In 1865, August Kekulé claimed the ring structure of benzene came to him in a dream about a snake biting its own tail. The story is famous—and philosophically loaded. Should we care how a scientist arrives at a theory, or only whether the theory survives testing?

This question cuts to the heart of what counts as good science. For decades, philosophers insisted on a strict wall between discovery and justification. But that wall may be less solid than it looks. Understanding why matters for how we think about scientific reasoning itself.

Here's a principle that feels obviously right: it doesn't matter where a theory comes from. What matters is whether it holds up under testing. Kekulé's dream, Newton's apple, Fleming's contaminated petri dish—these are great stories, but they tell us nothing about whether benzene actually has a ring structure, gravity follows an inverse-square law, or penicillin kills bacteria.

This intuition has a name. Philosophers call it the discovery-justification distinction, and it was formalized most influentially by Hans Reichenbach in the 1930s. The context of discovery—the psychology, the accidents, the cultural moment—belongs to biography. The context of justification—the evidence, the predictions, the logical structure—belongs to philosophy and science proper. Mixing them up, the argument goes, is a category error.

The logic is compelling. A theory proposed by a genius and a theory proposed by a coin flip face the same tribunal of evidence. Truth doesn't care about origins. This principle protects science from a kind of genetic fallacy—judging ideas by where they come from rather than what they say. And for a long time, most philosophers of science accepted it without much fuss.

Takeaway

A theory's truth doesn't depend on how someone thought of it. Judging ideas by their origins rather than their evidence is the genetic fallacy—and science rightly resists it.

But here's where things get interesting. Suppose a scientist develops a theory not from a dream but from a deep structural analogy with a well-confirmed theory in another domain. Does that context tell us nothing about whether the new theory is worth pursuing? It seems like it should count for something—not as proof, but as a reason to take the theory seriously.

This is the idea of heuristic value. The path to a theory can carry information about its promise. A theory born from systematic patterns in data, from productive analogies, or from the unification of previously separate phenomena arrives with a kind of epistemic recommendation letter. It hasn't been proven, but its origins suggest it's the kind of thing that might be true. Scientists make these judgments all the time when deciding which hypotheses to test first.

Consider how Darwin arrived at natural selection partly through analogy with artificial selection by breeders. That analogy wasn't evidence that natural selection was true. But it wasn't irrelevant either—it showed the idea had roots in something real and observable. Discovery contexts can function as informal evidence about a theory's plausibility, even if they don't constitute formal justification. The strict wall between discovery and justification starts to look more like a useful simplification than an absolute truth.

Takeaway

How a theory was generated can be informative about whether it deserves serious attention. Origins aren't proof, but they aren't nothing either—they can serve as rough guides to plausibility.

The deepest challenge to the discovery-justification distinction comes from historians and sociologists of science who argue it doesn't describe what scientists actually do. In practice, discovery and justification aren't separate phases. They're intertwined. Scientists often develop theories and the standards for evaluating them simultaneously. The criteria for what counts as good evidence can shift as new theories emerge.

Thomas Kuhn's work on scientific revolutions made this point forcefully. When Copernicus proposed heliocentrism, the available evidence didn't clearly favor it over the Ptolemaic system. The justification only became compelling as new instruments, new observations, and new standards of explanation co-evolved with the theory. Separating the clean logical structure of justification from the messy process of discovery gives us a rational reconstruction—a tidy version of what happened, but not what actually happened.

This matters because rational reconstructions can hide important features of scientific reasoning. If we only look at the final logical structure, we miss how scientists weigh competing considerations, how social trust in research communities shapes what gets tested, and how the process of inquiry itself generates new evidential standards. The discovery-justification distinction remains useful as an analytical tool. But treating it as a hard boundary risks making science look more mechanical and less intelligent than it really is.

Takeaway

Separating discovery from justification is a helpful analytical tool, but treating it as a rigid boundary can make us miss how real scientific reasoning works—where the process of finding ideas and the process of evaluating them constantly shape each other.

The discovery-justification distinction captures something real: a theory's truth doesn't depend on its biography. But treating this distinction as an absolute wall oversimplifies how science actually operates. Origins carry heuristic weight, and discovery and justification intertwine in practice.

The philosophical lesson is that scientific reasoning is richer than any single distinction can capture. Recognizing this doesn't weaken science—it reveals a more accurate picture of how careful inquiry actually navigates from hunches to knowledge.