I’m not a mathematician by training, but the language and (for want of a better term) the *sport* of geometry has always had a special appeal for me. I wasn’t a whiz at algebra in high school, but I aced geometry. As a homeschooling parent, I had a wonderful time teaching geometry to our three kids. I still find geometry intriguing.

When I was in high school, I spent hours trying to figure out how to trisect an angle with compass and straightedge. I knew that nobody had found a way to do it. As it turns out, in 1837 (before even my school days) French mathematician Pierre Wantzel proved that it was impossible for the general case (trisecting certain special angles is trivial). I’m glad I didn’t know that, though, since it gave me a certain license to hack at it anyway. Perhaps I was motivated by a sense that it would be glorious to be the first to crack this particular nut, but mostly I just wondered, “Can it be done, and if not, why not?”

Trisecting the angle is cited in Wikipedia as an example of “pseudomathematics”, and while I will happily concede that any claim to be able to do so would doubtless rely on bogus premises or operations, I nevertheless argue that *wrestling* with the problem honestly, within the rules of the game, is a mathematical activity as valid as any other, at least as an exercise. I tried different strategies, mostly trying to find a useful correspondence between the (simple) trisection of a straight line and the trisection of an arc. My efforts, of course, failed (that’s what “impossible” means, after all). Had they not, my own name would be celebrated in different Wikipedia articles describing how the puzzle had finally been solved. It’s not. In my defense, I hasten to point out that I never was under the impression that I had succeeded. I just wanted to try and to know either how to do it or to know the reason why.

My failed effort might, by many measures, be accounted a waste of time. But was it? I don’t think it was. Its value for me was not in the achievement but in the striving. Pushing on El Capitan isn’t going to move the mountain, either, but doing it regularly will provide a measure of isometric exercise. Similarly confronting an impossible mental challenge can have certain benefits.

And so along the way I gained a visceral appreciation of some truths I might not have grasped as fully otherwise.

In the narrowest terms, I came to understand that the problem of trisecting the angle (either as an angle or as its corresponding arc) is fundamentally distinct from the problem of trisecting a line segment, because curvature — even in the simplest case, which is the circular — fundamentally changes the problem. One cannot treat the circumference of a circle as if it were linear, even though it is much like a line segment, having no thickness and a specific finite extension. (The fact that π is irrational seems at least obliquely connected to this, though it might not be: that’s just a surmise of my own.)

In the broadest terms, I came more fully to appreciate the fact that some things are intrinsically impossible, even if they are not obvious logical contradictions. You can bang away at them for as long as you like, but you’ll *never* solve them. This truth transcends mathematics by a long stretch, but it’s worth realizing that failing to accomplish something that you want to accomplish is not invariably a result of your personal moral, intellectual, or imaginative deficiencies. As disappointing as it may be for those who want to believe that every failure *is* a moral, intellectual, or imaginative one, it’s very liberating for the rest of us.

Between those obvious extremes are some more nuanced realizations.

I came to appreciate iterative refinement as a tool. After all, even if you can’t trisect the general angle with perfect geometrical rigor, you actually *can* come up with an imperfect but eminently practical *approximation —* to *whatever degree of precision you require*. By iterative refinement (interpolating between the too-large and the too-small solutions), you can zero in on a value that’s demonstrably better than the last one every time. Eventually, the inaccuracy won’t matter to you any more for any *practical* application. I’m perfectly aware that this no longer pure math — but it is the very essence of engineering, which has a fairly prominent and distinguished place in the world. Thinking about this also altered my appreciation of precision as a pragmatic real-world concept.

A more general expression of this notion is that, while some problems never have perfect solutions, they sometimes can be *practically* solved in a way that’s *good enough* for a given purpose. That’s a liberating realization. Failure to achieve the *perfect* solution needn’t stop you in your tracks. It doesn’t mean you can’t get a *very good* one. It’s worth internalizing this basic truth. And only by wrestling with the impossible do we typically discover the limits of the possible. That in turn lets us develop strategies for practical work-arounds.

Conceptually, too, iterative refinement ultimately loops around on itself and becomes a model for thinking about such things as calculus, and the strange and wonderful fact that, with limit theory, we can (at least sometimes) achieve exact (if occasionally bizarre) values for things that we can’t measure directly. Calculus gives us the ability (figuratively speaking) to bounce a very orderly sequence of successive refinements off an infinitely remote backstop and somehow get back an answer that is not only usable but sometimes actually *is* perfect. This is important enough that we now define the value of *pi* as the limit of the perimeter of a polygon with infinitely many sides.

It shows also that this is not just a problem of something being somehow too difficult to do: difficulty has little or nothing to do with intrinsic impossibility (*pace* the Army Corps of Engineers: they are, after all, engineers, not pure mathematicians). In fact we live in a world full of unachievable things. Irrational numbers are all around us, from *pi* to *phi* to the square root of two, and even though no amount of effort will produce a perfect rational expression of any of those values, they are not on that account any less real. You cannot solve *pi* to its last decimal digit because there *is* no such digit, and no other rational expression can capture it either. But the proportion of circumference to diameter is always exactly *pi*, and the circumference of the circle is an exact distance. It’s magnificently reliable and absolutely perfect, but its perfection can never be entirely expressed in the same terms as the diameter. (We could arbitrarily designate the circumference as 1 or any other rational number; but then the diameter would be inexpressible in the same terms.)

I’m inclined to draw some theological application from that, but I’m not sure I’m competent to do so. It bears thinking on. Certainly it has at least some broad philosophical applications. The prevailing culture tends to suggest that whatever is not quantifiable and tangible is not real. There are a lot of reasons we can’t quantify such things as love or justice or truth; it’s also in the nature of number that we can’t nail down many concrete things. None of them is the less real merely because we can’t express them perfectly.

Approximation by iterative refinement is basic in dealing with the world in both its rational and its irrational dimensions. While your inability to express *pi* rationally is not a failure of your moral or rational fiber, you may still legitimately be required — and *you will be able* — to get an arbitrarily precise approximation of it. In my day, we were taught the Greek value 22/7 as a practical rational value for *pi,* though Archimedes (288-212 BC) knew it was a bit too high (3.1428…). The Chinese mathematician Zhu Chongzhi (AD 429-500) came up with 355/113, which is not precisely *pi* either, but it’s more than a thousand times closer to the mark (3.1415929…). The whole domain of rational approximation is fun to explore, and has analogical implications in things not bound up with numbers at all.

So I personally don’t consider my attempts to trisect the general angle with compass and straightedge to be time wasted. It’s that way in most intellectual endeavors, really: education represents not a catalogue of facts, but a process and an exercise, in which the collateral benefits can far outweigh any immediate success or failure. Pitting yourself against reality, win or lose, you become stronger, and, one hopes, wiser.