This action might not be possible to undo. Are you sure you want to continue?

# Little Mathematics Library

I.M.YAGLOM

AN UNUSUAL

ALGEBRA

Mir Publishers. Moscow

ПОПУЛЯРНЫЕ ЛЕКЦИИ ПО МАТЕМАТИКЕ

II. М.

Яглом

НЕОБЫКНОВЕННАЯ

АЛГЕБРА

И З Д А Т Е Л Ь С Т В О « Н А У К А » МОСКВА

L I T T L E MATHEMATICS LIBRARY

I.M.Ya^lom

AN UNUSUAL

ALGEBRA

Translated

from

the

Russian

by

1. G .

MIR

Volosova

P U B L I S H E R S

\1< ) S ( '.( ) \ V

F i r s t p u b l i s h e d 1978 Second printing 1984 Ha amAuùcitoM naut. Mir P u b l i s h e r s .e © E n g l i s h translation. 1978 .

in the second reading of the book it is advisable to study Sec. The book contains some optional material (in particular. 7 and Appendix which are starred in the table of contents) that can be omitted in the first reading of the book. — Tr. 1 ) *) The present translation incorporates suggestions m a d e b y the author. The corresponding parts of the text of the book are marked by one star at the beginning and by two stars at the end. The bibliography given at the end of the book lists some books which can be of use to the readers who want to study the theory of Boolean algebras more thoroughly. However. I. At the end of the book are placed answers and hints to some of the problems. 7 since it contains some material important for practical applications of the theory of Boolean algebras. The reader is advised to solve most of the problems. G. The distinction between the material of the lecture and that of the book is that the latter includes exercises at the end of each section (the most difficult problems in the exercises are marked by an asterisk).Preface The present book is based on the lecture given by the author to senior pupils in Moscow on the 20th of April of 1966. Kizner for the thoroughness and initiative in editing the book. . because only after the problems have been solved can the reader be sure that he understands the subject matter of the book. Gindikin for valuable advice and to F. if not all. Sec. The author is grateful to S.

"Laws of Thought". Further E x a m p l e s of A p p l i c a t i o n of R u l e s for D e d u c t i o n . Principle of D u a l i t y . . B o o l e a n E q u a l i t i e s and I n e q u a l i t i e s 4. D e f i n i t i o n of a B o o l e a n Algebra Answers and Hints 7 23 37 54 63 B9 89 100 117 120 Bibliography 125 Name Index 126 Subject Index 127 The starred i t e m s indicate those sections w h i c h m a y be o m i t t e d in the first reading of the book. P r o p o s i t i o n a l Algebra 5. Algebra of N u m b e r s and Algebra of S e t s 2. Further Properties of B o o l e a n Algebras. N o r m e d B o o l e a n Algebras A p p e n d i x * . S e t s and Propositions. R u l e s for D e d u c t i o n 6. P r o p o s i t i o n s and S w i t c h i n g Circuits 8. Implication 7* 1 ). B o o l e a n Algebra 3.Contents Preface 5 1.

Algebra of Numbers and Algebra of Sets When studying arithmetic and algebra at school pupils deal with numbers of various nature. in this case we cannot say that the owner of the books handles them well!) The moment we get used to the fact that a fractional number of objects in a collection can make sense we pass to negative numbers. such n u m b e r s as 1 or — 3 ' 2 or Y 2 are termed real numbers). irrational numbers etc. there cannot of course be 33 1 / 2 boys in a class or 3 x / 4 plates on the table. When the pupils get used to a new class of numbers they are no longer puzzled by them. and there can even be 6 1 / 2 books on a book shelf. but at the same time. (This would be quite unnatural!) But a thermometer can read —5° and it even makes sense to say that a person has —50 copecks (the latter situation may worry the person but this is of no importance to mathematics!). there cannot be —3 books on a book shelf. for instance. However. By the way. if the reader is not yet familiar with irrational and imaginary numbers.1. Of course. As to the fractions. how many apples there are in a basket or how many pages there are in a book or how many boys there are in a class. Senior pupils study still more "frightening" numbers: first the so-called irrational numbers such as 1^2 and then the imaginary numbers such as 1 + 2i1) (the terms "irrational" and "imaginary" clearly indicate how strange these numbers seemed to the people until they got used to them). but yet at every stage of the extension of the notion of a number they lose some of their illusions. there can be 4V2 apples on the table. A whole number gives information on how many objects there are in a given collection. At the beginning the pupils study whole numbers whose understanding causes no difficulties because most pupils are to some extent familiar with these numbers before they start going to school. this will not prevent x ) In modern m a t h e m a t i c s the n u m b e r s of the t y p e of 1 + 2i are called complex numbers while b y imaginary (or pure imaginary) numbers are m e a n t such numbers as 2i or —~\f 2i (in contrast to t h e m . (Of course. a film can last for l 3 / 4 hours. in the further course of the study of mathematics the pupils come across new and still new "numbers" (such as fractional numbers.). 7 .

when we add together two positive integers a and b we find. For instance. w e c a n n o t s u b t r a c t the n u m b e r 5 f r o m the n u m b e r 3 and if we k n o w o n l y t h e w h o l e n u m b e r s we c a n n o t d i v i d e the n u m b e r 7 b y the n u m b e r 4. If there are 35 pupils in a class and 39 in another class then there are 35 + 39 — 74 pupils in both classes (also see Fig. Similarly. Rational and Irrational. It can easily be noticed that the main common feature of all these kinds of numbers is that all the numbers can be added together and multiplied by each other 2 ). R a n d o m H o u s e . for i n s t a n c e . Numbers'. 8 . 1 û ô û j ô o i o û i ô ô Û Û I O Ù I Û Û I Û Ù b apples b apples b apples b apples a groups of apples Fig. N e w York._ a*b apples Ù Û ! 0 ÔIO 0 ! 0 Ù ÛÛOlOOOO o o o i û û û û û û ô a apples ' û ù û b apples Fig. this similarity between the various kinds of numbers is conditional: the matter is that although we can perform addition and multiplication of all kinds of numbers these operations themselves have a different sense in different cases. 1). 2 ) B u t n o t s u b t r a c t e d or d i v i d e d : i f . N i v e n .ab apples j. x ) A n e l e m e n t a r y r e p r e s e n t a t i o n of v a r i o u s n u m b e r s c a n be f o u n d in t h e b o o k : I. when we multiply two positive integers a and b we find the number of objects in a union of a collections each of which contains b obfects. the number of objects in the union of two collections the first of which contains a objects and the second b objects. However. 1961. So it is natural to ask what is the common feature of all these kinds of numbers which allows us to apply the term "number" to all of them. w e are f a m i l i a r o n l y w i t h p o s i t i v e n u m b e r s . 2 him from reading this book 1 ). Although the concepts of irra tional and imaginary numbers have little in common with the primary idea of a whole number as a quantitative characteristic of a collection of objects we nevertheless speak of them as "numbers".

