PLATO'S BUTCHER, MENDELEEV'S GAPS, AND THE DISCOVERY OF STRUCTURE
How the right ontologies generate knowledge
In the Phaedrus, Plato has Socrates describe the philosopher's method through the image of a butcher carving a carcass. The task is to cut at the natural joints, "not breaking any part as a bad butcher might." The distinction Socrates draws is between cutting where the divisions already exist in the animal's structure and hacking arbitrarily through bone, imposing separations that correspond to nothing in the thing itself.
This image has stayed with me because it captures something essential about how understanding works, not only in philosophy but in any domain where one is attempting to make complexity graspable. The right categories reveal structure that was present all along but remained invisible until someone found the language for it. The wrong categories obscure that structure, and no amount of reasoning conducted within them can compensate for the distortion they introduce at the foundation.
Mendeleev's Gaps
Consider what happened when Dmitri Mendeleev arranged the known chemical elements by atomic weight and noticed that their properties recurred periodically. Before the periodic table, chemistry consisted of a long catalogue of individually described substances. Chemists possessed substantial knowledge about each element in isolation but lacked any representation that made the relationships between them visible. Mendeleev's contribution was not the discovery of new data but the identification of a pattern that had been obscured by how the existing data was organized. He found a way to arrange what was already known such that chemistry's joints became apparent.
What made the periodic table extraordinary was not its elegance but its generativity. Gaps in Mendeleev's arrangement predicted elements that had not yet been isolated. When gallium and germanium were later discovered and found to match his predictions with remarkable precision, the table demonstrated something a mere taxonomy could never accomplish. The representation was performing cognitive work, making the unknown knowable in advance by revealing where the structure demanded something to exist.
The same pattern recurs wherever a discipline undergoes genuine transformation. Carl Linnaeus replaced folk categories like "creeping things" and "things that fly" with a hierarchical taxonomy grounded in shared morphological features, a system that worked so well that Darwin, nearly a century later, was able to explain why the categories Linnaeus had identified were tracking actual lines of biological descent. The good representation came first; the theoretical justification followed. This sequence matters because it reveals that the joints are real, that finding them constitutes a discovery rather than an invention.
The Interactive Grid
The principle extends beyond natural science to the design of tools. When Dan Bricklin created VisiCalc, the first electronic spreadsheet, he did not invent new mathematics or enable calculations that were previously impossible. Every operation a spreadsheet performs could already be done by hand or with a calculator. What Bricklin discovered was a representation, the interactive grid of cells, formulas, and references, that carved quantitative reasoning at its natural joints.
The grid made it trivial to ask "what if" questions, to see how changing one assumption propagated through an entire financial model, to reason about scenarios that would have required days of manual recalculation. The representation did not add computational power in any absolute sense. What it did was make existing power cognitively accessible by organizing it in a way that aligned with how people naturally think about relationships between quantities.
Plato's butcher image is memorable because it is both violent and precise, a stark reminder that categories are not neutral containers but cuts into reality that either align with its structure or violate it. What the history of intellectual progress reveals is that finding the right cut is not preparatory work that precedes understanding but the very substance of it. To name something correctly, to identify the categories that correspond to genuine distinctions rather than arbitrary ones, is to make it available for thought and action in a way it was not available before. The joints are there, waiting to be found.