Umbrellaology is not a science. It is a fictional discipline invented in 1941 by philosopher John Somerville to illustrate exactly what makes something a science versus a mere collection of facts. The thought experiment, published in the journal *Philosophy of Science*, describes a researcher who meticulously gathers data about umbrellas: their colors, sizes, locations, materials, and owners. The data is detailed and carefully organized, but it lacks something fundamental. Understanding why it falls short reveals a lot about what science actually requires.
What Umbrellaology Describes
In Somerville’s thought experiment, a fictional “umbrellaologist” devotes serious effort to cataloging everything about umbrellas. The data is precise. The methods are systematic. On the surface, it looks like research. But the entire exercise lacks a testable question. There is no hypothesis about why umbrellas behave in certain ways, no prediction that could be proven wrong, and no explanatory framework tying the observations together. It is pure data collection for its own sake.
This is what makes it such an effective teaching tool. When students or readers encounter umbrellaology, they instinctively feel something is off, but they often struggle to articulate exactly what’s missing. That struggle is the whole point. It forces you to think carefully about the line between organized knowledge and actual science.
Why Collecting Data Isn’t Enough
The core lesson of umbrellaology is that gathering facts, no matter how carefully, does not make a field scientific. Science requires explanation, not just description. A biologist doesn’t just count birds in a forest. She asks why certain species thrive there, forms a hypothesis, and designs observations or experiments that could prove her wrong. The data serves a purpose: testing an idea about how the world works.
Philosopher Karl Popper, the most influential modern thinker on this question, argued that the defining feature of science is falsifiability. A scientific claim must be structured so that evidence could, in principle, show it to be false. Umbrellaology has no claims to falsify. It simply records what exists. You could catalog every umbrella on Earth and still not have done science, because no explanatory claim was ever at stake.
Popper’s criterion isn’t the only one philosophers use. A broader analysis identifies at least five qualities that separate science from non-science: internal logical consistency, empirical confirmation of predictions, specific proposed mechanisms for how things work, non-arbitrary application of ideas, and honest use of all available data rather than cherry-picked results. Umbrellaology fails on nearly every count. It makes no predictions, proposes no mechanisms, and has no theoretical claims to be consistent or inconsistent with.
The Demarcation Problem
Somerville’s thought experiment taps into one of philosophy’s oldest debates: the demarcation problem. Where exactly is the boundary between science and non-science? This question has occupied thinkers since antiquity. The Roman philosopher Cicero warned against hasty assent to ideas that are false or insufficiently substantiated, arguing instead for careful comparison of arguments. That impulse, skeptical and evidence-driven, runs through every attempt to define science.
Modern philosophers generally agree that no single criterion perfectly separates science from pseudoscience in every case. Falsifiability captures something important, but it doesn’t resolve every edge case. Some philosophers emphasize that science is also a social process: a theory counts as scientific partly because it belongs to a research tradition maintained and scrutinized by a community of practitioners. Umbrellaology has no such community, no tradition of peer challenge and refinement, and no evolving body of theory.
Thomas Kuhn, another towering figure in philosophy of science, focused on the role of paradigms. In his view, a mature science is organized around exemplary achievements that define what good research looks like and what problems are worth solving. He pointed to landmarks like Newton’s *Principia Mathematica*, Lavoisier’s chemistry texts, and Maxwell’s equations for electromagnetism. These works didn’t just present data. They offered powerful frameworks that generated new questions and guided future research. A field without any such framework, like umbrellaology, is not even in the running.
Why It Still Gets Taught
More than 80 years after Somerville published it, umbrellaology remains a staple in introductory science and philosophy courses. UC Berkeley’s Understanding Science project recommends it as a starting activity for students aged 13 to 16, specifically because it works well for drawing out students’ assumptions about what science is. The exercise asks students to read a description of umbrellaology and then argue whether it qualifies as science. The discussion that follows tends to surface misconceptions, like the belief that any systematic observation counts as scientific research.
The thought experiment is also relevant to real methodological debates. In modern research, high-throughput data collection (genome-wide association studies, large-scale nutritional surveys, commercial data mining) can drift toward umbrellaology’s trap. When researchers test thousands of variables without a guiding hypothesis, they can generate statistically significant but meaningless results. Data mining software is sometimes designed to find patterns in massive datasets regardless of whether those patterns reflect anything real. Fields that rely on commonly agreed analytical methods tend to produce more reliable findings than those still experimenting with their tools, precisely because shared methods and theory keep the work anchored to testable ideas.
What Separates Real Science From Imitation
Umbrellaology helps clarify a set of features that genuine sciences share. They propose explanations for why things happen, not just descriptions of what exists. They generate predictions that can be checked against reality. They build on theoretical frameworks that connect individual findings into a coherent picture. And they operate within communities that challenge, replicate, and refine each other’s work.
A phone book is full of carefully organized data. So is a stamp catalog. Neither is science. Umbrellaology sits in the same category: meticulous, systematic, and entirely pointless as a scientific enterprise. The value of Somerville’s invention is that it makes this distinction vivid and hard to forget. If you can explain why umbrellaology isn’t science, you understand something important about what science actually is.