2). For instance. there are 3 classes in each of which there are 36 pupils then there are 3-36 = 108 pupils in all these classes (also see Fig.If. here we have gone too far: it turns out that there is in fact a great similarity between the operation of addition of the whole numbers and the operation of addition of the fractions. c and d are whole numbers). for instance. It is however evident that this interpretation of addition and multiplication applies neither to the operations on fractions nor to the operations on negative numbers. for the numbers of any nature we always have the identities the commutative law tor addition of numbers ') Wo do not discuss in detail irrational and imaginary n u m b e r s here and o n l y note t h a t the complex numbers are added together and m u l t i p l i e d b y each other according to the rules (a + bi) + (c + (a + bi) (c + di) = (a + c) + (6 + d)i and di) = (ac — bd) -f (ad + 6c)i These rules m a y seem strajige to the reader who is not yet familiar with c o m p l e x numbers but they are m u c h simpler than the d e f i n i t i o n s of the sum and of the product of irrational numbers. for instance. b. For instance. We also know that. ). However. the sum and the product of rational numbers (fractions) are defined by the following rules: a . y . Thus. for the signed numbers there exists the rule (—«)•(—6) = ab 1 etc. More precisely. the definitions of these operations are different but the general properties of the operations are completely similar. we can draw the following conclusion: the term "number" is applied to numbers of different kinds because they can be added together and multiplied by each other but the operations of addition and multiplication themselves are completely different for different kinds of numbers. c ad-\-bc bd and a c ~b'~d ac ~bd (here a.

and ab = ba the commutative law for multiplication of numbers and also the identities (a + b)+c = a + (br>rc) the associative law for addition of numbers and (ab) c = a (be) the associative law for multiplication of numbers Also. in all the cases there exist two "special" numbers 0 and 1 such that the addition of the first of them to any number and the multiplication of any number by the second one do not change the original number: for any number a we have a + 0 = a and a -1 = a What has been said accounts for the point of view of modern mathematics according to which the aim of algebra is to study some (different) systems of numbers (and other objects) for which the operations of addition and multiplication are defined so that the above laws and some other laws which will be stated later are fulfilled. It is due to this analogy that so many people often confuse the notion of the (additive) inverse of a number a (which is the number —a whose addition to the given number a results in 0) and the notion of the recip10 . For instance. b and c the identity (a-\-b) c = ac-\-bc the distributive law for multiplication over addition must hold. for any numbers a. There is a certain kind of analogy between the operations of addition and multiplication which is particularly noticeable because the properties of addition are in many respects similar to the properties of multiplication. For instance. if we set the unusual "proportion" addition subtraction multiplication ? then everybody will substitute the word "division" for the interrogation sign even without analysing the meaning of the "proportion".

if we take the distributive law (a -(. Further.b) c = ac + be and interchange addition and multiplication we get the "equality" ab + c = (a e) (b + c) with which nobody of course can agree. from the last identity it follows that a number different from 0 cannot be divided by 0). 11 . we see much similarity between the properties of an arithmetic progression (which is a sequence of numbers for which the difference between any member and the preceding member is one and the same) and a geometric progression (which is a sequence of numbers for which the ratio of any member to the preceding one is one and the same). (Since we obviously have (a + c) (b + c) = ab + ac + be + c2 = = ab -f c (a + b -f c) it follows that (a + c) (b + c) — ab -f c only when c — 0 or when a + b + c = 1. By the same reason. the number 0 plays a special role not only in the addition but also in the multiplication because we have o-0 = 0 for any number a (in particular.) There are however some other algebraic systems whose elements are not numbers such that it is also possible to deiine the operations of addition and multiplication for them and the similarity between the addition and the multiplication of these elements is even closer than the similarity *) If the e q u a l i t y a + 1 = 1 were fulfilled for any a then it would be impossible to subtract 1 from a n y number different from 1. However.(tïie multiplicative inverse) of a number a (which is the number —whose product by the given number a rocal is equal to 1). For instance. I n reality t h i s is not the case: for instance. 3 — 1 = 2. this analogy is not complete. But if we replace in this identity multiplication by addition and zero by unity we arrive at the meaningless "equality" a + 1 = 1 which can hold only when a = 0 1 ).

3 and the set В consists of the points in the area shaded by inclined lines then the set A + В is the whole area shaded in Fig. 14. 10. if A is the set of all even positive integers and В is the set of the positive integers divisible by 3 then {2. It is quite evident that the addition of two sets can be defined in the following way: by the sum A В of a set A and. a set В we shall simply mean the union of these sets.between the same operations on numbers. the set of points in a square. It is clear that the addition of positive numbers and the addition of negative numbers are completely different operations: for instance. 3. the addition of fractious performed according to the rule alb + -f. The fact that we have defined a completely new operation and have called it "addition" must not seem strange: it should be remembered that earlier when we passed from numbers of one kind to numbers of another kind we defined the operation of addition in a different way. the set of points bounded by a circle. 3. 8. For instance.. For instance. 22. 6. the set of elephants in India.which are called the elements of the set. if A is the set of the boys in a class and В is the set of the girls in that class then A + В is the set of all the pupils in the class. 9. the sum of the numbers 5 and (—S) is equal to the difference between the (positive) numbers 5 and 8. 3 the set of even numbers. let us consider a very important example of an algebra of sets. If the set A consists of the points belonging to the area shaded by horizontal lines in Fig. the set of elements in the periodic system. 4. similarly. 12. 15.} is the set A + В which consists of both groups of numbers. the set of grammar mistakes in your composition etc. Fig. . we can consider the set of the pupils in a given class. 20. 21. 16. By a set is meant any collection of arbitrary objects..c/d = {ad bc)!bd differs from the addition of whole 12 . Similarly. 18. For instance.

numbers: the definition of the addition of the positive integers (see Fig. To facilitate the analysis it is convenient to consider special diagrams demonstrating operations on sets. Petersburg. Such diagrams are often called Venn's diagrams after the English mathematician John Venn (1834-1923) who used these diagrams in his study of mathematical logic. Euler 1 ) used such diagrams much earlier than J. a famous Swiss mathematician who spent most of his life in Russia and died in St. It follows that the set (A + B) -f С (or. we have A + В = В + A for any two sets A and B. Venn 2 ). which means that the commutative law holds for the addition of sets. the numbers 3 and 5 or the pupils Peter and Mary). 1 ) Leonard Euler (1707-1783). Further. 2 ) In his studies in mathematical logic L. 3. The usage of one and the same term "addition" in all the cases was accounted for by the fact that the general laws for the operation of iddition of whole numbers remained valid when we passed. the set of all whole numbers or the set of all p upils in a school) will be represented as a square. В and С there always holds the identity (A + В) + С = A + (В + C) This means that the addition of sets obeys the associative law. the corresponding •liagrams (which. that is for the addition of sets. 1) given on page 8 is inapplicable to the description of the addition of fractions. In this representation the sets consisting of some of the elements of the given set (for instance the set of even numbers or the set of excellent pupils) are represented by some parts of that square. therefore. 13 . within this square we can mark different points representing some concrete elements of the set (for instance. which is the same. it is obvious that for any sets А. for instance. Let us make the following convention: the set of all the elements under consideration (for instance. It would be more precise to call them Euler's diagrams because L. to fractions: in both cases the operation of addition turned out to be commutative and associative. Euler represented differ1'iit sets of objects by circles in the plane. Now let us check whether these laws remain valid for the new operation of "addition". do not differ from Venn's diagrams) "t 1 often referred to as Euler's circles. in principle. As is clearly seen from Fig.

