Reading the Comics, October 22, 2015: Foundations Edition

I am, yes, saddened to hear that Apartment 3-G is apparently shuffling off to a farm upstate. There it will be visited by a horrifying kangaroo-deer-fox-demon. And an endless series of shots of two talking heads saying they should go outside, when they’re already outside. But there are still many comic strips running, on and on Comics Kingdom. They’ll continue to get into mathematically themed subjects. And best of all I can use a Popeye strip to talk about the logical foundations of mathematics and what computers can do for them.

Jef Mallett’s Frazz for the 18th of October carries on the strange vendetta against “showing your work”. If you do read through the blackboard-of-text you’ll get some fun little jokes. I like the explanation of how “obscure calculus symbols” could be used, “And a Venn diagram!” Physics majors might notice the graph on the center-right, to the right of the DNA strand. That could show many things, but the one most plausible to me is a plot of the velocity and the position of an object undergoing simple harmonic motion.

Still, I do wonder what work Caulfield would show if the problem were to say what fraction were green apples, if there were 57 green and 912 red apples. There are levels where “well, duh” will not cut it. In case “well, duh” does cut it, then a mathematician might say the answer is “obvious”. But she may want to avoid the word “obvious”, which has a history of being dangerously flexible. She might then say “by inspection”. That means, basically, look at it and yeah, of course that’s right.

Jeffrey Caulfield and Alexandre Rouillard’s Mustard and Boloney for the 18th of October uses mathematics as the quick way to establish “really smart”. It doesn’t take many symbols this time around, curiously. Superstar Equation E = mc2 appears in a misquoted form. At first that seems obvious, since if there were an equals sign in the denominator the whole expression would not parse. Then, though, you notice: if E and m and c mean what they usually do in the Superstar Equation, then, “E – mc2” is equal to zero. It shouldn’t be in the denominator anyway. So, the big guy has to be the egghead.

Peter Maresca’s Origins of the Sunday Comics for the 18th of October reprints one of Windsor McCay’s Dreams of the Rarebit Fiend strips. As normal for Dreams and so much of McCay’s best work, it’s a dream-to-nightmare strip. And this one gives a wonderful abundance of numerals, and the odd letter, to play with. Mathematical? Maybe not. But it is so merrily playful it’d be a shame not to include.

Zach Weinersmith’s Saturday Morning Breakfast Cereal for the 20th of October is a joke soundly in set theory. It also feels like it’s playing with a set-theory-paradox problem but I can’t pin down which one exactly. It feels most like the paradox of “find the smallest uninteresting counting number”. But being the smallest uninteresting counting number would be an interesting property to have. So any candidate number has to count as interesting. It also feels like it’s circling around the heap paradox. Take a heap of sand and remove one grain, and you still have a heap of sand. But if you keep doing that, at some point, you have just one piece of sand from the original pile, and that is no heap. When does the heap move away?

Daniel Shelton’s Ben for the 21st of October is a teaching-arithmetic problem, using jellybeans. And fractions. Well, real objects can do wonders in connecting a mathematical abstraction to something one has an intuition for. One just has to avoid unwanted connotations and punching.

Doug Savage’s Savage Chickens for the 21st of October uses “mathematics homework” as the emblem of the hardest kind of homework there might ever be. I saw the punch line coming a long while off, but still laughed.

The Sea Hag is dismissive of scientists, who try to take credit for magic even though 'they can't even THINK! They have to use machines to tell them what two plus two is!! And another machine to prove the first one was right!'

Bud Sagendorf’s Popeye for the 22nd of October, 2015. The strip originally ran sometime in 1981. This would be only a few years after the Four-Color Theorem was solved by computer. The computer did this by trying out all the possibilities and reporting everything was OK.

Bud Sagendorf’s Popeye began what it billed as a new story, “Science Vs Sorcery”, on Monday the 19th. I believe it’s properly a continuation of the previous story, though, “Back-Room Pest!” which began the 13th of July. “Back-Room Pest!”, according to my records, originally ran from the 27th of July, 1981, through to the 23rd of January, 1982. So there’s obviously time missing. And this story, like “Back-Room Pest”, features nutty inventor Professor O G Wotasnozzle. I know, I know, you’re all deeply interested in working out correct story guides for this.

Anyway, the Sea Hag in arguing against scientists claims “they can’t even think! They have to use machines to tell them what two plus two is!! And another machine to prove the first one was right!” It’s a funny line and remarkably pointed for an early-80s Popeye comic. The complaint that computers leave one unable to do even simple reasoning is an old one, of course. The complaint has been brought against every device or technique that promises to lighten a required mental effort. It seems to me similar to the way new kinds of weapons are accused of making war too monstrous and too unchivalrous, too easily done by cowards. I suppose it’s also the way a fable like the story of John Henry hold up human muscle against the indignity of mechanical work.

The crack about needing another machine to prove the first was right is less usual, though. Sagendorf may have meant to be whimsically funny, but he hit on something true. One of the great projects of late 19th and early 20th century mathematics was the attempt to place its foundations on strict logic, independent of all human intuition. (Intuition can be a great guide, but it can lead one astray.) Out of this came a study of proofs as objects, as mathematical constructs which must themselves follow certain rules.

And here we reach a spooky borderland between mathematics and sorcery. We can create a proof system that is, in a way, a language with a grammar. A string of symbols that satisfies all the grammatical rules is itself a proof, a valid argument following from the axioms of the system. (The axioms are some basic set of statements which we declare to be true by assumption.) And it does not matter how the symbols are assembled: by mathematician, by undergrad student worker, by monkey at a specialized typewriter, by a computer stringing things together. Once a grammatically valid string of symbols is done, that string of symbols is a theorem, with its proof written out. The proof is the string of symbols that is the theorem written out. If it were not for the modesty of what is claimed to be done — proofs about arithmetic or geometry or the like — one might think we had left behind mathematics and were now summoning demons by declaring their True Names. Or risk the stars overhead going out, one by one.

So it is possible to create a machine that simply grinds out proofs. Or, since this is the 21st century, a computer that does that. If the computer is given no guidance it may spit out all sorts of theorems that are true but boring. But we can set up a system by which the computer, by itself, works out whether a given theorem does follow from the axioms of mathematics. More, this has been done. It’s a bit of a pain, because any proofs that are complicated enough to really need checking involve an incredible number of steps. But for a challenging enough proof it is worth doing, and automated proof checking is one of the tools mathematicians can now draw on.

Of course, then we have the problem of knowing that the computer is carrying out its automatic-proof programming correctly. I’m not stepping into that kind of trouble.

The attempt to divorce mathematics from all human intuition was a fruitful one. The most awe-inspiring discovery to come from it is surely that of incompleteness. Any mathematical system interesting enough will contain within it statements that are true, but can’t be proven true from the axioms.

Georgia Dunn’s Breaking Cat News for the 22nd of October features a Venn Diagram. It’s part of how cats attempt to understand toddlers. My understanding is that their work is correct.