0% found this document useful (0 votes)
226 views680 pages

H. Sagan Advanced Calculus

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
226 views680 pages

H. Sagan Advanced Calculus

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 680
Hans Sagan North Carolina State University Advanced Calculus of Real-Valued Functions of a Real Variable and Vector-Valued Functions of a Vector Variable Houghton Mifflin Company - Boston Atlanta » Dallas - Geneva, Illinois Hopewell, New Jersey « Palo Alto - London Numbers 1 1.1. Set Language—A Mathematical Shorthand This section is devoted to a discussion of certain abstract concepts, namely sets, and some manipulations with these concepts, namely the elementary set op- erations. It is not intended as an introduction to set theory. It merely serves to introduce some terminology, the language of sets, that will prove to be helpful in our studies. A command of this language is not absolutely essential for a study of analysis but is of considerable assistance in such a study, as it provides for short and concise formulations as well as for an abstract structure for argu- ments which will occur repeatedly in the study of analysis. Such arguments would be awkward and overburdened with imprecise and excessive verbiage were it not for the language of sets. We shall take a naive approach and say, in the spirit of Georg Cantor,+ that a set is a collection of well-distinguished objects of our conception or imagination. (We note that this idea of a set, when carried too far, leads to grave difficulties, as we shall point out in section 3.) The collection of coins in your pocket is a set—it may be the empty set. Definition 1.1 The empty set @ is the set which contains no elements. The collection of all matter in the universe is a set—for those who are material- istically inclined, this may well be the universal set. The universal set is the set of all objects that are under consideration in a specific investigation. For a number theorist the universal set may be the set of all integers; for an analyst, the universal set may be the set of all real numbers, or the set of all functions; for a geometer, the universal set may be the set of all points on a line, or the set of all lines in a plane, or what have you. + Georg Cantor, 1845-1918, 2 Chapter 1 Numbers In this and the next four sections, we shall discuss a variety of examples in order to illustrate the new concepts that are being introduced. To make these examples sufficiently interesting and to provide for enough variety, we shall assume, for the time being, that the reader has some intuitive notion of the real- number system, functions, graphs, coordinate systems, and other basic ideas. The concepts and operations which we are about to introduce do not, however, depend on the knowledge of, or even on the existence of these notions. In due course, we shall establish a sound foundation for the real-number system and the attendant concepts. We use capital letters A, B, C, X, Y, Z, etc., to denote sets and lower-case letters a, b, c, x, y, Z, etc., to denote elements of sets. If a is an element of A, we write acd and if a is not an element of A, a¢A. With the understanding that N denotes the set of all natural numbers (positive integers), we describe the set A that contains the elements 2, 3, 4, 5 as follows: (1.1) A={xeEN|1 1}, and if F is defined as in (1.6), c(F) =R\F ={xe R[x? >0}=R. We note that the sets 4, E, O in (1.1), (1.2), and (1.3) are contained in the set N; that the sets 3, 3 in (1.4) and (1.5) are contained in R; that E and O when com- bined yield N; and that E and O have no common elements. We shall introduce some more set language that will enable us to make these statements crisp and precise: Definition 1.3 X is a subset of Y (X is contained in Y) xXxoY means that if xe X, then xe Y. Y is a superset of X if and only if X is a subset of Y. X is a proper subset of Y (X is properly contained in Y) xey 4 Chapter 1 Numbers means that X is a subset of Y and there exists at least one element ye Y such that y ¢ X. Y is a proper superset of X if and only if X is a proper subset of Y. The sets X and Y are equal, X=yY, means that YS Yand Yc X. We have, for the sets in (1.1) to (1.6) and N and R, that AGN, EcN, OcN, NEN, JeJcR, Fed, RER, and, of course, We adopt the convention that (1.7) Bex for all sets X. (See also exercise 1.) Definition 1.4 (Set Union and Intersection) The union of XY and Y (in symbols: X u Y) is the set XU Y= {x|xeX or xe ¥}. (Note that “or” is not exclusive, i.e., x may be in X as well as in Y.) The intersection of X and Y (in symbols: X 7 Y) is the set Xo Y= {x|xeX and xe Y}. We have, for the sets in (1.1) to (1.5), that AuUcA)=N, EUO=N, EnNO=Q, Jut=s, InNF=3. We have, directly from the definition of a set complement and from definition 1.4 that, for every set X, XuUdX)=U, XadX)=o where U denotes the universal set. We call two sets X and Y disjoint if and only if Xn Y=. In order to simplify our notation, we shall employ special symbols for certain subsets of the set R of real numbers, the so-called intervals: {xe Rlaa} = [a, 0), {xeR|x , that is, ¥, © X}. Then show the converse: if y € X>, then y € X;, that is, Y, © X,. Prove: (a) e(c(A)) =A (b) XUU=U () XnU=X @) Xug=x © X0S=o Prove: (a) XEXUYYeNXuY (bt) XO YEXXAYSY Given U = N (natural numbers), A = E (even numbers), B = {1, 3, 5, 6, 8}, C= (2, 6, 10, 14}. Find: (cA) U BAC The set operations have a simple and compelling representation, as in- dicated in Fig. 1.2. Such a figure is called a Venn diagram. Draw Venn dia- grams for the following set operations: (a) (AUB)AC (b) (AN BUC () AUB (d) AU cB) AUB ANB Fig. 1.2 6 Chapter 1 Numbers 6. Prove: (a) AUB=BUA; ANB=BOA (bt) (AU B)UC=AU(BUC); (AN B)NC=AN(BNO) () An(BUC)=(ANB)U(ANC); AU(BAC) =(AUB)n(AuO) 7. Prove: A\B = An c(B) 8. Prove: (A U B)\C =(A\C)U(B\C); and (An B)\C =(A\C) 9 (B\C). 9. Prove: (a) If AN B=Q, then A\B=A; — (b) If A SB, then A\B=Q; (©) If A\B=Q, then ASB. 1.2 Generalized Unions and Intersections; DeMorgan’s Law If J denotes a given set (finite or infinite) and if with each j € J there is associated a set X;, we can generalize the operation of union and intersection as follows: (2.1) UX) = (xlxe X, for some je J} is and ’ (2.2) (VX) =flxeX, forall jet. jeu In such context, J is referred to as an index set. Note that the union is the set of all elements that are contained in at least one of the sets and that the intersection is the set of all elements that are contained in all the sets. Let ¥, = {x €R|a/2 Jj}. Find: UX; and) X;. JEN jen 3. Let 3, = {xe R|0 (a) by DeMorgan’s law, (b) by direct analysis. Prove: If 4 & B, then c(B) < c(A). Prove: A\(BU C) =(A\B) A (A\C); A\(Ba C) = (A\B) v (A\C). Generalize the result of exercise 7 to the following: For a given index set J, 4\U% = (\A\8). 4\ne = Ua\s). 9. Prove: An B= A\(A\B). Pn 1.3 The Russell Paradox* Bertrand Russell,t in his examination of the foundations of mathematics, found them to be quite shaky. The following example, due to him, has led to a thorough reexamination of these foundations. *This section is optional and may be skipped without loss of continuity. +tBertrand Russell, 1872-1970. 8 Chapter 1 Numbers One would not expect a respectable set to contain itself as an element. Accord- ingly. we shall call sets that do not contain themselves as elements respectable sets. Respectable sets are certainly objects of our imagination, if not conception We consider the set of all respectable sets and call it B. /s B a respectable set? Suppose that B is respectable, Since B is respectable and is the set of aif respect- able sets, it has to contain itself as an element. Hence, B cannot be respectable Suppose that 8 is not respectable. A set that is not respectable has to contain itself as an element (otherwise it would be respectable). As the set of all respect- able sets, being itself an element, 8 has to be respectable. A situation in which an object (of our conception or imagination) has a certain property if it does not have that property and vice versa, is called a paradox. Clearly, a set B such as we have attempted to construct