12. 18. if A is the set of the chess-players in your class and В is the set of the swimmers in your class then А В is the set of those chess-players who can also swim. It is quite clear that the multiplication of sets defined in this way obeys the commutative law. if A is the set of the positive integers and В is the set of numbers divisible by 3 then the set А В is {6. .} This set consists of all positive integers divisible by 6. that is for any two sets A and В we have AB = BA (see the same Fig. 24. 4 where the set A -f. 4 Fig. . it is also obvious that "the set of the chess-players who can swim" and "the set of the swimmers 14 . В and С (see Fig. that is the intersection of these sets (also called the meet of A and В). . Now let us agree that by the product А В of two sets A and В will be meant their common part.C)) can simply be denoted as A + В + С without using parentheses.В -f С coincides with the whole area shaded in this figure). For instance. 5 by horizontal lines and the set В consists of the points in the area shaded by vertical lines then the set AB is the area in the figure which is cross-hatched. If the set A consists of the points lying in the area shaded in Fig. 5.Fig. the set A + В + С is nothing but the union of all the three sets A. 5 the set A + (B -f.

В and С there also holds the distributive law (A + В) С = AC + ВС which means that intersection distributes over union. Further. which is the same. that is divisible by 30. A is the set of the chess-players in your class and В is the set of the pupils who can play draughts while С is the set of the swimmers then the set A + В is the union of the set of the chess-players and the set of the pupils who can play draughts. for instance. On the other hand. The set (A -(. В the set of integers divisible by 3 and С the set of integers divisible by 5. A remarkable fact is that for any three sets A. the set ABC is the common part (the intersection) of the three sets A. the set А (ВС) simply as ABC without using parentheses. 6 the set ABC is covered by the network of the three families of hatching lines 1 ).who can play chess" coincide: this is simply one and the same set). ВС is the set of integers divisible by 15 and A(BC) is the set of even integers divisible by 15. 15 . Then AB is the set of integers divisible by 6 and (AS)C is the set of integers divisible both by 6 and 5. hence. it is quite evident that the associative law also holds for the multiplication of sets.В) С can be obtained from the set A + В if we choose from the latter only those pupils who can swim. Let A be the set of integers divisible by 2. we see that А (ВС) coincides with the set (AB)C of all integers divisible by 30. that is the set of those pupils who can play chess or draughts or both. В and С (in Fig. that is for any three sets A. В and С we have (AB) С = А (ВС) The associative law allows us to denote the set (AB)C or.' if. But it is quite clear that exactly the same set can be obtained if we form the union AC + ВС of the set А С of the chess-players Here is one more example demonstrating the associative law for the product of sets. Indeed.

Indeed. The verbal explanation of the distributive law is rather lengthy. 7. if this set (we shall denote it 0) is added to an arbitrary set the latter must not change and therefore the set О must contain no elements at all.a the set A + В is shaded by horizontal lines and the set С by vertical lines so that the set (A + В) С is covered by a network of hatching lines. One may think that since the set О is empty and contains no elements there is no need to take it into account. If we did so this would resemble the exclusion of the number 0 from the number system: a "collection" containing zero elements is also "empty" and it may seem senseless to speak of the "number" of elements contained in such a collection. In Fig. 7.a. Buttt would be in fact quite unreasonable to exclude the empty set from the consideration. Thus. 7. It is easy to understand which "set" plays the role of "zero" in the "algebra of sets". О is an empty set (it is also referred to as a void or null set). it is quite meaningful. In Fig. If we did not introduce the number 0 it would be impossible 16 .b the sets AC and ВС are shaded by lines with different inclination.(b) Га) Fig.b does not differ from the area {A + В) С covered by cross-hatching in Fig. 7 who can swim and the set ВС of the pupils who can play draughts and can swim. But it is clearly seen that the area AC + ВС in Fig. the set А С + ВС in the figure coincides with the whole shaded area. But it is in fact by far not senseless. 7. moreover. To explain this law it is also possible to use the graphical demonstration.

(For instance. the difference 3—3 would be equal to no number in this case). I lie intersection of the set of the girls in your class and the set of all those pupils whose height exceeds 2. eight ones and no tens! There are many other important tilings which would be Fig. the intersection of the sets A and В shown in Fig.5 m is empty. -420 17 . it would be impossible to speak of "the set of the pupils in a class whose name is Peter" because this set may not exist at all. That is why the introduction of the number zero is considered one of the most remarkable events in the history of the development of arithmetic. for instance. the intersection of the set of excellent pupils in your class and the set of elephants is also empty. It is quite clear that if О is an empty set then we have A + О = A for any set A . It is also evident that for any set A we always have AO — О because the intersection of an arbitrary set A and the set О (which contains no elements) must be empty. Similarly. If we did not have the notion of an "empty" set it would be even impossible to mention some sets: for instance. that is it may turn out to be an empty set. if we did not introduce the notion of an empty set О it would be impossible to speak of the product (the intersection) of any two sets: for instance. Without having the number 0 at our disposal it would be very difficult to write the number 108 in the decimal number system because this number contains one hundred. 8 is empty.(о subtract any number from any other number (because.) The last identity is known as one of the intersection laws (the other intersection law will be presented later). 8 impossible without the number zero. Analogously.

as the set / we can take the set of all the pupils of a given school or class or the set of all positive integers or the set of all the points of 18 . 9 Now we proceed to a more complicated question concerning a set playing the role analogous to that of the number 1 in the system of numbers. From this requirement it follows that the set I must contain all the elements of all the sets A\ It is clear however that such a set can only exist if we limit ourselves to those sets whose elements are taken from a definite store of "objects": for instance.) Thus. by / we shall always mean some underlying basic set containing all the objects admissible in a particular problem or discussion. Similarly. We can also consider the sets of points forming various geometrical figures lying within a definite square such as those represented in Figs. we can limit ourselves to the positive integers (then A can be. for instance. if we limit ourselves to the sets of pupils of one definite school or of one definite class (such a set A may be the set of excellent pupils and another set В may be the set of the chess-players). the set of the even positive integers and В can be the set of prime numbers which have 110 divisors except themselves and unity). (We remind the reader that when we introduced Venn's diagrams on page 13 we stipulated the existence of such a "set of all the elements under consideration".This set (we shall denote it as I) must be such that the product (that is the intersection) of the set I and any set A coincides with A. 3-8.Fig. For instance.

the set / contains all the objects under consideration and therefore it cannot be enlarged: when we add an arbitrary set A (this set A must of course belong to the class of sets we deal with) to the universal set I we always obtain the same set I. in the general case the rule obtained from the equality a -0 = 0 by replacing multiplication by addition and zero by unity does not hold for numbers: for almost all numbers a we have a -f 1 Ф 1. For instance. almost all basic laws which are known for numbers also hold for the algebra of sets but the algebra of sets also has some other completely different laws which may seem strange when they are first encountered. Indeed. As has been shown. In "algebra of sets" the set / is referred to as the universal sel. we see that the operation laws for the "algebra of sets" we have constructed are in many respects similar to the laws of elementary algebra dealing with numbers but at the same time the former laws do not completely coincide with the latter laws. let. As to the algebra of sets. A be again the set of the chess-players. It is evident that for any "smaller" set A (and even for the set A coinciding with /) we have the second intersection law AI = A The last identity resembles the well-known arithmetic equality defining the number unity. by definition. В and C) the equality AB + С = (A + С) (В + С) expressing the second distributive law of set theory (this is the distributive law for addition over multiplication). 9). the situation is quite different: in this case we always have A + I = I Indeed. But in the algebra of sets the situation is reverse: in this case we always have (for any sets A. as was already mentioned. В the set of the pupils who play draughts and С the set of the 19 . for instance. Further.a square (Fig. if we take the distributive law (a + b) с = — ас be which holds for numbers and interchange addition and multiplication in it we arrive at the meaningless "equality" ab + с = (a + c) (b + c) which turns out to be wrong for numbers in almost all cases. Thus.

