Can a Spiral Be True Geometry? Descartes’ Controversial Answer
A 2,000-Year-Old Puzzle: Which Curves Belong in Geometry?
It is 1588. A newly printed book of ancient Greek mathematics lands on scholars’ desks across Europe. Its author, Pappus of Alexandria, had lived around 300 CE, and in his Collection he described how the ancient Greeks divided geometrical problems into three kinds. Some problems, he reported, could be solved using only straight lines and circles—those were called plane problems. Others required cutting cones to get curves like the parabola or ellipse; those were solid problems. Still others needed stranger curves, such as spirals and quadratrices; the Greeks called these line-like problems.
But Pappus left a burning question unanswered: were solid and line-like problems really geometry? The ancient geometers seemed to accept only straightedge-and-compass constructions as beyond doubt. Everything else—conic sections, spirals, curves that required a point to move along a rotating line—was in a fuzzy zone. The book did not say clearly whether those complex curves met geometry’s rigorous standards.
For early modern mathematicians like Christoph Clavius (1537–1612) and François Viète (1540–1603), the puzzle was not just about dusty texts. Outstanding challenges like trisecting an angle or squaring a circle could be solved only if certain non‑straightedge‑and‑compass curves were allowed. The whole map of what problems counted as solved depended on which curves you let into the club.
Why the Spiral Caused So Much Trouble
Take the problem of squaring the circle: to construct a square whose area equals a given circle’s area. The ancient Greek Archimedes had answered it by using a curve called the spiral, generated when a point moves at a steady speed along a rotating line. Pappus’s own quadratrix, another curve meant to square the circle, was described by two moving line segments marching in different ways at once. But many mathematicians worried. To generate such curves, you had to think about two motions happening simultaneously—a circle rotating and a point sliding—and the relationship between a straight line and a curved one seemed impossible to nail down with perfect exactness.
Clavius tried to rescue the quadratrix with a pointwise construction. Instead of imagining two motions that had to stay in sync, he divided a quarter‑circle and its arc into tiny pieces, found piercing points that were easy to construct with straightedge and compass, and then connected the dots. At first he claimed this method made the quadratrix “truly geometrical,” but later he backed off, calling it only geometrical “in a certain way.”
Viète took a different path. He proposed a new fundamental move, the neusis postulate: suppose you can always draw a line through a fixed point that cuts off a given length between two existing lines. If you accept that move, you can solve the ancient problems of trisecting an angle and constructing two mean proportionals without ever touching a spiral or a quadratrix. Yet Viète still refused to admit the spiral or quadratrix themselves into genuine geometry. For him, if a curve could not be generated by that neat neusis step, it was not described “in the way of true knowledge.” Squaring the circle, which required those banned curves, remained unsolved.
Descartes’ First Big Idea: Only Clear Motions Count
In March 1619, a 23‑year‑old René Descartes (1596–1650) wrote to his friend Isaac Beeckman about an “entirely new science” that would solve any problem about quantity. In geometry, he said, some problems could be solved by straight lines and circles alone; others required curves that come from a single motion, like those drawn by new kinds of compasses; and still others needed curves, such as the spiral and quadratrix, that are certainly only imaginary. That third pile, Descartes warned, does not belong to legitimate geometry—just as some arithmetic problems “can be imagined but not solved.”
Descartes was doing something bold. He took the ancient descriptive label “line‑like” and turned it into a normative border: if a curve could not be traced by a single motion or by a chain of motions each completely determined by the one before, it was not genuinely geometrical. A clear and distinct motion became his badge of entry.
To prove his point, he built extraordinary compasses. One, the mesolabe, looked like a series of hinged rulers that slid and pushed one another; when you changed one angle, the intersection points of the rulers swept out ever more complex curves. Although the instrument was mechanical, the motion it produced was a single, smooth unfolding that, Descartes argued, the mind could grasp just as clearly as the motion that draws a circle. So curves made this way deserved a place in geometry. The spiral did not, because its description forced you to think about a straight‑line speed and a circular speed at the same time, with a ratio between straight and curved that, he said, “I believe cannot be discovered by human minds.”