here cannot exist, because its existence would lead to a paradox which is logically unaccept- able. The lesson we learn from this experience is that in the construction of sets one has to proceed with the utmost care. We have, in the construction of B, abused the concept of sets by stretching it beyond its capacity. There are some escape routes from this dilemma. One could attach safeguards to a suitably worded definition of the concept of a set, or, better still, define the concept of set not explicitly at all but rather with respect to its properties and behavior under the various set operations. The latter approach, the so-called axiomatic approach, is the one most widely followed and the one that provides the most comfort to mathematicians We wish to emphasize that we shall not encounter any difficulties in our treat- ment because the simple sets which we shall consider behave in an acceptable manner under the elementary set operations. 1.3. Exercises 1. Show that the set of all sets that can be defined with less than 20 words is not a respectable set. 2. Take a blank page and inscribe the following statement on that page: “The only statement on this page is false.” Is this a true statement? A false statement? 3. Definition. The village barber is the only man living in the village who shaves all men in the village who do not shave themselves. Does the village barber shave himself? Does he not shave himself? 4. Find a common feature of the Russell paradox, exercise 2, and exercise 3. 1.4 Cartesian Products and Functions Although we shall not deal with specific functions for some time to come, it is useful to have this concept, some of its properties, and the technical language associated with it at our disposal in our discussion of the real-number system in sections 6 to 12. The notation of a function is based on the concept of the Cartesian product: 1.4 Cartesian Products and Functions 9 Definition 4.1 The Cartesian} product X x Y of two nonempty sets X and Y is the set of all ordered pairs (x,y), where x € X, y € Y: Xx Y=((x yilve X ye ¥}. ((x, 3) is called an ordered pair whenever (x,) = (u,t) if and only if x =u, y =v. Note that, in particular, (x,") = (1.x) if and only if x = y.) a Example 7 Let 4 = {1,2,3}, B = {1,3,5,7}. Then, AxB = {1 D1. 3) (1 9), (1. 2), (2, 1), (2, 3), (2. 5), (2, 7), (3. Ds (3, 3), 3, 5), 3, D} and BxA = {(1, 1), (1,2). (1, 3), 3, 1), (3, 2),(3, 3), (5, 1), (5, 2). (5, 3), (7, 1), (7, 2), (7, 3) }- Note that A x BA Bx A. a Example 2 Let R denote the set of all real numbers. Then R x R = {(x, »)|xeR, ye R} is the set of all ordered pairs of real numbers. If one interprets (x, y) as the Cartesian coordinates of a point in the plane, then R x R may be viewed as the set of all points in a nvo-dimensional plane R?. We may consider subsets of Cartesian products, such as (4.1) f= f(xy eR x Rly =3x?,xe R} The representation of this set as a set of points in a two-dimensional plane R? is called the graph of f. (See Fig. 4.1.) i, y) ERX Riv s x2} Fig. 4.1 + Cartesius is the Latinized pen name of René Descartes, 1596-1650. 10 Chapter 1 Numbers fx ye RX Rix? ty? = 1) Fig. 4.2 The graph of (4.2) G = {(x,y)eR x R[x? +9? = 1} is given in Fig. 4.2. We have, for the sets 4, B in example 1, that A CR, BCR. Hence, A x B R by I forx#1,x#—-1, SO) = hv 0 forx = +1. (Note that the square-root operation is defined by \/a? = |a| for all ae R—the “ positive” square root.) 4 Chapter 1 Numbers (a) Verify that fis a function. (b) What is the supply set of f and what is the range of f? (c) Is fonto R? (d) Is fone-to-one on [—1,1]? 2. For the function in exercise 1, find the following sets: {fOlxe 10,1}; {fO)|xe [-1,0]}; {f@) xe [= 1,0) vO, 1]}. 