10. which proves the second distributive law. the unions A + С and В + С of the sets A and С and of the sets В and С respectively consist of the pupils who can play chess or can swim or both and of the pupils who can play draughts or can swim or both. This verbal explanation may seem too lengthy and therefore we shall also present the graphical demonstration of the second distributive law of the set theory. the intersection (A + С) (В + C) of these two unions is covered by the "network" of hatching lines. perhaps.6 exactly coincides with the whole area shaded in Fig. 10. obviously. fa) гь) Fig. But it is readily seen that the area covered by the network of horizontal and vertical lines in Fig. the intersection А В of the sets A and В consists of all the pupils who can play both chess and draughts and the union A B -f.a. draughts and can swim). 10.a the intersection А В (of the sets A and B) and the set С are shaded by hatching lines with different inclinations. It is clear that the intersection {A -f С) {В + С) of these two unions includes all the pupils who can swim and also those pupils who cannot swim but can play both chess and draughts. Then. 10. In Fig. which means that this intersection coincides with the set A B + C. the whole shaded area in the figure represents the set A B + C. 10 20 . On the other hand.С of the sets AB and С consists of all the pupils who can play both chess and draughts or can swim (or. In Fig.swimmers in your class.6 the union A + С of the sets A and С is shaded by horizontal lines and the union В -} С of the sets В and С is shaded by vertical lines. can play chess.

when we pass from numbers to sets we encounter a completely different situation: it turns out t h a t there are a number of laws of algebra of numbers which are inapplicable to the algebra of sets 1 ). in all such cases related to algebra of numbers we have to learn some new facts and laws but the facts learned previously remain valid. But it is nevertheless expedient to write down here the basic laws of set algebra using the standard settheoretic notation: A[} B = B \) A and A ft B = B A the c o m m u t a t i v e l a w s (A U В) и c = A (J {В и С) and (А П В) П С = А Г) (В Г) С) the associative laws 21 . therefore for our aims it is natural to use the ordinary symbols -f. when we pass from whole numbers to fractions or to signed numbers supplied with the signs "plus" or "minus"). The fact that the general laws of algebra retain one and the same form for all kinds of numbers is very important: this makes i t possible to use our experience acquired in studying numbers of one kind when we pass to another kind of numbers (for instance. However. but in a completely different way: the union of sets A and В is denoted as A U В and the intersection of these sets as А П B.In conclusion we present two more laws of algebra of sets which essentially differ from what is known from elementary algebra dealing with numbers. In other words. The usage of the symbols + and • makes it possible to indicate in a visual way the similarity between elementary algebra and the new algebraic systems. In the present book we shall deal not only with algebra of sets but also with some other algebraic systems in which the operations of "addition" and "multiplication" obey the same laws as in algebra of sets. l ) It is this distinction between the laws of algebra of sets and the laws of algebra of numbers that accounts for the fact that in many books the addition and the multiplication of sets (that is the operations of forming the union and tne intersection of sets) are denoted not by the usual symbols + and -.and •. It is easy to understand that for any set A the union of this set and another replica of the same set and also the intersection of the set A with itself coincide with the original set A: A + A = A and AA = A These two identities are called idempotent laws (the first of them is the idempotent law for addition and the second is the idempotent law for multiplication). instead of the symbols [J and f| used in set theory.

A = A n times for any A and n. that is why. We first of all mention the relation A + I = I which indicates a significant distinction between the universal set / and the number 1. we have (A +B)(B + C) (C'+A) = ABC + = AAB = (ABC + ABC) A\j 0 = A ACC + + BBC + ABB + AAC + + ВС С + ABC = + (AB + AB) + and A [] 1 = A A U / = / and A f\ 0 = 0 the properties of the empty set О and of the universal set 1 (А и В) n с .С For instance..(А П С) и (В n С) and (A f] В) U С = = (А и С) П (В и С) the distributive laws A U the А ^ idempotent A and Аlaws Г\ Л = А 22 . the idempotent laws A + A = A and + D AA = A are completely new to us. . Another peculiarity of algebra of sets is the way in which the "parentheses are opened" using the second distributive law (A + C) (B + C) = AB -f.. for instance.Let us enumerate these new laws. For in the algebra of sets we have A + A+.. we can express the meaning of these laws verbally by saying that the algebra of sets involves neither exponents nor coefficients.+A = A • n times and A-A .. in algebra of sets we have (A +D)(B + D) (C + D) = [{A + D) (В +/))] (С + D) = = (AB + D) (C + D) = (AB) С + D = ABC Finally..

(A + В) (В + С) (C + D) = AC + ВС + 8. A (A + I) {В + 0)= AB 6. Boolean Algebra Let. (A+B) (A + 1) + (A + В) (В + О) = A 9. us write down all the general laws of algebra of sets we have established: A+ B= B+ A and AB = В A Uk* C o m m u t a t i v e l a w s . Exercises 1'rove the following equalities in which capital letters denote sets (the letter О always denotes the empty set and (lie letter I denotes the universal set): 1. A (A + В) = A 3.+ (AC + AC) + (ВС + ВС) = + = ABC + AB AC + ВС (cï. (A + B)(B + I)(A +0) = A 10. (A + В) (В + С) (C + A) = AB + ВС + 7. (A + В + С) (В + С + D) (С + D + А) = AB + AD + BD + С Example: A (A + 6) (В + С) = + ВС СA BD + В - A[(A + C) X the a s s o c i a t i v i t y of m u l t i p l i c a t i o n X (B -f C)] = A(AB + C) the second distributive law = (AB)A-{~CA the first d i s t r i b u t i v e law = (AB + C)A = the e o m m u t a t i v i t y of m u l t i p l i c a t i o n = (AA)B + AC = the e o m m u t a t i v i t y and the a s s o c i a t i v i t y of multiplication = AB + AC t h e idempotent l a w for m u l t i p l i c a t i o n 2. Exercise 6 below). A (A + С) (B + C) = AB + AC 5. AB + A = A 4. (A -f В) (A + С) (B + D) (C + D) = AD 2.

