The Infinite Radius
May 30th, 2010Not all my thoughts on infinity involve trying to teach the concepts to a six-year-old. My daughter’s questions are definitely what got me thinking about infinity again, but this is a question that has been intriguing me for years – since high school, in fact. I’ve tried to pose it to others many times over the years. Some “don’t want to think that hard”, and others, many of whom are much better at math than I am, dismiss it without giving me a satisfactory reason why. Of course, what may be satisfactory to a math expert may not be satisfactory to me.
Anyone with knowledge of high school algebra should be able to follow this. Whether you want to or not is another matter. But, to those who do choose to take an interest, I have a request. Can you tell me what you think? If you have no opinion, can you refer this article to someone who might? I’d love to have a genuine mathematical perspective on this, if one exists.
It’s a well-known fact that you can’t get a meaningful answer dividing any number by zero. Even dividing zero by zero is problematic. So, look at dividing 1 by any number greater than zero: the smaller the denominator (the number on the bottom), the greater the result. It would appear that 1/x approaches infinity as x approaches zero. Now, the fact that infinity isn’t a real-world number might be enough reason to question that answer. You can’t fit an infinite number of anything in the known universe, and you can’t plot infinity meaningfully on any conventional graph.
But it’s worse than that. If you approach x = 0 from the negative side, the result approaches negative infinity. The graph of y = 1/x looks roughly like the drawing below. It illustrates a bizarre result, scornfully challenging intuition to get any grasp on it whatsoever. If you approach zero from the positive side, 1/x approaches positive infinity. If you approach zero from the negative side, 1/x approaches negative infinity. This is why, when you ask most people what 1/0 is, they’ll say that it’s “undefined”. Math functions in microchips have a special error condition for dividing by zero – they won’t even attempt it. Who can blame them? What two “numbers” could be further apart than positive infinity and negative infinity?
Now, let me digress for a moment, and talk about the most common way to produce flat glass. It was developed in the mid-twentieth century, and involves letting the glass solidify on a pool of molten tin, or some other metal with a melting point lower than that of glass. The glass comes out a uniform thickness, and very, very flat. The glass isn’t actually flat, however. It’s only as flat as the pool of tin, which has the curvature of the earth. However, even on the scale of a very large window, the curvature of the earth is very slight. By a quick calculation, you’d need a window about 70 feet long for it to dip a tenth of an inch. That is very, very flat.
The point is, the greater the radius of curvature, the flatter the curve. So, what might an infinite radius of curvature yield? Might we not get a perfectly flat curve? A circle with an infinite radius might be the same as a line. A sphere with an infinite radius might be equivalent to a plane. Certain non-Euclidean geometries might become Euclidean again.
To further clarify (I hope) the issue, consider if a two-dimensional space were mapped to the surface of a sphere, as shown below:
This becomes a finite two-dimensional space, but it’s one that I’ve seen quite often in lay discussions of non-Euclidean spaces. The X and Y axes here are great circles on the sphere, perpendicular to each other, and they meet both at the “origin” (arbitrarily chosen) and at the maximum distance from the point, halfway around any great circle passing through that point. We can’t call this point infinity, because it’s a finite space. It should be noted that this is a two-dimensional space, equivalent to a plane (or part of a plane), and traveling through the sphere is not possible – entities in that space can travel only along the outside of the sphere.
Along the sphere, we can map a function like y = 1/x, which meets at the maximum distance on the other side of the sphere. I won’t bother figuring out what the function is – we can define it very artificially if we want to. But such a function would plot something like this:
Now, I think those of you who’ve kept reading probably know where I’m going with this. The bigger the sphere gets, the flatter the curve gets, and the more the actual function mapped can resemble y = 1/x. If the radius is infinite, then each great circle could be a straight line, and the sphere could be a plane. Granted, this is a fudge. I don’t think that the appearance of a totally flat surface is the only possibility. Multiply the infinite radius by two pi to get circumference, and you get the exact same infinity. It’s hard to get a definite shape from that.
This reminds me of problems I heard of in some quantum theory models, where infinities are canceled out by dividing them by other infinities. It’s mathematically possible for them to work out, but not mathematically required – so it feels messy.
But, all messiness aside, if you do think of the number line as a circle of infinite radius, is it not possible for infinity and negative infinity to occupy the same point on a number line – and thus, in effect, to be the same “number”? If we allow this, it either makes better intuitive sense of the y = 1/x equation as x approaches zero, or it wreaks havoc with the concept of infinity, or at least the intuitive sense of it. Maybe it does both.
Is it possible that the transfinite numbers transcend positive and negative? Is infinity just too big to have a plus or minus sign attached to it? What other implications might such a trans-Euclidean geometry have? Anyway, that’s about all I have for now. So, what do you think?




