KininarunoteA notebook of curious things

This site is only a doorway

Everything here is gathered from things other people actually researched, thought through, and built: books, papers, historical records, and the people behind them. If something here catches your interest, go look at the original sources listed at the bottom of the page. They carry more depth than a page like this ever can.

How Did Mathematicians Solve a 100-Year-Old Question About the Shape of Space, and Know the Answer Was Right?

Category: History & Culture

Stretch a rubber band around an apple and you can slide it, without lifting it off the skin, until it shrinks to a point. Try the same on a doughnut, with the band threaded through the hole, and you cannot. In 1904 the French mathematician Henri Poincaré asked whether this “always shrinks” test is enough to recognize a ball in three dimensions, not just in the flat skin of an apple. That question took about a century to answer.

The short version: yes. A closed 3D space in which every loop shrinks must be the 3D sphere.Grigori Perelman, a Russian mathematician, sketched a proof in papers posted online in 2002 and 2003. His method built on an idea by Richard Hamilton for smoothing out a shape over time. Then other mathematicians spent years filling in the missing steps, and only after that was the result widely accepted. The rest of this page follows how the question was asked, how the proof worked, and how anyone could be sure.

What Did Poincaré Actually Ask in 1904?

The Clay Mathematics Institute, which listed this problem among its seven million-dollar “Millennium Prize” problems, explains it with a similar rubber-band picture. An ideal sphere, like the surface of a perfectly smooth ball, is called simply connected, because every loop on it can be shrunk to a point. A doughnut’s surface is not simply connected: a loop threaded through its hole cannot shrink to a point. Poincaré knew that a two-dimensional sphere is essentially characterized by this property, and he asked the matching question for the three-dimensional sphere [1].

A “three-dimensional sphere” is not a ball you can hold. A ball’s skin is two-dimensional, and the 3D sphere is the 3D version of it, the next step up: a space with no edge and a finite size, in which every small neighborhood looks like ordinary 3D space. We cannot see such a space from outside, so mathematicians work with equations and careful definitions instead of pictures. The question then reads: if a closed space like that has the “every loop shrinks” property, must it be the 3D sphere? This is the Poincaré conjecture [1].

Why Was the 3D Case the Last One Solved?

You might guess that higher dimensions would be harder. But the higher-dimensional cases were solved first. According to Wikipedia’s summary of the history, the matching question was settled for five dimensions and up by Stephen Smale in the early 1960s, and for four dimensions by Michael Freedman in 1982 [2]. Three dimensions came last.

Why? Science News says the extra dimensions give more “wiggle room” [6]. Take that as a picture, not a full explanation: moving furniture is easier in a big room than in a small one.

How Does Smoothing a Shape Over Time Help?

In 1982 Richard Hamilton introduced a method called Ricci flow. A simple way to picture it: take a lumpy space and let its bumps and dents even out gradually, the way heat spreads until the temperature is the same everywhere. Wikipedia notes that Ricci flow behaves similarly to the heat equation, though the two are not the same equation [2]. If the lumps always smoothed away, a space with the “every loop shrinks” property would settle into a perfectly round 3D sphere.

The catch is that the smoothing sometimes goes wrong. In places, the curvature can grow without limit in a finite time, and the equation can no longer be followed. These trouble spots are called singularities [2]. For years that was the wall: the smoothing idea was promising, but it broke before it finished.

How Did Perelman Get Past the Breaking Points?

Perelman posted his first paper on arXiv, a free site where researchers share work before it is formally published, in November 2002. Its abstract describes a quantity that changes in only one direction as the flow runs, in any dimension, and that he interprets as an entropy. It also sketches a route to the larger geometrization conjecture of William Thurston [3].

In March 2003 came a second paper, which he calls technical and a continuation of the first. In it he builds Ricci flow “with surgeries” [4]. A simple way to picture this: when the flow is about to break at a thin neck, cut the neck, cap the open ends, and keep going. (That sentence is my picture, not his wording.) A third short paper, from July 2003, shows that for certain spaces the flow with surgery ends in finite time [5].

The Clay Institute’s summary is that Perelman’s solution, announced in preprints on arXiv in 2002 and 2003, was based on Hamilton’s theory of Ricci flow, and that he also proved Thurston’s conjecture, of which Poincaré’s is a special case [1].

Who Checked the Proof, and How Long Did It Take?

Perelman never sent his papers to a journal; he only posted them online. So the checking fell to others. Science News reported that dozens of mathematicians worked for several years to fill in details and verify the proof, producing several book-length expositions. Three are named: by Bruce Kleiner and John Lott, by John Morgan and Gang Tian, and by Huai-Dong Cao and Xi-Ping Zhu [6].