is there only one realization of such an algebra or is there also a number of these "algebras" which differ from one another in the elements on which the operations are performed and in the definitions of the operations (as before we shall call these operations a d d i t i o n ' a n d multiplication) b u t at the same time are similar in the basic properties of the operations? The reader can undoubtedly anticipate the answer to the question: there are in fact many algebras similar to the algebra of sets (in these algebras the same general operation rules hold). First of all.{A + B) + C ^ A + {B + C) and (AB)C = A ( B C ) the associative laws (A + В) С = AC + ВС and AB-\-C — (A -f С) (В -f. also an "algebra" but it is new to us and rather unusual. In this connection it now appears natural to ask.C) the distributive laws A + A = A and AA = A the idempotent laws Besides. of course. the "algebra of complex numbers" (that is of the real and the imaginary numbers) and so on. what is the situation in the unusual algebra of sets? In other words. the algebra of sets is. the "algebra of rational numbers" (by the rational numbers are meant both the integers and the fractions). the "algebra of signed numbers" (that is of the positive and non-positive numbers) and also the "algebra of real numbers" (that is of the rational and irrational numbers). All these "algebras" differ from one another in the numbers on which the operations are performed and in the definitions of the operations of addition and multiplication but the general properties of the operations remain the same in all the cases. we can consider the "algebra of positive integers". there are a variety of algebras 24 . Now it should be noted t h a t we have in fact not one ordinary algebra of numbers but m a n y such "algebras": indeed. tlie set algebra contains two "special" elements (sets) О and / such that A + О = A and AI = A A + I = / and AO — О Those laws (identities) are similar to the ordinary laws of the algebra of numbers but they do not coincide completely with the latter.

b. we shall call these elements 25 . Some definitions given below may seem strange because they are new. Algebra of two numbers (elements). we can consider the "algebra of sets of pupils in your class". a completely different algebra!). 2 on page 8). respectively. t h a t is when they satisfy definite general logical requirements. these new rules and definitions may of course seem strange to the pupils until they get used to them. For instance.of sets themselves: for instance. Example 1. But after these rules have been chosen we have no right to ask. and new things always seem strange when they are first encountered. we define the sum a -f b as being the element с and. But there are also completely different examples of algebras having similar properties. means to indicate some rules according to which we attribute to any two objects a and b two other objects с and d called. below we shall discuss some examples of this kind. . 1 on page 8) and t h a t the product ab is equal to the number of objects in the union of a collections each of which contains b objects (Fig. the "algebra of sets of animals in a zoo" (this is. Before proceeding to these examples the reader should realize that to define the operations of addition and multiplication in a set of objects (elements) a. Later the pupils are taught fractions and are told that the sum and the product of fractions alb and dd are defined according to the rules given on page 9. Let us assume that the algebra consists of only two elements. 3-10). why the sum of a and b is equal to c? Indeed. the sum and the product of a and b: с — a -f b and d = ab These rules must be chosen so that all the laws characteristic of the algebra of sets are fulfilled. Now we proceed to the examples. pupils are first told that the sum of two numbers a and b is the number of objects in the union of a collection containing a objects and a collection of b objects (see Fig. of course. . . the "algebra of sets of numbers" (these numbers can be of different kinds). as is known. the definitions must not be discussed when they are consistent. the "algebra of sets of points lying within a square" (see Figs. the "algebra of sets of books in a library" and the "algebra of sets of stars in the sky". 'or the sake of simplicity.

It is a little more difficult 26 . the "addition table" in this new algebra has the form 0 i + 0 1 1 1 0 1 It is obvious that in the algebra thus defined both commutative laws hold: a + b = b + a and ab = ba for any a and b It can readily be verified that the associative laws also hold for this algebra: [a -}. b and с There is no need to verify the associative law for multiplication because the "new" multiplication completely coincides with the multiplication of numbers for which. It is also clearly seen that the idempotent laws also hold for this algebra: a + a —a and aa = a for any a that is for a = 0 and for a = 1 (now we see why it was necessary to put 1 + 1 = 1!). Thus. we shall define it in an "almost ordinary way" with the only distinction from the ordinary arithmetic that now the sum 1 + 1 is not equal to 2 (this "algebra of two numbers" does not contain the number 2 at all!) but is equal to 1. as we know. the associative law holds."numbers" and denote them by the familiar symbols 0 and 1 (but in the case under consideration these symbols have a completely new meaning).(b c) and (ab) с = a (be) for any a.b) -f. that is by means of the following "multiplication table": 0 1 • 0 1 0 0 0 1 As to the addition.с = a -f. We shall deline the multiplication of these numbers in exactly the same way as in ordinary arithmetic.

a -—. that is we shall always have (for a = 0 and for a = 1) a + 0 = a and a-1 = a\ a + 1 = 1 and a . Algebra of four numbers (elements). Suppose that the elements of the algebra are four "numbers" which we shall denote as the digits 0 and 1 and the letters p and q. b and с For instance. l = l . a — q and a = 1) 27 . if we agree that the number 0 plays the role of the element 0 and the number 1 plays the role of the element I. l = 1 and (1-1) + (1-1) = 1 + 1 = 1 (1-1) + 1 = 1 + 1 = 1 and (1 + 1)-(1 + 1) = M = 1 Finally.p.0 = 0 Example 2.b = b + a and ab = ba for any a and b (a + b) -f с = a + (b + c) and (ab) с = a (be) for any a.be and ab \-c -= (a 4 c) (b + c) a for any a.to verify the distributive laws: (a + b) с = ас + be and ab + с = (a + c) (b + c) for any a. the rules concerning the "special" elements О and I will also hold. b and с -f. in this algebra of two elements we have (1 l ) . b and с ta -L Ь) с — ас -f. The addition and the multiplication in this algebra will be defined with the aid of the following tables: + 0 p 1 1 0 p 4 1 и p я 1 p p 1 1 4 1 4 1 1 1 1 1 0 p q 1 0 0 0 0 0 p 0 p 0 q 0 0 <7 p я l 0 p q l As can readily be checked by means of the direct computation.a = a and aa — a for any a (that is for a = 0. in this case we also have a -j. This is a slightly more complicated example of the same kind.

2/3 and 1 then the "addition table" and the "multiplication table" for these numbers have the form 1 3 1 2 2 3 2 2 3 3 1 2 2 3 1 3 1 3 1 2 2 3 1 2 1 2 1 2 2 3 2 3 2 3 2 3 2 3 1 1 1 1 1 Ф 0 1 3 i 0 0 i 1 ® 0 1 3 1 2 2 3 1 0 0 0 0 0 0 1 3 1 3 1 3 1 3 1 3 1 2 1 2 1 о 1 3 1 2 2 3 2 . . By the product x ® у of two numbers x and у we shall mean the least (the minimum) of these numbers in case x Ф у and any of them in case x = y. Namely. As the elements of the algebra let us take the numbers contained in an arbitrary (bounded) number set.. let us agree that the elements of the algebra are some (or. . v..1 = a\ a 1 = 1 and a . z] and the minimum number among these 28 . . we shall define them in a completely new manner. To avoid the confusion between the ordinary addition aild multiplication and the new operations we shall even denote the latter by the new symbols ® (addition) and ® (multiplication). 1/3. As to the operations of addition and multiplication.0 = 0 Example 3. if the elements of the algebra are the numbers 0. let us assume that the sum i f f i у of two numbers x and у is equal to the greatest (the maximum) of these numbers in case x Ф у and to any of them in case x = y. v. .. the numbers 0 and 1 play the roles of the elements О and I of the algebra of sets respectively because for any a we have a + 0 = a and a . perhaps. that is numbers lying between 0 and 1 and the numbers 0 and 1 themselves. Algebra of maxima and minima. 1/2.Besides. for instance. .3 1 3 1 2 2 3 1 \i 1 1 1 3 1 2 2 3 1 0 0 0 0 1 I The maximum number among two or several numbers u. For instance. 2 is often denoted in mathematics as max Iu. all) numbers x satisfying the condition 0 x ^ 1.

