Fundamentals of University Mathematics

Chapter 1

Preliminaries

All of the mathematics in this book is based on real numbers. Historically, mathematics began with the set of natural numbers (or positive integers) on which we have the operations of addition and multiplication. Much later, this set was extended by adding zero and the negative integers, thus giving the set of integers. In this set, subtraction is always possible, but division is only possible in certain cases. For example, we can divide 20 by 5, but not by 6. Mathematicians were thus led to introduce the rational numbers (or fractions), so division was always possible (except by zero). Eventually it was realised that yet more ‘numbers’ were needed to do even quite simple calculations. For example, the diagonal of a unit square has length , but cannot be written as a fraction (see Section 1.8). To resolve this problem, mathematicians introduced irrational numbers, thus arriving finally at the set of real numbers.

In this first chapter, we recall results about the set of real numbers and some subsets which are important in their own right. The reader should be familiar with most of the material in Sections 1.1 to 1.6, but should read these sections carefully since they introduce notation which will be used in the remainder of the book. The final sections may well be new. Section 1.7 looks in detail at the idea of divisibility in the set of integers. Section 1.8 considers the numbers which cannot be written as fractions—the irrational numbers. These sections also introduce the idea of proof by contradiction, a technique much used throughout mathematics.

1.1 Number Systems

Informally, a number system consists of a set of numbers which is closed under operations of addition and multiplication, i. e. if a ∈ and b ∈ then a + b ∈ and ab ∈ . The number systems introduced in this section are all subsets of the real numbers and should be familiar to the reader. Indeed, it is assumed that the reader has a working knowledge of these systems. They are:

, the set of positive integers or natural numbers

, the set of integers

, the set of rational numbers or fractions

, the set of real numbers, represented by the points on an infinite straight line. We shall further assume that the reader is familiar with the language of arithmetic, including such terms as reciprocal and quotient.

Example 1.1.1 Assuming that and are number systems, prove that is a number system. Show, also, that if q ∈ with q ≠ 0 then 1/q ∈ .

Solution Let p, q ∈ . Then p = h/m and q = k/n where h, k ∈ and m, n ∈ . So

Since and axe closed under addition and multiplication and  ⊆  we have nh + mk ∈ , hk ∈ and mn ∈ . Hence p + q ∈ and pq ∈ as required.

Now let q ∈ with q ≠ 0. Then q = k/n where k ∈ , k ≠ 0 and n ∈ . Hence 1/q = n/k = (–n)/(–k) ∈ since n and – n are in and k or –k is in .

Problem 1.1.2 Let = {1, 3, 5, …} and = {2, 4, 6, …}. Is either or a number system? Can you find a number system such that ⊂ ⊂ ?

The reader is probably aware that

As number systems, the four sets differ in a more fundamental way—in terms of the solutions of equations. The equation x + 5 = 3, which is expressed in terms of the number system , has no solution in . It does have a solution, namely x = –2, in the larger system . Similarly, the equation 2x – 3 = 0, which is expressed in terms of the system has no solution in but does have a solution, x = 3/2, in . We shall see later that x² = 2, an equation in , has no solution in . It has a solution in . Even in there are simple equations with no solution. For example, x² = –1. There is a yet larger number system , the set of complex numbers, in which all such equations have solutions. We shall deal with in Chapter 6.

In addition to the arithmetic in the number systems , , and , there is also the idea of order. For example, 1 < 2 and > 0. This is not a feature of every number system. There is, for example, no sensible meaning of order for the number system .

A real number x can be positive, written x > 0, negative, written x < 0, or zero. We call x non-negative and write x ≥ 0 if x > 0 or x = 0, and we call x non-positive and write x ≤ 0 if x < 0 or x = 0.

When we write x > 0, x ≤ 0, etc. without specifying to which number system x belongs, it will be assumed that x ∈ .

We shall see more of order in Section 1.6.

As has been mentioned, each real number can be regarded as a point on an infinite straight line called the real line. Figure 1.1.1 illustrates the way in which each of the isolated numbers –2, 1 and 3 is marked on the line by •. The arrowhead indicates the positive side of 0 and the positive direction on the line.