3. State a sufficient condition on ag, a,,@),a3 for the function / in example 7 to be one-to-one on R. 4. Same as in exercise 3, so that the range of / consists of one element only. 5. Given ae R and f: RR by f(x) =ax. Under what condition on a is f onto R? 6. Givena function f: N > R by f(k) = a,. Such a function is called a sequence. Show: If a, Y,D¢ X, is such that {(y,x)|(x.y) ef} is a function with domain Rc Y, then f= (Oi Gy efye R} is called the inverse function of f on R, and f is called invertible on R. In this case, f-: R>X. 1.5 The Inverse Function 15 We shall show first that, for a function to be invertible on its target set Y, it is necessary and sufficient that the function be one-to-one and onto. A criterion for invertibility on a suitable subset of the target set will then follow easily. A function f which is one-to-one from D onto Y, fo“sy is called a bijective map. (A map that is surjective and injective is bijective.) We shall henceforth employ the symbol fiDeY to indicate that the function is one-to-one on D and onto Y rather than the bulky notation onto SDs y, Theorem 5.1 Let X, ¥ denote nonempty sets and let D © X,D # @. The function f:D—> ¥ is invertible on Y if and only iff: Y. In this case, the inverse function (1) ST =OD& NES} Y > X exists. Proof (a) Clearly, f~' € Y x X; since f # @, we see that f~' # . Since f is one-to- one, we have, from definition 4.4, that (x,,y), (x2,)) €f implies x, = x). Hence, (x1), Q.x2) €f7! implies x, = x2; that is, f~" is a function. Since f is onto Y, we have, for all ye Y, an xD such that (x,y) ef. Hence, for all ye Y, there is an xD such that (y,x) €f~'; that is, Y is the domain of f~}. (b) Suppose that (5.1) represents a function from Y into X. Then, for all ye Y, there is an xX such that (y,x)¢f~'; that is, (x.y) efi Hence, f is onto Y. Since f~! is a function, (y,x1), (¥.x2)€f7! implies x, =x. Hence, (x1. (2.3) € fimplies x, = x2; that is, fis one-to-one on D. By the definition of range, /: “+ R(f). Hence if f:D ——+ ¥, then f: Do K(f), and we have: Corollary to Theorem 5.1 Iff:D ——+ ¥ where De X, then fis invertible on its range R(f) and the in- verse function f~!: R({) > X exists. (See also exercise 12.) So we see that every one-to-one function is invertible on its range. The range of f becomes the domain of the inverse function f~’, the target set of f becomes the supply set of f~!, the domain of f becomes the range of f~', and the supply set of f becomes the target set of f~*. 16 Chapter 1 Numbers a Example 1 Let 1:R > R be given by I(x) = ax, a # 0. / is one-to-one on R and onto R, and hence, is invertible. The inverse function 17! :R > R is given by!” *(y) = (I/a)y. Theorem 5.2 If f:D— Y is invertible on RY, then f7?: ROD, Proof Let x € D. Then (x,f(x)) €f and therefore, (f(x), x) €f~'. Hence f~! is onto D. If (1,2), 02,4) Ef 1, then (x, ¥1), (92) Ef. Since f is a function, y; = y2. Hence f~! is one-to-one. Throughout this book we shall make liberal use of the elementary transcen- dental functions (exponential function, logarithm, and the trigonometric functions and their inverses) and their basic properties in examples and exercises. We make the necessary assumption that the reader has acquired a working knowledge of these functions in a calculus course. Our informal use of these functions does not prejudice the rigorous definitions of these functions which we shall give in section 4.52 (logarithm) and sections 14.142 (exponential function) and 14.143 (trigonometric functions and their inverses). The next example deals with the exponential function and the logarithm, for the purpose of illustrating definition 5.1 and the corollary to Theorem 5.1. 1.5 The Inverse Function W7 a Example 2 The exponential function E: RR is given by E(x) =e*. The reader knows from calculus that the exponential function assumes all positive real values. Hence, the range of E is R* where R* = (0, 0) denotes the set of positive real numbers. Each such value is assumed for one and only one real x. Hence, E is one-to-one. (See Fig. 5.1.) Since E: R> R* is onto and one-to-one, E~! exists and E~!:R* oR. The inverse function of the exponential function is the natural logarithm: E~"(y) = log y. In example 1, we have for all xeR I") and 1 1G) = aly) = a7 y =y forall yeR. Similarly, we have in example 2, E7"(E(x)) = log E(x) = loge* =x forallxeR and E(E“"(y)) =e& ' = el? =y forall yeR*. In general: Theorem 5.3 If f: D+ Y is invertible