Х9у

<У

АГ®У

—О—

X

Oh

У

Fig. i l

**numbers as min [и, v, . . ., zj. Thus, in this "algebra of
**

maxima and minima", we have, by definition,

х ф y = max [x, y]

and

x <8> y = min [x, y 1

**We can also agree to represent numbers as points on the
**

number line. Then the numbers x satisfying the condition

0 ^ x ^ 1 are represented by the points of a horizontal

line segment of length 1, the sum x ф y of two numbers

x and y is represented by the rightmost of the points x and

y and their product x ® y is represented by the leftmost point

(Fig. 11).

It is clear that the new operations of addition and multiplication we have defined satisfy the commutative laws:

x Ф у —у Ф x

and

x <g> y = y <8> x

**The associative laws
**

(x Ф у) Ф z = x ф (у Ф z) and

(x <g) i/) <S> z = .r ® (?/ ®

z)

**also obviously hold because the number (x ® у) ф z (or,
**

wiiich is the same, the number x Ф (у Ф z)) which can

simply be denoted as x Ф у Ф z without the parentheses

is nothing but max [x, y, z] (Fig. 12) while the number

(x ® y) ® z (or, which is the same, x <g> (у ® z)) which

can simply be written as x <8> y ® z without the parentheses is nothing but min [x, y, z] (see the same Fig. 12).

It is also quite clear that the idempotent laws also

hold here:

x Ф x = max [x, x] — x and x <S> x = min [x, x] = x

Finally, let us check the validity of the distributive laws

(x Ф y) ® z = (x О z) Ф (y (8) z)

01

о

y 97

О

«®у-- y Ф / : * ® y ®Z

-OУ

7

Fig. 12

29

**(x®y)®z=(x®z)@(y®z)
**

y®Z

x®z

о h-о

о

у

Z

y

О Н /

X

y®Z

О I—о

у

(а)

**(x®y)®z=(x®z)g)(y®z)
**

x©y=x®z

о

О

х

Z

11

(Ь)

Fig. 13

and

H } ) 0 z

=

(iez)»(!/®z)

**It is evident that the number
**

(x © y) ® z = min {max [x, y], z}

is equal to z if at least one of the numbers x and у is greater

than z and is equal to the greatest of these numbers if both

x and у are less than z (Fig. 13,a and b). It is also clear that

the number

(x <8> z) © (у ® z) = max {min [x, z], min [y, z]}

is equal to the same value (see again Fig. 13). Analogously,

the number

(x <8> y) © z = max {min [x, y], z}

is equal to z if at least one of the numbers x and у is less than z

and is equal to the minimum of the numbers x and у if

both x and у exceed z (Fig. 14,a and b). As is seen from the

same Fig. 14, the number

(x © z) <8> (у © z) = min {max [x, z], max [y, z]}

is also equal to the same value.

Now to make sure that all the laws of the algebra of sets

hold for the n:w unusual algebra of maxima and minima

it is sufficient to note that the role of the elements О and I

of the algebra of sets is played by the smallest number 0

among all the numbers under consideration and by the greatest number 1 respectively. Indeed, for any number x satis( х ® у ; ф z=(x®zmy@z)

x®z = y®z

x®y

oi—o-!.

у

х

о

Z

ii

oi—о

(a)

о

X

(b)

Fig. 14

30

Z

**(x®y)®z=(x®z)®(y®z)
**

x®y=x®z

У®г

о

y

II

**the condition 0 ^ x ^ 1 we always have
**

x © 0 = max [x, 0] = x

and x ® 1 = min [x, 1] = x

x Ф 1 — max [x, 1] = 1 and x ® 0 = min [x, 0] = 0

lying

**Example 4. Algebra of least common multiples and greatest common divisors. Let N be an arbitrary integer. As the
**

elements of the new algebra we shall take all the possible

divisors of the number N. For instance, if N = 210 =

= 2 -3 -5 -7 then the elements of the algebra in question are

the numbers 1, 2, 3, 5, 6, 7, 10, 14, 15, 21, 30, 35, 42,

70, 105 and 210. In this example we shall define the addition and the multiplication of the numbers in a completely

new way: by the sum m © n of two numbers m and n we

shall mean their least common multiple (that is the smallest

positive integer which is divisible by both numbers m and n)

and as the product m ® ж of the numbers m and n we shall

take their greatest common divisor (that is the greatest integer

by which both m and n are divisible). For instance, if N = 6

(in this case the algebra contains only the four numbers

1 , 2 , 3 and 6) the addition and the multiplication of the

elements of the algebra are specified by the following tables:

1 2 3 6

e

i

2

3

6

1

2

3

6

2

2

6

6

3

6

3

6

6

g

<8>

and

6

6

1

2

3

6

1 2 3 6

1

1

1

1

1

2

1

2

1

1

3

3

1

2

3

6

**In "higher arithmetic" (number theory) the least common
**

multiple of two or several numbers m, n, . . ., s is often

denoted as [m, n, . . ., s] and the greatest comtaon divisor

of the same numbers is denoted as (m, n, . . ., s). Thus, for

this algebra we have, by definition,

m Ф n = \m, n\

and

m ® n = (m, n)

**For instance, if the algebra contains the numbers 10 and 15
**

! lien

Ю e 15 = [10, 151 = 30

and

10 ® 15 = (10, 15) = 5

**It is evident that in this algebra we always have
**

m ф n = n ф m

and

m <8> n -= n <8> m

31

re. p). 105).Further. (14. the number (m ® n) © p = [(m. 105) = (70. 105)] = = [5. m] = m and m ® m = (m. p) (the latter number can simply be denoted as m ® n ® p). The idempotent laws m © m = [m. in 32 . n]. m) — m are also quite evident. p)] and therefore we always have (m © re) ® p = (m ® p) + (re <8> p) For instance. re. p] is the least common multiple of the number p and the greatest common divisor of the numbers m and re. (n. 7] = 35 Analogously. p) is nothing but the greatest common divisor of the number p and the least common multiple of the numbers m and n (think carefully about this expression!). we have (m © re) © p = m © (re © p) = [to. The verification of the distributive laws is a little more lengthy. This number contains those and only those prime factors which are contained in p and are simultaneously contained in at least one of the numbers m and n. The number (m © re) ® p = ([m. if we limit ourselves to the divisors of the number 210 then we have (10 © 14) ® 105 = (110. re). p] (we can denote this number simply as m © re © p without the parentheses) and also (m ® re) ® p = m ® (re ® p) = ( m . 14]. it contains those and only those prime factors which are contained in p or in both numbers m and re or in all the three numbers p. 105) = 35 and (10 ® 105) © (14 ® 105) = [(10. But it is evident that these (and only these) prime factors are also contained in the number (m ® p) © (n <g> p) = [(m.