Figure 1.1.1

When we wish to indicate more complicated sets of points we use a thick line to indicate a continuous set of points and o for a point which is specifically excluded. Thus, in Figure 1.1.2, we have represented the set S = {x ∈ : – 2 ≤ x < 0} ∪ {1, 3}.

Figure 1.1.2

1.2 Intervals

Definitions 1.2.1 The connected subsets of , those represented on the real line by a continuous set of points, are called intervals. In the different types listed below, we let a, b ∈ with a < b. Their real line representations are shown in Figure 1.2.1. They are:

Figure 1.2.1

We shall refer to the empty set and singletons as trivial intervals and all other types as non-trivial.

Remark The symbols ∞ and –∞ which are read as ‘infinity’ and ‘minus infinity’, respectively, are not real numbers and are not included in any interval. Sometimes + ∞, read as ‘plus infinity’, is used instead of ∞.

Example 1.2.2 Let A = (–2, 3), B = (–∞, –5] [–2, 5) and C = (4, ∞). Express A’ (B – C) as a union of intervals.

Solution Figure 1.2.2 shows the steps taken to calculate this expression. It shows that

Figure 1.2.2

Frequent use is made, in mathematics, of sets of real numbers ‘near’ a given number. We make this idea precise as follows.

Definitions 1.2.3 Let c ∈ . A neighbourhood of c is any set of the form (c – δ, c + δ) where δ > 0, i.e. any bounded open interval centred on c. A punctured neighbourhood of c is any set of the form N – {c} where N is a neighbourhood of c, i.e. a set of the form (c – δ, c) (c, c + δ) where δ > 0. See Figure 1.2.3.

Figure 1.2.3

A left neighbourhood of c is any set of the form (c – δ, c] where δ > 0 and a punctured left neighbourhood of c is any set of the form L –{c} where L is a left neighbourhood of c. A right neighbourhood R and a punctured right neighbourhood R –{c} are defined in the obvious way. See Figure 1.2.4.

Figure 1.2.4

Notes 1.2.4 (1) If M and N are both neighbourhoods [both punctured neighbourhoods, etc.] of c ∈ then M N is also a neighbourhood [punctured neighbourhood, etc] of c.

(2) If c ∈ (a, b) then (a, b) contains a neighbourhood of c. See Figure 1.2.5. More generally, if c ∈ (a, b) ⊆ A then A contains a neighbourhood of c.

Figure 1.2.5

(3) Let c ∈ and let δ > 0. Let N = (c – δ, c + δ) and I = (– δ, δ). Then N and I are neighbourhoods of c and 0, respectively, and, if x = c + h,

See Figure 1.2.6.

Figure 1.2.6

1.3 The Plane

The elements of the set

are often regarded as the points on a plane. We call it the x, y-plane or simply the plane. Figure 1.3.1 shows how subsets of ² may be represented. The subsets shown are

Figure 1.3.1

Hazard When x, y ∈ with x < y, the notation (x, y) can mean both an open interval and a point in the plane. The context usually makes the meaning clear.

1.4 Modulus

Definition 1.4.1 For x ∈ the modulus or absolute value of x, denoted by |x|, is defined by

For example,

The curves y = x and y = |x| are shown in Figure 1.4.1.

Figure 1.4.1

More generally, Figure 1.4.2 shows the relationship between curves y = f(x) and y = |f(x)|. The two curves coincide where f(x) ≥ 0 and they are mirror images of each other in the x-axis where f(x) < 0. For more on this see Section 2.2.

Figure 1.4.2

For x ∈ , the number |x| can be thought of as the distance on the real line from 0 to x. So x is the point on the line reached by starting at 0 and taking a step of length |x|, in the positive direction if x > 0, or in the negative direction if x < 0.

More generally, we can think of x or ‘adding’ x as a step of length |x| in the appropriate direction starting at any point on the line. Then, for x, y ∈ , x + y is represented by a step x followed by a step y. Since x = y + (x – y) it follows that |x – y| is the length of the step from x to y, i.e. |x – y| is the distance from x to y on the real line.

These ideas are illustrated in Figure 1.4.3.

Figure 1.4.3

Example 1.4.2 Express each of the sets

as an interval.

