every war changes the people who lived through it — so a second, similar event does not happen to the same people in the same state; it happens to people who remember the first one. Add to this that no two historical episodes ever match exactly on every condition, and that a participant who knows they are part of an "experiment" behaves differently than they would in ordinary life — and the method that works in chemistry and physics simply does not transfer.
None of this means nothing reliable can be known about people and societies. It means a different method of knowing is needed — not the one that works in the natural sciences.
Two ways of thinking about society
Social thought has long contained two large traditions. The first is empirical: gather more cases, compare them, extract a pattern. This is how statistical sociology works, and much of journalism's appeal to "historical examples" as well. The second is deductive: start from a few statements about what a human being and human action actually are, and derive the consequences step by step, the way geometry proceeds from axioms.
The second path can look less "scientific" at first, since it cannot be tested by experiment. Yet this is the path economist Ludwig von Mises took in the twentieth century when he built praxeology — a theory of human action derived from a single axiom: man acts, meaning he chooses means to achieve ends under conditions of scarcity. From that one sentence, without a single statistical observation, a substantial part of economic theory can be logically derived — why exchange exists, why money is needed, how prices form.
The method this article describes follows the same path but extends it beyond economics: not only purposeful action, but all human activity mediated by a sign — language, number, law, ritual — can be understood deductively, step by step, from a few basic claims about how distinction, sign, and levels of abstraction work.
The trap of "historical examples"
There is a second, less obvious reason to distrust induction in talk about society. The mass image of history is shaped not by archival documents but by novels, theatre, and film. Most people's image of Spartacus comes not from reading sources on the revolt of 73 BCE but from Giovagnoli's novel and Soviet cinema. The image of Napoleon comes from Tolstoy and Stendhal. Julius Caesar comes from Shakespeare. The Middle Ages, in ordinary consciousness, is mostly Walter Scott, Dumas, and Hollywood costume dramas.
This means that when someone in a debate about how society works reaches for a "historical example," they are usually reaching not for a document but for a literary image retold as historical fact. This does not devalue literature — it can carry a deep structural truth about human beings — but it undermines its use as an empirical base for testing a theory, because the selection of "telling" examples has already been made by an artist, not a researcher.
Deduction and three tests
If empirical testing is this limited when it comes to society, what tests a deductive theory instead? Not direct correspondence to facts, but three other criteria.
The first is internal coherence: does the theory contradict itself, do its conclusions hold together from start to finish. The second is explanatory power: does the theory clarify what other approaches leave unexplained or write off as coincidence. The third is fruitfulness: does the theory open new questions and deepen the analysis, or does it hit a dead end after its first application.
This is not a softer standard than empirical testing — it is a different standard, suited to the subject. A geometer does not verify the Pythagorean theorem by measuring a million triangles; they check whether it follows logically from the axioms. A deductive theory of society works the same way — except its axioms concern not points and lines, but how a human being distinguishes, marks with a sign, and holds levels of understanding together.
What about the texts the theory draws on?
The method has a feature that can surprise at first: it draws on classical texts, including biblical ones, not as historical evidence but as structural records. The difference matters. Historical evidence is used to prove "this happened." A structural record is used differently: as a system in which the relations between levels of the human are fixed with unusual density and preserved unchanged across many generations.
In this sense such a text serves structural theory the way Euclid's axioms serve geometry: not an object of belief, but a compact, stable form from which structural relations can be read repeatedly and reliably. A reader who does not treat the text as a religious authority can use it exactly as an atheist can study Euclidean geometry without sharing its Platonic assumptions.
Why this matters in practice
None of this is abstract methodological posturing; it comes with a concrete working rule. A researcher following this path must state their starting axioms openly, derive consequences from them rigorously, and accept that another researcher, starting from different axioms, may reach different conclusions. Disputes between such theories are not settled by trying to catch each other out against a single fact — that path is deceptive in social theory — but by comparing which system explains more, more coherently, and more fruitfully.
None of this cancels or diminishes empirical work — gathering data, studying sources, fieldwork. It is only a reminder that data says nothing by itself until it has been selected, described, and interpreted within some theory — which means a theory always comes first, even when the person using it has never stated it out loud.