The Famous Test: The Pappus Problem and the New Geometry
Seventeen years later, Descartes put his ideas to work in the Geometry (1637). The great challenge was the Pappus problem, a locus problem: given a bunch of lines and angles, find the curve all of whose points satisfy a certain ratio of distances. Pappus had solved the cases with three or four lines, but the general case with any number of lines had stumped everyone.
Descartes attacked it with algebra. He treated unknown lengths as variables, just like known ones, and boiled the problem down to an equation in two unknowns. On paper, he could then generate as many points of the solution curve as he liked by plugging in values. In Book One he showed that a solution existed. In Book Two he promised the synthesis: to prove the curves themselves were geometric.
To be geometric, a curve had to be traceable by a continuous motion or successive motions that were fully determined. Descartes claimed that such curves always corresponded to a single algebraic equation, and he classified them by the degree of that equation. The simplest class—circles, parabolas, ellipses, hyperbolas—came from equations of degree 2; the next class from degrees 3 or 4, and so on. Solving the four‑line Pappus problem, for example, yielded a second‑degree equation, so the locus was merely a conic section, already a member of Class I. For more lines, the degree rose, but the curve always landed in one of his legitimate classes. Algebra, he thought, handed him a perfect fence between the curves that reason could accept and the murky “mechanical” ones.
Where Descartes’ Plan Started to Crumble
Yet cracks soon appeared. For the five‑line Pappus problem, Descartes actually constructed the curve by plotting many individual points—a pointwise construction—rather than by a single continuous sweep. He insisted this was totally different from the pointwise construction of a spiral. With a spiral, he argued, you can only find a limited set of points using straightedge and compass (for instance, you can only divide an arc into 2, 4, 8, and so on, parts). With his “geometric” curves, you can determine any arbitrary point by solving the equation. That, he declared, was as good as being drawn by a single motion.
The trouble is that he never proved this jump. He simply equated a curve you can study point‑by‑point with one produced by a clear motion, leaving a gap in his own demonstration. If the pointwise method counted, why not admit other pointwise curves like Clavius’s quadratrix? Descartes’ answer—that some pointwise constructions are “generic” and others are not—was clever but not airtight.
Some scholars suspect there was a deeper reason. In letters after the Geometry, Descartes admitted that he could, in fact, clearly conceive a straight line becoming curved, like a piece of string. Yet in his book he had said the ratio between a straight and a curved line “cannot be discovered by human minds.” Why the flip‑flop? A likely explanation is that Descartes was determined to keep the Archimedean spiral out of geometry because accepting it would mean squaring the circle was possible after all—and he believed that problem was impossible. His official “clear motion” rule may have served a hidden agenda.
Why It Still Matters When You Draw a Curve Today
Next time you use a graphing calculator or drag your finger to draw a swirling line on a screen, you are stepping into Descartes’ old debate. Computers often draw curves by calculating thousands of tiny points—a massive pointwise construction. Is that curve “exactly geometrical,” or is it just a clever approximation? Descartes would have insisted that a curve’s real legitimacy hangs on whether the motion that produced it can be grasped with perfect clarity. The algebra alone, he thought, wasn’t enough.
Even now mathematicians argue about whether computer‑assisted proofs are fully rigorous, or whether a curve described only by an infinite series of points is as solid as one traced by a flowing movement. The fight that started in a 1588 translation of an ancient book still echoes whenever we ask what counts as an acceptable answer. So when you slide your pencil along a ruler and then try to sketch a smooth spiral freehand, you might feel a tiny version of Descartes’ puzzle: which one feels more real to you?
Think about it
- If you could program a robot to trace a curve with a single smooth motion, would that curve be “geometric” even if nobody fully understood how the robot’s gears worked? What matters more—the motion or your understanding of it?
- Take a spiral staircase: when you trace its shape with your finger, you mix a circular motion with an upward motion. Does that mixture feel less clear and exact than drawing a simple circle? Why or why not?
- A modern computer plots a heart‑shaped curve by lighting up millions of separate pixels. To you, does it still feel like a single, wholehearted shape—or does knowing how it is made change your sense of its “exactness”?