Solution We use the geometrical interpretation of modulus. Since |x + l| = |x –(– 1) |, the set A consists of those x in whose distance from –1 is less than or equal to 3, i.e. A = [–4, 2]. The set B consists of those x in nearer to 8 than to 0, i.e. B = (4, ∞). See Figure 1.4.4.

Figure 1.4.4

Properties 1.4.3 The following properties of modulus follow directly from the definition. Let x, y ∈ .

(1) |x| ≥ 0 and |x| = 0 ⇔ x = 0.

(2) |–x| = |x|.

(3) |x|² = x².

(4) |x – y| = 0 ⇔ x = y.

Lemma 1.4.4 Let x ∈ and a ≥ 0. Then

Proof Using the definition of modulus, we have

the last equivalence being immediate.

Theorem 1.4.5 (The Triangle Inequality) Let n ∈ and suppose that x1, x2,…, xn ∈ . Then

Proof For i = 1, 2, …, n, it is obvious that |xi| ≤ |xi| so, by Lemma 1.4.4, – |xi| ≤ xi ≤ |xi|. Then adding gives

This is justified in Theorem 1.6.5 (2).

Applying Lemma 1.4.4 again gives

Corollary 1.4.6 Let x,y ∈ . Then

Proof (1) This is just a special case of Theorem 1.4.5 together with the observation that |x – y| = |x + (–y)| ≤ |x| + | – y| = |x| + |y|.

(2) We have

so that

Thus

and hence, by Lemma 1.4.4, ||x| – |y|| ≤ |x – y|. Replacing y with –y completes the proof.

Theorem 1.4.7 Let n ∈ and x1, x2,…, xn ∈ . Then

Proof We shall prove the special case that, for x, y ∈ , |xy| = |x||y|. The general case is dealt with in Example 1.5.12.

(1) Case 1. Let x, y ≥ 0. Then xy ≥ 0 and |xy| = xy = |x||y|.

(2) Case 2. Let x ≥ 0 and y < 0. Then xy ≤ 0 and |xy| = –xy = x(–y) = |x||y|.

(3) Case 3. Let x < 0 and y ≥ 0. Then xy ≤ 0 and |xy| = – xy = (– x)y = |x||y|. [Since ab = ba for a, b ∈ this is essentially the same as Case 2.]

(4) Case 4. Let x, y < 0. Then xy > 0 and |xy| = xy = (– x)(– y) = |x||y|.

Thus, in all four cases, |xy| = |x||y|.

Remark The above proof of Theorem 1.4.7 is an example of proof by cases. It is important to ensure that the cases considered cover all possibilities.

Corollary 1.4.8 Let x, y ∈ with y ≠ 0. Then .

Proof Using Theorem 1.4.7, |x/y||y| = |(x/y)y| = |x| and the result follows immediately.

Definitions 1.4.9 For x, y ∈ , we define

More general definitions of min and max are given in Section 2.4.

Example 1.4.10 Show that, for x, y ∈ ,

(1) min{x, y} = [(x + y) – |x – y|],

(2) max{x, y} = [(x + y) + |x – y|].

Deduce that, for x ∈ , |x| = max{x, –x}.

Solution We shall leave (1) to the reader and establish (2).

Case 1. Let x ≥ y. Then x – y ≥ 0 and

Case 2. Let x < y. Then x – y < 0 and

Thus, in both cases equation (2) holds.

Now let x ∈ . Then, using (2),

as required.

1.5 Rational Powers

Our aim in this section is to define, as far as is possible, the rational powers of real numbers so that they obey certain index laws. More general real powers must wait until Chapter 15. For q ∈ and x ∈ , the qth power of x, when it exists, is denoted by xq. We proceed in four stages.

Stage 1. For n ∈ and x in any number system, we define

For example,

The general shape of the curve y = xn is shown in Figure 1.5.1.

Figure 1.5.1

Properties 1.5.1 These properties of positive integer powers follow directly from the definition. Let n ∈ and x ∈ .

(1) x = 0 ⇔ xn = 0.

(2) x > 0 ⇒ xn > 0.