on RY, then (52) FY) =x forallxeD and f(f-'Q)) =y forall ye R. Furthermore, f~!: R + X is the only function with this property. Proof For all xe, there is a ye Y such that y=/(x) and x =f7'(y). Hence, S'@) =f 71) =x. For all ye R, there is an xe X such that x =f~'(y) and y = f(x). Hence SEO) = fC) = Suppose there is another function g: R + X with the property (5.2): (53) gf) =x forallxeD and f(g(y))=y forall ye R. We obtain from (5.2) and (5.3) that I) = 9 FF *0)) =f7'Q)_ for all ye R, 18 Chapter 1 Numbers that is, g=/f~! and we see that f~! is indeed the only function with the property (5.2). The exponential function is ssrictly increasing. Such functions are always invertible. Definition 5.2 S:D—>R, where D € R, is called strictly increasing (decreasing) if and only if x1

f(a). Theorem 5.4 Iff: D> R, where D ¢ R, is strictly increasing (decreasing), then f~!: R(f) > R exists and is strictly increasing (decreasing). Proof (We formulate the proof for strictly increasing functions only. The proof for strictly decreasing functions is analogous.) Tf (41,9), (ea, 9) Ef, then f(x,) = f(x2). If x; A x2, then xy x). Hence, f(x.) f(x2), contrary to our assumption. Hence, (1, ¥), (259) €f implies that x, = x2, that is, f is one-to-one on D. By the Corollary to Theorem 5.1, f~! exists on R(/). If yy < yz where yy, y2 € R(f) and f-"(y,) >f-"(2), then, with yy, = f(x), ¥2 =f (%2), we have x, > xz while f(x,) O, defined by f(x) =2x-1 is one-to-one and onto. Hence, N ~ O. 1.5 Exercises 1. Given f:R* > R by f(x) =x’. Show that f:R* +R* and find f~*(y). 2. Given f:R* > R by f(x) = 1/x. Show that f:R* oR* and find f~'(y). 20 Chapter 1 Numbers » Given f:R > R by f(x) = x°. Show that f: ROR and find f~'(y). 4. Given f:R—R by f(x) =x? —x. Find three subsets D,,D,,D; of R such that D, UD, U9; =R, 0,0 D,=@ for i#k, and such that f:9; Sarak R,i=1,2,3. 5. Given the sets A = {0,1,2}, B = {—5, —7, —9}. Find three different func- tions fi.fr,fy such that f,: Ao B for i = 1,2,3. 6. Let A,B denote finite sets. Prove: A ~ B if and only if A and B have the same number of elements. 7. Show that R* ~ R™ where R™ denotes the set of negative real numbers. 8. Prove: (a) A ~ A; (b) If A~ B then B~ A; (c) If A~ B, B~C, then An~C. 9. Given the sets (0, 1) and R*. Show that (0,1) ~ R*. 10. Show that (0,1) ~ R. Il. Let f:D—Y where DCX. Show: If D, <9, if fy,(x) = f(x) for all xeD, and fin, D, + Y, then fy, the restriction of f to D,, is in- vertible on {f(x)| x € D,}. (See also exercise 9, section 1.4.) 1.6 The Field of Rational Numbers Leopold Kroneckert said that God created the natural numbers and that every- thing else is the work of man. Be that as it may, we shall take the natural numbers for granted and take it from there. We denote the set of all natural numbers by N: N = {1,2,3,...}. We know that the sum and the product of natural numbers are again natural numbers: ifa,beN, thena+b,a-beN; and that addition and multiplication are commutative: a+b=b+ta, a-b=b-a; associative: (a+b) +c =a+(b + 0), (ab)c = albe); and distributive: a(b +c) = ab + ac. Subtraction, the inverse operation to addition, cannot be carried out freely within N (e.g., 5 — 7 is not a natural number). In order to free ourselves from any restrictions, we adjoin new elements to N, namely 0 and the negative integers, and thus arrive at the set of all integers Z: Z = (0,1, —1,2, —2,3, —3,...}. + Leopold Kronecker, 1823-1891. 1.6 The Field of Rational Numbers 21 We require that addition and multiplication within Z be governed by the commutative, associative, and distributive laws and, guided by experience, we impose the following rules on subtraction: 0 is the only element such that a+0=a forallaeZ and for each aé Z there is a unique element (—a) in Z such that a+(-a)=0. (We refer to 0 as the additive identity in Z and to (—a) as the additive inverse of a.) The system Z which we have obtained is closed under addition, multiplication, and subtraction; i.e., these operations can be carried out freely within Z and will always result in an element from Z. Division, the inverse