The Axiomatic Method

1. The Axiomatic Method The 2004 Isuzu Axiom “Old Gassy” (Greenhouse Gassy, that is)

2. As far as the laws of mathematics refer to reality, they are not certain, and as far as they are certain, they do not refer to reality. Albert Einstein Mathematics may be defined as the subject in which we never know what we are talking about, nor whether what we are saying is true. Bertrand Russell

3. What Is It? The Axiomatic Method is the procedure by which an entire science or system of theorems is deduced in accordance with specified rules by logical deduction from certain basic propositions (axioms), which in turn are constructed from a few terms taken as primitive. These terms may be either arbitrarily defined or conceived according to a model in which some intuitive warrant for their truth is felt to exist. (http://concise.britannica.com/ebc/article?tocId=9356243)

4. Tell Me More! To axiomatize a subject, we need to specify three things: TERMS: These are the fundamental objects we’re talking about. We usually choose them to be meaningful to us, but the point of the method is that we don’t get to import our intuitions into our proofs. AXIOMS: These are the statements (about the terms) that we assume from the get-go are true. They are theorems; they are also freebies. Again, we usually pick axioms that make sense to us, but we shouldn’t make ANY assumptions that aren’t explicitly covered as axioms.

5. That Was Only Two Things RULES OF INFERENCE: These are the rules that tell us how to use theorems to make other theorems. If we want to, we can usually get by with just one rule of inference: Modus Ponens (MP). However, doing so means that we have to introduce lots more axioms. In general, that’s the trade-off: fewer rules means more axioms. We’ll usually want more rules around (since we know more valid rules than just MP), but it’s sometimes useful – and interesting – that we could get by with just the one.

6. Just Any Old Rules of Inference? It’s important that our rules be reasonable, that is, that they don’t produce inconsistencies. Here’s an example of an unreasonable rule of inference: RULE OF IDIOCY: If A is a theorem, then so is not A. If we have even one axiom A, this rule makes our system inconsistent; the system claims that A and not A are both true, which is impossible.

7. Euclidean Geometry • Terms: Point, Line, Plane. • Axioms: (usually called Postulates in this case) • It is possible to draw a straight line from any point to any point. • It is possible to produce a finite straight line continuously in a straight line. • It is possible to describe a circle with any center and radius. • All right angles are equal to one another. • Given a line and a point not on the line, there is exactly one parallel through the point (that is, exactly one line through the point that is parallel to the given line).

8. More Euclidean Geometry It’s not quite true that we have to specify the rules of inference. Euclid didn’t; he just used normal reasoning (“common notions”) and didn’t get into trouble. (Actually, he did get into trouble with 19th-century mathematicians, but it had nothing to do with rules of inference; he tacitly assumed several facts that he should have written down as axioms. Hilbert’s Foundations of Geometry (1899) rectified those problems by making the tacit assumptions explicit.)

9. Non-Euclidean Geometry Terms: Point, Line, Plane. Axioms: 1-4 from before, and we replace 5 with something else. If we say there are NO parallels, we get Riemannian Geometry! If we say there are SEVERAL parallels, we get Hyperbolic Geometry! Note that these aren’t “real” geometry.

10. Arithmetic/Number Theory Terms: Number, 0, S, +, *. Axioms: 0 is a number. If x is a number, Sx (the successor of x) is a number. 0 isn’t Sx for any number x. If Sx = Sy, then x = y. Induction: If a set A of numbers contains 0 and contains Sx for every x in A, then A contains every number. For all numbers x and y, we have: x + 0 = x; x + Sy = S(x+y) x * 0 = 0; x * Sy = (x * y) + x

11. Principia Mathematica This is another famous example, in which Russell and Whitehead tried to set down axioms for Mathematical Logic (and, by extension, for all of Mathematics). As you know from reading GEB, it didn’t work out in quite the way they’d hoped. Gödel showed that even arithmetic isn’t fully axiomatizable, let alone all of Mathematics. Yikes! Still, it was a big step forward in mathematical thought; it was a “proof of concept” sort of deal that put Mathematics on sturdy axiomatic ground.

12. Nobody Said It Would Be Easy

13. Hofstadter’s MIU-System • Terms: M, I, U. • Axiom: MI. • Rules of Inference: • If xI is a theorem, so is xIU. • If Mx is a theorem, so is Mxx. • In any theorem, III can be replaced by U. • UU can be dropped from any theorem. • This looks different from the other examples because it’s only about producing strings. But we can make the other ones just as formal:

14. Arithmetic/Number Theory In Formal Attire Terms: Number, 0, S, +, *. Axioms: x (x = 0) x y (y = Sx) x (0  Sx) x y (Sx = Sy  x = y) [A(0) & x (A(x)  A(Sx))]  x A(x) x (x + 0 = x) x y (x + Sy = S(x+y)) x (x * 0 = 0) x y (x * Sy = (x * y) + x)

15. Hofstadter’s pq-System Terms: p, q, -. Axioms: xp-qx- is an axiom whenever x is composed of hyphens only. Rule of Inference: If x, y, and z are strings containing only hyphens, and if xpyqz is a theorem, then xpy-qz- is a theorem. So, for example, -p-q-- and --p-q--- are both axioms, and from them we can use the rule to get the theorems -p--q--- and --p--q----.

16. But What Does It All MEAN? How can we make sense of the theorems -p-q--, --p-q---, -p--q---, ---p-q----, and so on? - = 1; p = “+”; q = “=” - = 1; p = “=”; q = “taken from” - = “apple”; p = “horse”; q = “happy” So -p-q-- becomes “apple horse apple happy apple apple”. This is prolly a valid interpretation if you’re a horse. The point: if the axioms and rules make sense under an interpretation, then all theorems must hold under that interpretation.

17. This Talk Needs More Pictures Chicken Liver Bake: enjoy it with the ashes of a loved one. Or maybe what's left of the chickens are in that urn. Maybe the chickens were your loved ones. But chickens never love back enough. And that's why you have to KILL them. And eat their livers ritualistically. And then they're a part of you forever. Forever. http://www.candyboots.com/wwcards/chickenliverbake.html

18. Why Bother? Certainty: If you believe that your axioms are correct and you believe that your rules of inference are valid, then you can confidently believe the theorems of the system. Said another way, if you derive a theorem that’s false, then either one of your axioms is false or one of your rules of inference is faulty. Applicability: If you find another system that satisfies the axioms, then ALL results that follow from the axioms are true about your system.

19. One More Reason To Bother Mechanization: If we axiomatize a subject formally enough, we can have computers derive theorems for us. Sometimes (not very often, actually) they obtain results that humans hadn’t discovered yet. Another bonus: the axiomatic approach to Number Theory paved the way for Gödel’s Incompleteness Theorems. His idea was to encode theorems as numbers – but that only works if the theorems are formal objects, and his proof is specifically about formal systems.

20. As far as the laws of mathematics refer to reality, they are not certain, and as far as they are certain, they do not refer to reality. Albert Einstein Mathematics may be defined as the subject in which we never know what we are talking about, nor whether what we are saying is true. Bertrand Russell

21. FIN http://qwantz.com/