**;i!i(i п. But exactly the same factors are contained in the
**

number

(m фр) <8> (n ® p) = ([m, p], [n, p])

,IH1 therefore we always have

(m ® n) © p = (m Ф p) <S> {n ф p)

For instance,

(10 ® 14) ф 105 = [(10, 14), 105] = [2, 105] = 210

and

(10 e 105) ® (14 0 105) = ([10, 105], [14, 105]) ==

= (210, 210) = 210

Finally, in this case the roles of the elements О and 1

of the algebra of sets are played by the smallest number 1

among the collection of numbers we deal with and by the

greatest number N respectively. Indeed, this algebra only

contains the divisors of the number N and we obviously have

m Ф 1 = [m, 1] = m

m ф N = [m, N] = N

and

and

m ® N = (m, N) = m

m ® 1 = (m, 1) = 1

**We thus see t h a t for this algebra all the laws of the algebra
**

of sets are fulfilled.

llence, there are different systems of "objects" (of elements

of the algebra in question) for which it is possible to define

the operations of addition and multiplication

satisfying all

the known rules fulfilled in the algebra of sets: the two commutative laws, the two associative laws, the two distributive laws, the two idempotent laws and the four rules specifying the properties of the "special" elements whose role in

these algebras is close to that of zero and unity. Later on we

f-hall consider two more important and interesting examples

of such algebras.

Now we proceed to the study of the general properties

«>! all such algebras and our immediate aim is to give a general

"name to all of these algebras. Since algebras with such

grange properties were first considered by the distinguished

English mathematician George Boole who lived iu the

1'tth century, all the algebras of this kind are called Boolean

algebras1). For the basic operations on the elements of a Boo!

**\ Л rigorous definition of a Boolean algebra is stated in Appendix
**

117.

33

G E O R G E BOOLE

(1815-1864)

lean algebra we shall retain the terms "addition" and "multiplication" (but the reader should bear in mind that in

the general case these operations differ from the ordinary

addition and multiplication of numbers!). We shall also

sometimes refer to these operations as the Boolean addition

and the Boolean

multiplication.

In his "Laws of Thought" which first appeared in 1854,

that is more than a hundred years ago, G. Boole investigated

in detail this unusual algebra. The title of G. Boole's work

may first seem strange; however, after the reader has studied

this book it will become clear what is the relationship between unusual algebras considered in the book and the laws

of human thought. At present we only note that it is this

relationship between the Boolean algebras and the "laws of

thought" that accounts for the fact that the work of G. Boole

to which his contemporaries paid little attention is of such

great interest nowadays. In recent years the book by G. Boole

has been many times republished and translated into various

languages.

Exercises

1. Verify directly that for all triples of elements of the

"Boolean algebra containing two numbers" (see Example

1 on page 25) there hold both distributive laws.

2. Check the validity of both distributive laws for several

triples of elements of the "Boolean algebra with four elements" (Example 2 page 27).

3. (a) Let there be a family in which there is only one

schoolboy. Then all the "sets of the schoolboys in the family"

are the following: the set / containing one schoolboy and

the set О containing no schoolboys (the empty set). Compile

the "addition table" and the "multiplication table" for the

"algebra of sets of schoolboys in the family" (this algebra

consists of only two elements О and I) and compare these

tables with the tables on page 26. Proceeding from this

comparison show that in the "algebra of two numbers"

considered in Example 1 of this section all the laws of the

Boolean algebra hold.

(b) Let there be a family in which there are two schoolchildren Peter and Mary who go to school. Then the "algebra

of sets of schoolchildren in the family" consists of four elements: the set / containing both schoolchildren, the two

3

35

0 < а 2 < Az. 8) = [(12. у]. 8] = ([12. 30]. T]> T } = ([12. -j and 1 in which x © y = max [x. n] and m ® n — (m. 30).. Compile the "addition table" and the "multiplication table" for this algebra of sets and compare them with the tables on page 27. 0 <Cah^Ah . 8]. 1 ] . n). 6*. .. 4. . i]} = min{max[-i. ±]} and m a x { m i n (b) [ y . . 8]) 5. = = max{min[i. Show that some of the laws of the Boolean algebra are fulfilled for this algebra of the divisors.. (b) Compile the "addition table" and the "multiplication table" for the algebra of the divisors of the number 12 in which m Ф n = [m. Proceeding from this comparison show that in the "algebra of four elements" considered in Example 2 of this section there hold all the laws of the Boolean algebra. min[i. Check that (a) m i n { m a x [ 4 . m a x [ { . h phh Then any two divisors m and n of this number can be written as where 36 m = pVp? • • • Plh 0 < а 1 < Л 1 . (30. y] and x <8> y = min \x. 1 ] . I ] . Verify the validity of the laws of the Boolean algebra for this algebra of three elements.. 8). [30. 8)] and [(12. (a) Compile the "addition table" and the "multiplication tablé" for the Boolean algebra consisting of the three numbers 0. Let the decomposition of a (positive integral) number N into prime factors be of the form •r At N = рТрг Аг A .sets P (Peter) and M (Mary) each of which contains one of the two schoolchildren and the empty set 0 .

in a Boolean algebra there holds the equality A (A + С) (B + C) = AB + AC (which was proved earlier for the algebra of sets. We first of all see a complete parallelism between the properties of the Boolean addition and the Boolean multiplication.. 3.. bh can be equal to zero). Further Properties of Boolean Algebras. n) is a Boolean algebra. the similarity between the operations is so close that in every (correct) formula of a Boolean algebra we can interchange the addition and the multiplication: the equality resulting from the interchange remains valid. n\ and m (g) n = (m. .... see the exercises to Sec. I o n page 23). For instance. n] (the least common multiple of the numbers m and n) and of the numbers (m. . .. It should be taken into account t h a t when an equality fulfilled for a Boolean algebra involves the "special" elements 0 and I then the interchange of the Boolean addition and the Boolean multiplication in this equality must be followed by the internange of the elements 0 and I. . b2. . . Boolean Equalities and Inequalities Let us continue to study Boolean algebras. ah and b1. a2.О<А<Л„ (some of the numbers аг. W h a t form have in this case the decompositions into prime factors of the numbers [m.and bi i>2 . .. the validity 37 . On interchanging the addition and the multiplication in this equality we obtain A + AC + BC=(A + B)(A + C) • he latter equality is also valid (see page 39). n = pSpo b h p^ where о<А<Л„ 0<&2<л2. For instance. Principle of Duality. n) (the greatest common divisor of the numbers m and n)l Using these decompositions prove t h a t the set of all the divisors of the number N with the operations m © n = [m.