operation to multiplication, cannot be carried out freely in Z. (While 6/— 2 = —3€Z, 7/2 ¢Z.) Again, we adjoin new elements to Z, namely, all the results than can be obtained from carrying out the four elemen- tary operations with elements from Z except for division by 0. This leads to the set of rational numbers Q: m o- {2 n where (m,n) = 1 means that m,n are relatively prime (have no common factor other than | and —1), We require that all the rules which we have imposed up to now govern the operations in Q. This is the case if we define addition and multiplication of fractions in the customary manner by m,/n, + m,/n, = a/b, meZ.nenc(onn)=t where ann, = b(n + ym) and (a,b) = 1, and by (m,/n,) + (m3/n3) = eld, where cnyn, = dm,m, and (c,d) = |. Then, 1 is the only element in Q for which a-l=1-a=a forallaeQ, and for all ae Q, a #0, there is a unique (I/a) € Q such that (We call 1 the multiplicative identity in Q and (1/a) the multiplicative inverse of a.) Q is closed under the four elementary operations of addition, multiplication, subtraction, and division; i.e., these operations can be carried out freely within Q (except division by 0) without ever arriving at a result that is not in Q. A collection of elements with this property is called a field: Definition 6.1 A set F of elements, denoted by a,b,c,..., is called a field if and only if two operations, called addition (+) and multiplication (+), are defined such that 1. IfabeF, thena+b,a-beF; 2. a+b=b+a,a+b=b-a for alla,beF; 22 Chapter 1. Numbers io (a+b)+ce=a+(b+0), (a+b)+c=a+(b+c) for allab,ceF; a‘(b+c)=a+bh+are, forall a,b,ceF; 5. There is a unique additive identity in F, call it 0, such that cal a+0=a forallaeF; 6. Each ae F has a unique additive inverse in F, call it (—a), such that a+(-a)= 7. There is a unique multiplicative identity in F different from the additive identity, call it 1, such that a-l=a forallaeF; 8. Each ae F,a #0, has a unique multiplicative inverse in F, call it (1/a), such that Examples of fields abound in mathematics. We have seen that Q is a field and is, as a matter of fact, the “smallest” field that contains N as a subset. Many more examples of fields are treated in exercises 2, 3, 5,6, and 7. The rules of arithmetic which we are accustomed to can all be derived from the field axioms (1) to (8). This task is to be carried out in exercises 8 through 12. Henceforth we shall adopt the standard convention of writing ab for a: 6. 1.6 Exercises 1. Show that ,/2 is not a rational number. 2. Show that the set Q(,/2) = {a + \/26 | a,b € Q} is a field. (Addition and multiplication are defined in the usual manner.) 3. Show that the set Q(/) = {x + iy | x,y €Q.i = —1) isa field if addition and multiplication are defined by: X +x2 +i + ¥2), XX — YiV2 + i y2 + X2I1)- (x1 + iy) + G2 + ty2) (x1 + Hy)Q2 + 12) 4. Show that /p, where p is a prime number, is not a rational number. 5. Show that Q(\/p) = {a + \/pb| a,b € Q} is a field if p is a given prime number. 6. Show that the set L = {0,1} is a field, if addition and multiplication are defined by: +|0 1 0 1 0|0 1 0} 0 0 Vy 0 1}0 1 1.7 Ordered Fields 23 7. Same as in exercise 6 for M = {0, — 1,1} with + o -il 1 0 -l 1 0 0 -1 1 0/0 oO 0 -I]-l 1 0 -1)0 1-1 1 0 -l 0-1 1 8. Derive the following rules from the field axioms (1) to (8): (a) (a+ bc =ac + be, a,b,c eF. (b) Ifa+x=a,aeF, then x =0. () Ifa+b=0,4,beF, then a = (d) If ab =1,a,beF,a#0, then b Ifa. () a-0= (f) (-Da=-a (g) -(@@+b)=-a-b (hy -(-I=1 OO -D-D= 9. Use the field axioms, and whatever results of exercise 8 prove useful, to show that each of the equations at+x=b and cx=d, where a, b,c,de€ F, and c # 0, has a unique solution. 