Defn.The real numbers are defined to be a set IR with two binary operations (+,·) which satisfy the following properties: Given any a,b,c in IR
Notation:
|
it follows from Homework Problem 2.3 that a: = (-1)·a is an additive inverse for a. But additive inverses are unique (from your Homework Problem 2.2), so the conclusion of part (i) follows.To prove part (ii), we assume to the contrary, i.e. that 1 does not belong IP. By the definition, 1 ¹ 0, so (-1) Î IP and therefore 0 < -1. But (-1)·(-1) = -(-1) = 1 by part (i) and the HW Problem that additive inverses are unique. This shows that 1 Î IP by property (b) of the positive cone and the assumption that (-1) Î IP. Contradiction, by the tricotomy property (c).
For part (iii), observe that 0 < a means a Î IP. Since additive inverses are unique, then -(-a) = a, and so (0-(-a)) = -(-a) Î IP. This is equivalent to the statement (-a) < 0. [¯]
Defn. A real number a is said to be an upper bound for A Í IR if
|
Lemma. Suppose that A is a nonempty subset of IR, with least upper bound M, then for every e > 0, there exists a Î A such that
|
Pf. Since 0 < e, then M-e < M. This shows that M-e cannot be an upper bound for A. Hence there is a member of A, call it a, so that M-e < a. [¯]
Theorem. (Archimedean
Property) Suppose a,b are positive real numbers, then
there exists n Î IN such that b
< n·a.
Pf. Suppose to the contrary that na < b for all n Î IN, then it follows that a: = b/a is an upper bound for the natural numbers. Let M be the least upper bound. By the lemma, 1/2 > 0, so there exists a natural number N so that M-1/2 < N. But then, M < N+1/2 < N+1, which shows that M is not an upper bound for IN. Contradiction. [¯]
Corollary.
The natural numbers are not bounded.
Corollary. Given any e > 0, there exists n Î IN such that 1/n < e.
Notation: Next we define intervals of real numbers.
Pf. Define the length of I by L: = b-a. By the previous corollary, there exists no Î IN such that 0 < 1/no <L. Let A : = {k | k an integer and k/no < a}. A is nonempty, since the negative integers are not bounded from below. Let ko belong to A. Set B : = {k| k an integer and k ³ ko }ÇA. Also, A is bounded from above by a·no, which shows that B is in fact a finite set of integers. Let K be the largest member of B and therefore of A, then K+1 does not belong to A. Let r : = (K+1)/no, then
a < (K+1)/no < K/no + L £ a + (b-a)which shows that the rational r Î (a,b) Í I. [¯]
Corollary.
Each interval with nonzero length contains an infinite number of rationals.
Defn. A real number is said to be irrational if it is not rational.
Corollary. Each interval with nonzero length contains an uncountably infinite number of irrationals.
We establish a few other facts about irrational numbers and also prove directly that each interval of positive length contains an infinite number of irrationals.
Lemma.
The product of a nonzero rational with an irrational is irrational.
Pf. Suppose that q1 ·a = q2, where q1,q2 are rational and a is irrational. Since q1 ¹ 0, then a = q2/q1 and it follows that a is rational. Contradiction. [¯]Lemma. If m is an odd integer, then m2 is odd.
Pf. If m is odd, then there exists an integer k such that m = 2k+1. In this case m2 = 2(2k2 +2k)+1. [¯]
Lemma.Ö2
is irrational.
Pf. Suppose that Ö2 = m/n where m,n are integers with n > 0. We may assume that the rational is in lowest terms (i.e. m and n have no common factors). Squaring the equation and multiplying by n2, we obtain that m2 = 2 n2. This shows that m2 is even. By the lemma m must be even and equivalently that it contains 2 as a factor. This shows 4 k2 = 2 n2 for some integer k. Consequently, n is even and 2 appears as one of its factors. Contradiction, since m/n was supposed to be in lowest terms. [¯]Pf. (Another proof of the `density' of the irrationals) Let a,b be the endpoints of the interval I. Consider the interval (a/ Ö2, b/ Ö2). It has length (b-a)/Ö2 > 0, and so contains a nonzero rational number q. It follows that qÖ2 is between a and b and hence belongs to I. [¯]
Defn.
The absolute value of a real
number a is defined by
|
Lemma. The absolute value function has the following properties:
Pf. To prove property 1, `case it out.' Either
a ³ 0 or a < 0.
In the first case |a|
= a ³ 0. In the second case, |a|
= -a > 0, since a < 0.
Property 2 follows similarly. Properties 3 and 4 are left for the student.
To prove property 5, recall that either c : = a+b satisfies c ³ 0 or c < 0. In the first case, |c| = c = a+b £ |a| + |b|. In the second case, |c| = -c = -a + (-b) £ |a| + |b|. This establishes property 5.
To prove property 6, use the fact that |b| = |(b-a) + a| £|b-a| + |a| and subtract |a| from each side. This shows that |b| - |a| £|b-a|. By symmetry in a and b, one proves that |a| - |b| £|a-b| = |b-a|. [¯]