Two of them are on arXiv in full. Kleiner and Lott posted detailed notes on the first two papers in 2006 (216 pages in the final version, later published in the journal Geometry & Topology in 2008) [7]. Morgan and Tian posted a 493-page manuscript the same year, described as a detailed proof built on Perelman’s three preprints, revised after comments from referees [8]. Page counts are from the latest versions and may differ in earlier ones.

The pattern is worth noticing. Perelman published a concise proof that built on Hamilton’s work. Many other mathematicians, working for years, then made every step explicit enough that an expert could follow it.

When Did the World Treat the Problem as Solved?

In December 2006 the journal Science named the proof of the Poincaré conjecture its “Breakthrough of the Year” [9]. In March 2010 the Clay Institute announced that it was awarding Perelman its first Millennium Prize, worth one million dollars. The Institute’s rules require a wait of two years after publication before a prize, and it counted from when the expositions by others appeared [6]. Of the seven Millennium problems, this is the one with a verified solution [2].

The story has a quiet ending. Science News reports that Perelman had already declined the 2006 Fields Medal and withdrawn from the mathematical community [6], and Wikipedia says he also declined the prize, reportedly saying that Hamilton’s contribution was equal to his own [2]. We only have secondhand accounts of his reasons, so it is best not to guess further.

How Can You Test the “Shrinking Loop” Idea Yourself?

  • Loop a piece of string loosely around an apple, an orange, or a ball, or just draw a loop on a picture of one and imagine sliding it smaller without lifting it. Does it always shrink to a point?
  • Now find something with a hole, such as a mug handle or a bagel, and thread the loop through the hole. What stops it from shrinking?
  • Next time you solve a math problem, write down every step and why, and give it to someone to read. Wherever they get stuck is a step you skipped. Checking Perelman’s work was that, on a much larger scale.
  • If you want to look at the real thing, the arXiv abstracts below are free to read. They are written for specialists, but the first few lines of each are surprisingly clear.

Sources

What I could read: the Clay Mathematics Institute page, the four arXiv abstract pages, Science News and Wikipedia were fetched and read. The Science article page loaded only its summary line and a reader comment, so I did not read the article itself; I used it only for the title, the topic and the 2006 date. I could not open the Clay press release of March 18, 2010; the prize date and wording rest on Science News. The 1982 Hamilton paper and the Smale and Freedman dates were not checked against the original papers, so they rest on Wikipedia, a secondary source anyone can edit. The picture of cutting and capping a neck is my own plain-language image, not a quotation. Science News describes the conjecture in confusing terms about dimensions, so I used it only for the checking process, the prize rules, the “wiggle room” remark and Perelman’s withdrawal. The claim that Perelman declined the prize rests on Wikipedia alone and is worded cautiously.

  1. Clay Mathematics Institute, “Poincaré Conjecture.” https://www.claymath.org/millennium/poincare-conjecture/ (rubber-band explanation, 1904 question, Perelman’s preprints of 2002 and 2003, Hamilton’s Ricci flow, Thurston’s conjecture. Research institute page; overview read.)
  2. Wikipedia, “Poincaré conjecture.” https://en.wikipedia.org/wiki/Poincar%C3%A9_conjecture (summary of the history, heat-equation comparison, singularities, prize declined. Secondary source; opening part read.)
  3. G. Perelman, “The entropy formula for the Ricci flow and its geometric applications,” arXiv:math/0211159. https://arxiv.org/abs/math/0211159 (abstract only: monotonic quantity, entropy, geometrization sketch.)
  4. G. Perelman, “Ricci flow with surgery on three-manifolds,” arXiv:math/0303109. https://arxiv.org/abs/math/0303109 (abstract only: technical continuation, construction of flow with surgeries.)
  5. G. Perelman, “Finite extinction time for the solutions to the Ricci flow on certain three-manifolds,” arXiv:math/0307245. https://arxiv.org/abs/math/0307245 (abstract only.)
  6. Science News, “Million-dollar math prize awarded, but not necessarily accepted.” https://www.sciencenews.org/article/million-dollar-math-prize-awarded-not-necessarily-accepted (2010 prize announcement, two-year rule, other mathematicians’ expositions, Fields Medal declined. News report; text read.)
  7. B. Kleiner and J. Lott, “Notes on Perelman’s papers,” arXiv:math/0605667; Geometry & Topology 12 (2008). https://arxiv.org/abs/math/0605667 (abstract page only: 216 pages, journal reference.)
  8. J. Morgan and G. Tian, “Ricci Flow and the Poincaré Conjecture,” arXiv:math/0607607. https://arxiv.org/abs/math/0607607 (abstract page only: 493 pages, detailed proof, revised after referees.)
  9. “The Poincaré Conjecture—Proved,” Science 314 (2006). https://www.science.org/doi/10.1126/science.314.5807.1848 (summary line and page metadata only; article text not read.)

Update history

  • First published.