10. Use the field axioms, and whatever results from exercises 8 and 9 prove useful, to show that: (a) IfaeF,a#0, then a = I/(I/a). (b) If ab =0, a,beF, then a =0 or b =0. (co) (=a)(—6) =ab for all a,b € F. ll. Define a? =aa and, inductively, a**! =aa",neN. Prove that a"*" = aa" for allaeF, m.neN. 12. Generalize the result in exercise 11 to m,n € Z with the understanding that a" = I/a",neN. 1.7 Ordered Fields It is our goal to establish a unique characterization of the set of real numbers R. The real numbers, as we know them, are certainly a field but we know with equal certainty that they are not uniquely characterized by the field axioms. As we have seen, the rationals Q are a field and so are many other collections of numbers or other abstract elements for which addition and multiplication are suitably defined. (See section 1.6, exercises 2, 3,5, 6,7.) In our search for additional properties that might be useful for a characteriza- tion of R, we let experience be our guide. We know that any two real numbers are either equal or one is larger than the other one. Fields where this holds true are called ordered fields. 24 Chapter 1 Numbers Definition 7.1 F is an ordered field if and only if 1, Fisa field. 2. There exists a nonempty subset Pc F, called a positive class, with the following properties (a) Ifa,beP, thena+b, abeP. (b) If aeF, then one of the following is true: a = 0, or ae P, or —ae P. We have as an immediate consequence of this definition that, whenever ae F and a # 0, where F denotes an ordered field, then 1) @eP. This may be seen as follows: If ae P, then, by 2(a) of Definition 7.1, a- a= a’ a is equivalent to a 0, and the positive class P stands revealed as the class of all positive numbers in F —if F is a field of numbers. We obtain, as another consequence from definition 7.2, that (1.2) a>b,c>d_impliesthat atc>b+d because a > b, c > d means, by definition 7.2, that a— b > 0, c — d>0. Hence, by 2(a) of definition 7.1, a-—b+c—dz0, or equivalently, (a+c)—(+d)>0 from which (7.2) follows readily. The other familiar rules for operating with inequalities are left for exercises 4,9, 10, 11, 12. Definition 7.3 If F is an ordered field, then for all a ¢ F, lal = a ifa>0(@eP ora=0), “(=a ifa<0(-aeP) is called the absolute value of a. (Intuitively, the reader may view the absolute value of a as the “distance” from a to 0.) We obtain immediately (7.3) ta<|al. This can be seen as follows: If a>, then [al =a; if a=0, then |a| =0; if a<0, then —a>0 and |a| — a > 0, by 2(a) of definition 7.1. Hence, a < |a|. A symmetric argument applies to the case —a < |a|. (See exercise 1.) The fundamental properties of the absolute value are given in the following theorem: Theorem 7.1 1. Jal >0 for allaeF 2. |al =0 if and only ifa=0 3. Jab] = Jal 5] 4. |atb|<|al+|d| (Triangle inequality) Proof 1: follows directly from definition 7.3. 2: If a=0, then, by definition 7.3, ja] =0. If a #0, then, by definition 7.3, la| > 0. Hence, |a] = 0 implies a = 0. 26 Chapter 1 Numbers 3: Ifa >0, b> 0, then, by 2(a) of definition 7.1, ab > 0 and |ab| = ab = |a| |b]. If a>0, b <0, then —b >0 and a(—b) > 0. Hence, |ab| = a(—4) = |al |dI. If a =0, then ab =0 and |ab| =0 =0-b =0- |d| = lal A}. 4: Ifa +6 >0, then |a + 6| =a +5 < [al + (61, by (7.2) and (7.3). Ifa +b <0, then Ja +6] =—a—6<|aj + (bl, by (7.2) and (7.3). If a+b =0, then la + b| =0 <|a| + |B], because |al + |b| > 0 for all a,b € F. The case |a — 5] < [al + [b| is left for exercise 2. Corollary to Theorem 7.1 74) Ila — [d|| < la + 4 Proof Let x = —a, y =b +a. Then, x + y = and we have, from (4) of Theorem 7.1, that |b] = lx + yl < [xl + ly] =|—al + la + 5 = lal + la +5] if we also note that |—a| = |(—1)a| = |—1] |al =a]. Hence, |b] — Ja] < Ja + 5]. Interchange of a and b yields |a| — |b| < |a + b| and (7.4) follows readily. The case ||a| — |b|| < |a — d| is left for exercise 3. We shall now show that Q is the “smallest” ordered field in the sense that every ordered field F contains Q as a subfield. By subfield of F we mean a subset of F which is itself a field under the same operations as F. Theorem 7.2 Q is the smallest ordered field i.e., every ordered field F contains Q as a subfield. Proof By field axiom (7) of definition 6.1, F has to contain at least two distinct elements, the additive identity, 0, and the multiplicative identity, 1. It follows from the field postulates that 0 and 1 have to behave under addition and multiplication as follows: +/0 1 “|O 1 o;o 1 0 1}1 2? 