(3) x < 0 ⇒

(4) (–x)n ⇒

Lemma 1.5.2 Let x, y and let n ∈ . Then

and, if n is odd,

In particular,

Proof When the right-hand sides of the equations are multiplied out, the terms cancel in pairs except for xn and ± yn.

Lemma 1.5.3 Let xn = yn where n ∈ and either x, y ≥ 0 or x, y ≤ 0. Then x = y.

Proof If x = 0, Property 1.5.1 (1) gives xn = 0, so that yn = 0, and then y = 0, so that x = y(= 0). If y = 0 the proof is similar.

Let A = xn−1 + xn−2y + xn−3y² + · · · + x²yn−3 + xyn−2 + yn−1. Then, by Lemma 1.5.2,

If x, y > 0 then Property 1.5.1 (2) leads to A > 0 so we must have x – y = 0, i.e. x = y. If x, y < 0 then –x, –y > 0 and Property 1.5.1 (4) gives (–x)n = (–y)n so that – x = – y, i.e. x = y.

Stage 2. For n ∈ and x ∈ – {0}, we define

For example,

We have now defined all sensible integer powers of real numbers.

Lemma 1.5.4 (Index Laws for Integer Powers) Let h, k ∈ and x, y ∈ . Then, whenever both sides of the equation are defined,

Proof This is left to the reader. For each index law, various cases must be considered. For example, let h ∈ and –k = n ∈ with h < n. Then, for x ≠ 0,

which proves (1) in this case.

Stage 3. For n ∈ and x ≥ 0 we define

and for n ∈ , n odd, and x < 0 we define

When n = 1, these definitions agree with the Stage 1 definition. When n ≥ 2, the existence of u and v is an assumption but their uniqueness follows from Lemma 1.5.3.

The number x¹/n is called the nth root of x. Traditionally, x¹/² is called the square root and x¹/³ the cube root. The nth root of x is also written or, in the case n = 2, . For example,

Hazard Remember that x¹/² is always non-negative. Thus, for example, 4¹/² and mean 2; they never mean –2.

Lemma 1.5.5 Let n ∈ and x ∈ such that x¹/n is defined. Then

Proof Since, by definition, x¹/n = w where wn = x, it follows immediately that (x¹/n)n = wn = x. This proves the first equation.

When x ≥ 0, xn ≥ 0 and (xn)¹/n = u where u ≥ 0 and un = xn. Then Lemma 1.5.3 gives u = x. When x < 0 and x¹/n exists, n must be odd so that xn < 0 and (xn)¹/n = v where v < 0 and vn = xn. Then Lemma 1.5.3 gives v = x. This proves the second equation.

Example 1.5.6 Solve the equation

(1.5.1)

Solution Squaring both sides of equation (1.5.1) gives

i.e.

i.e.

(1.5.2)

Then squaring both sides of equation (1.5.2) gives

i.e.

i.e.

i.e.

Substituting these values of x in equation (1.5.1) we find that only x = 5 satisfies the original equation and is therefore its only solution.

Hazard The check by substitution at the end of the last example was essential. By squaring both sides of an equation we create a new equation which may have more solutions than the original. For example, x = – 2 implies that x² = 4 which has solutions x = – 2 and x = 2. In Example 1.5.6 it was squaring equation (1.5.2) which caused the problem.

Lemma 1.5.7 Let h, k ∈ and m, n ∈ with h/m = k/n and let x > 0. Then

Proof Let a = (x¹/m)h and b = (x¹/n)k. Then, using Lemmas 1.5.4 and 1.5.5,

Similarly, bmn = xmk. Since hn = mk, it follows that amn = bmn. Also, a, b > 0. Hence, Lemma 1.5.3 applies to give a = b.

Stage 4. For k ∈ , n ∈ and x > 0 we define

When k = l, k = n or k = –n, this definition agrees with the definitions of Stages 1 and 2. Lemma 1.5.7 guarantees that this definition does not depend on the fractional representation of k/n when x > 0.

The definition also applies when x = 0 provided k ≠ 0 and when x < 0 provided k and n have no common factor and n is odd. For example,

In the definition we could stipulate throughout that k and n have no common factors, thereby avoiding the need for Lemma 1.5.7. However, it is convenient to have the less restrictive definition for x > 0.

