ags@pucc-i (Seaman) (02/25/84)
There is a way to define exponentiation which sidesteps some of the problems I outlined in my previous note. I think it is worth mentioning, because it leads to a simple proof that n < 2**n for all n, using a variation of Cantor Diagonalization, which is what started the whole discussion. For openers, let's consider the Zermelo-Fraenkel model of the natural numbers. According to this model, (0) Zero is the empty set, 0 = {}. (1) One is the set {0} = {{}}. (2) Two is the set {0,1} = {{}, {{}}}. (3) Three is the set {0,1,2} = {{} ,{{}} ,{{{},{{}}}}. ... (n) Is the set {0,...,n-1}. In this model, numbers are sets of a particular form. The successor of a number n is the set n+{n}, "n union singleton n", where + represents set union. It is rather pleasant to see how everything can be built out of nothing. Exercise: Show that the Peano Axioms are satisfied by this model. Exercise: Show that any finite set can be placed in one-to-one correspondence with exactly one of the natural numbers, viewed as a set. This number is called the CARDINALITY of the set. The concept of cardinality can be carried to transfinite sets, but we don't need that here. Exercise: Show that multiplication of natural numbers can be defined in terms of the cardinality of the CARTESIAN PRODUCT of the two numbers, viewed as sets. (If you don't know what a Cartesian product is, skip this exercise. It's just a curiosity which won't be needed.) Exercise: Think of a similarly clever way to define exponentiation in terms of sets. Answer to follow in next article. -- Dave Seaman ..!pur-ee!pucc-i:ags "Against people who give vent to their loquacity by extraneous bombastic circumlocution."