1 (See also exercise 8(e), section 1.6.) If | + 1 =1, then we do not have a field, since then 1 does not have an additive inverse. Hence, 1 + 1 #1. If | + 1 =0 we do have a field, as we have seen in exercise 6, section 1.6, but not an ordered field as we have seen in example 2. Hence, | + 1 #0. aoe 1.7 Exercises 27 If Fis an ordered field we have from (7.1) that | = 1? € P. Hence, if k € P, then k+leP. By letting 1 +1 =2, 2+1 =3,..., we obtain NF and, by field axiom (6), also Z&F. Consider m/n, where me Z, neN, (m,n) =1. Since neN CF, we have, from field axiom (8) that I/n € F. Since m € Z, we have, from field axiom (1), that m(I/n) =mjneF. Hence, QF. Since Q itself is an ordered field, we see that Q is the “smallest” ordered field. Note that if we had denoted the additive and multiplicative identities by other symbols, such as @ and e, and introduced other symbols for e + e, e+e+e,..., we still would have obtained Q, except in a new and awkward notation. 1.7. Exercises 1, Prove: —a< al for all ae F. 2. Prove: |a — 6] < lal + 61. 3. Prove: |la| — [dll <|a — a. 4. Let F represent an ordered field. Prove, for a,b,c,de F: (a) Ifa>b,b>c, thena>e. (b) Ifa,beF, then either a > 6 or a=6, orabandb>a, thena=b. (d) Ifa>b6,thena+e>b+e forallceF. (e) Ifa>bandc>d,thena+c>b+d. (f) Ifa>band c> 0, then ac > be. (g) Ifa>band c <0, then ac < be. (h) Ifa>0, then 1/a>0. (i) Ifa<0, then I/a <0. (j) If ab>0, then either a > 0, b > 0 ora <0,b <0. (k) Ifa? +b? =0, then a =b =0. 5. Show: If ay,a2,... .a, € F, where F is an ordered field, then lay + ay + + ay] Slay] + faz] +--+ + Lay]. 6. Show: An ordered field has no smallest positive element. 7. If F is an ordered field, show that if a < b, then a < (a+ b)/2 0, then (1 + a)">1+na,neN. ll. Show: If c>1, then c™>c" for m>n, m,néN and if 0n, mneN. 28 Chapter 1 Numbers 12. Show: If a 0, then a" < b" for nN, and vice versa. 13. An ordered field F is called Archimedean if, for any x € F, there exists an néNCF such that n> x. Prove: The ordered field Q of rationals is Archimedean. 14. Given the collection QE) = {£0 GDS 0) = 0X + XE + Fy, GR) = Bo x" + yx"! fe + Bq, where m,meN, a,b €Q, and (bo, by, ... bm) # (0,0, ... 5 0)3 f(x), g(x) have no linear factor in common}, (a) Show that Q(x) is a field. (b) Show that Qe Q(x). (c) Show that Q(x) is an ordered field with the positive class containing all elements f(x)/g(x) for which ag by > 0 and only those elements. (a) Show that Q(x) is not Archimedean. 1.8 The Real Numbers as a Complete Ordered Field When we plunged from Z into Q we adjoined, in effect, the solutions x of all linear equations ax = 6, a,b€ Z, a # 0, to Z—except the ones that were already in Z. This might mislead us to expect that the real numbers can be obtained by a similar process, namely, by adjoining to Q the solutions of all possible algebraic equations dy x" + a,x""} + ++ +4,_-,x +4, =0, where a,eZ (or a,€Q, for that matter—see exercise 1), and ne N, and where not all of the coefficients Go, @;, ... , G1 Vanish. (See also exercise 2, section 1.6.) Such a process does not yield enough and, at the same time, yields too much. We shalll see later (exercise 10, section 1.13) that this process leads to a subset of the set of complex numbers. Since the rational numbers already form an ordered field, the field axioms and the order axioms do not suffice to characterize the real numbers. It appears that some important, but seemingly elusive, property of the real numbers is not captured in these axioms. A study of the set A= {1, 1.01, 1.01001, 1.010010001, 1.01001000100001, 1.01001000100001000001, ...} of rational numbers will help us to flush out this elusive property. Clearly, the rational number u, = 2 is an upper bound of A, that is, a < u, for all ae A. Also, u, = 1.02 € Q is an upper bound of A, a 42. Likewise, u, = 1.01002 € Q has the property that a 2 >u 3. We continue in this manner and obtain rational numbers ug = 1.010010002, us = 1.01001000100002, ... such that for all j¢ N, a uz >u3>uU,>us>---. The question arises: Is there a smallest rational number wu such that a

You might also like