We have now defined all sensible rational powers of real numbers.

Note 1.5.8 Let q ∈ with q = k/n where k ∈ , n ∈ and k and n have no common factor. Then, in Stages 1 to 4, we have defined xq for x ∈ D(q) where

Theorem 1.5.9 (Index Laws for Rational Powers) Let p, q ∈ and x, y ∈ . Then, whenever xp and xq are defined,

and whenever xp and yp are defined,

In particular, (1), (2) and (3) hold for p, q ∈ and x, y > 0.

Proof (For the case x, y > 0) Let p = h/m and q = k/n where p, q ∈ and m, n ∈ .

(1) We have

(2) Since mn ∈ and x > 0, there exists a > 0 such that x = amn. Then x¹/,mn = a and we have

(3) We have

Hazard Note that index law (2) in Theorem 1.5.9 requires that both xp and xq are defined. Otherwise the index law may not hold. For example, since (–5)³/² is not defined, one cannot deduce that ((–5)²)³/² = (–5)³ = –125. In fact, ((–5)²)³/² = 25³/² = 125.

Corollary 1.5.10 Let p, q ∈ and x, y ∈ . Then, whenever both sides of the equation are defined,

Proof We have

Remark We shall assume the ‘obvious’ extensions to the index laws. For example,

for suitable p, q, r, … ∈ and x, y, z, … ∈ .

Example 1.5.11 Simplify

where means .

Hazard Whenever is written without any brackets it always means not (ab)c. If the latter meaning was intended it would be written as abc.

Solution We have

Example 1.5.12 Show that, for x ∈ , = |x|. Deduce that, for n ∈ and x1, x2, …, xn ∈ ,

Solution Since |x| ≥ 0 and |x|² = x² it follows that = |x|. Then we have

1.6 Inequalities

We have already made considerable use of the order relation for real numbers based largely on ‘common sense’. Here we put things on a sound footing.

Notation 1.6.1 For x ∈ , we write x > 0 to mean x is positive and x < 0 to mean x is negative.

All inequalities for real numbers stem from four rules.

Rule 1. For each x ∈ exactly one of the following holds:

x > 0, x = 0, x < 0

Rule 2. Let x ∈ . Then x > 0 ⇔ – x < 0.

Rule 3. Let x, y ∈ . Then x, y > 0 ⇒ x + y > 0.

Rule 4. Let x, y ∈ . Then x, y > 0 ⇒ xy > 0.

Theorem 1.6.2 Let x ∈ . Then

x² ≥ 0  and  x ≠ 0 ⇒ x² > 0.

Proof First note that 0² = 0. Next, by Rule 1, if x ≠ 0 then x > 0 or x < 0. When x > 0, Rule 4 gives x² > 0. When x < 0, Rule 2 gives – x > 0 and then Rule 4 gives x² = (–x)(–x) > 0.

Notation 1.6.3 For x, y ∈ , we write x > y or y < x to mean x – y > 0, and we write x ≥ y or y ≤ x to mean x – y > 0 or x = y.

Example 1.6.4 Prove that, for all a, b ∈ , a² + b² ≥ 2ab.

Solution Let a, b ∈ . Using Theorem 1.6.2 we have

and the result follows immediately.

From the four rules we can deduce further principles for manipulating inequalities.

Theorem 1.6. 5 Let a, b, x, y ∈ and q ∈ . Then

(1) x > y ⇒ a + x > a + y,

(2) a > b and x > y ⇒ a + x > b + y,

(3) a > 0 and x > y ⇒ ax > ay,

(4) a < 0 and x > y ⇒ ax < ay,

(5) a > b > 0, and x > y > 0 ⇒ ax > by,

(6) x > 0 ⇒ 1/x > 0,

(7) x > y > 0 ⇒ 1/x < 1/y,

(8) q ≠ 0 and x > 0 ⇒ xq > 0,

(9) q > 0 and x > y > 0 ⇒ xq > yq,

(10) q < 0 and x > y > 0 ⇒ xq < yq,

Proof The reader is invited to decide which rule is being invoked at each stage.

(1) Let x > y. Then x – y > 0 so that (a + x) – (a + y) > 0, i.e.