the idempotent law for addition and the idempotent law for 1 ) The "new" equality obtained from a formula of a Boolean algebra by means of the interchange of the addition and the multiplication and of the elements О and I may sometimes coincide with the original relation. the associative law for addition and the associative law for multiplication. if we take the correct equality (A + В) (В + С) (C + A) = AB + ВС + CA (see Exercise 6 on page 23) and interchange the addition and the multiplication in it we obtain the equality AB + ВС + CA = {A + В) (В + С) (С + A) coinciding with the original equality. For instance.e. The just stated property of the Boolean algebras which allows us to obtain automatically (that is without proof) from any equality a new one1) is called the Principle of Duality and the formulas which are obtained from each other with the aid of this principle are called dual formulas. Examples of dual pairs of laws are the commutative law for addition and the commutative law for multiplication. 38 .of the equality (A + B)(A + / ) + (A + В) (В + 0)= A + В (see Exercise 8 on page 23) implies that the equality {А В + AO) (AB + В I) = AB must also hold. The principle of duality follows from the fact that the list of the basic laws of a Boolean algebra (when proving various Boolean relations we can only proceed from these laws) is completely "symmetric". Similarly. together with every law it also includes another law dual to the former. In this case we deal with a self-dual relation and the application of the duality principle does not give us a new formula. the application of the principle of duality to the correct equality (A + В) (В + С) (С + D) = AC + ВС + BD (see Exercise 7 on page 23) results in the equality AB + ВС + CD = (A + С) (В + С) (В + D) which only slightly differs from the original relation (it simply turns into the original relation if we interchange the letters В and C). that is the law which is obtained from the former law by interchanging the addition and the multiplication and by interchanging simultaneously the elements О and I. i.

A + B)C + A the c o m m u t a t i v i t y of addition = the second distributive law = [(A + B) + A](C + A) the c o m m u t a t i v i t y and the associativ i t y of addition = l(A + A) + B](A + C) = {A + B)(A + C) the idempotent law of addition (cf. under the "bar" operation the element A goes into the original element A. the equalities А Ц. the proof of the equality A (A + С) (B + C) = . this operation (we shall refer to it as the "bar" operation) is such that A + B = AB and ÂB = Â + B Further. Similarly. Example.(AC + ВС) the associativity of addition = A + {A + B)C = the first distributive law = (. That is why when we prove an equality by using some basic laws of a Boolean algebra we can similarly prove the dual equality by using the corresponding dual. finally. An alternative proof of the principle of duality is connected with a special operation defined in Boolean algebras which transforms every element A of a Boolean algebra into a new element A and under which addition and multiplication are interchanged.laws.О = A and A .multiplication. Let us prove the equality A + AC + ВС = (A + В) (A + C) It can readily be seen that this equality is the dual of the relation A {A + С) {В + С) = AB + AC Indeed. this operation possesses the properties 0 = 7 and 1=0 Finally. the first and the second distributive laws are dual and.AB + AC on page 23). A + AC + BC = A -f. In other words.f I = I are dual to the equalities AI = A and AO = О respectively. that is for every element A 39 .

they are also referred to as the De Morgan formulas 40 . 15 By A we mean the so-called complement of the set Л wnich is.of a Boolean have algebra we A = (A) = A In the algebra of sets the "bar" operation (this operation makes it possible to form a new element of a Boolean algebra from one given element of the algebra and not from two given elements as in the case of ddition or multiplication) has the following meaning. It is also evident that 0 = 1 and 1= 0 Finally. 15). The definition of the complément. if we take the set of all the pupils in your class as the universal set and if A is the set of those pupils who got at least one bad mark then A is the set of those pupils who got no bad marks. let us prove that in the set algebra there hold the following highly important properties of the "bar" operation : A+B=AB and ÂB = Â + B The last two relations express the so-called laws of dualization. by definition. A of the set A directly implies that It also follows from the definition that A+ A= I and AA = 0 (see Fig. 15. For instance. the last two equalities expressing the so-called complementation laws can even be taken as the definition of the set A). Fig. the set containing those and only those elements of the universal set I which are not contained in the set A (see Fig.

166 is the complement of the set represented by the whole shaded area in Fig. 16a the area representing the set A is shaded by hatching lines inclined to the left. 16a repre166 repretwo areas that is АВ = 1 + В 41 . 166 the complement A of the set A (with respect to the whole square I) is shaded • it. The first of these relations is also called the De Morgan theorem for union-complement and the second is called tlie De Morgan theorem for intersection-complement. a contemporary and associate of George Hoole. sents the set AB. 16a represents the set A + В while the cross-hatched area in Fig.English mathematician Augustus De Morgan (1806-1871). 166 show the complement В of the set B.Fig.1 ts) are also the complements of each other. the whole shaded area in Fig. which proves the first of the De Morgan formulas: 4+7? = Jfl " " the other hand. The whole shaded area in Fig. in Fig. the vertical lines in Fig. The comparison of Fig. 16a and Fig. 16a. It is evident. the cross-hatched area in Fig.h hatching lines inclined to the right. t h a t these l s. 166 indicates t h a t the cross-hatched area in Fig. sents the set A + B. The horizontal lines in Fig. 166 represents the set A B. In Fig. 16 (or iules) after the . 16a cover the area representing the set B.

For the algebra of four elements (Example 2 on page 27) we put 0=1. For the algebra of two elements (Example 1 on page 25) we put 0= 1 and 1= 0 It is quite evident that for any element a of this algebra (that is for a = 0 and for a = 1) we have a = a. q= p and 1=0 In this case it is also quite clear that a = a for any element a of this algebra. De Morgan's second rule is verified in an analogous way: ab = a + b.which proves the second of the De Morgan formulas. the comparison of the "addition table" and the "multiplication table" compiled for the numbers 0 = 1 and 1 = 0 which have the form + 0 1 (I 1 0 1 1 1 and ô=i 1=0 0=1 1=0 1 0 0 0 shows that a + b = ab in all the cases. As above. p=q. * Now let us discuss the meaning of the "bar" operation for the other examples of Boolean algebras considered earlier. Further. to prove the relation a b = ab it suffices to compare the following two tables: + 0 p <1 l 0 p 4 1 0 p 1 1 p p 1 1 4 1 g l i l l l 0 = 1 p = q q=p 0=1 p=q q=p 1=0 The relation ah = a + manner. 42 1 4 P 0 g q 0 0 P 0 P 0 1= 0 0 0 0 b can be checked in a similar .

1=0 and x= x In this case the De Morgan rules also obviously hold: x@y~x®y and x ® y = x © y see Figs. 0=1.X 01 о * * 1 о * 2 Fig. y] and it is also necessary that the "bar" operation should reverse the order of elements. that is it is necessary t h a t the condition x ^ y should imply x ^ y (why?). obviously. unfortunately. Then. what is said in this connection on page 117). we can put x= 1—x In other words. y] For the De Morgan rules to hold in this algebra we must have _ г © y = x ® y and x <g) y — x © y This means that there must be max [x.x = 1 and xx = 0 do not hold here (cf. we can assume t h a t the points x and x are symmetric about the midpoint 1/2 of the closed interval [0. when the elements of the algebra are all the numbers x satisfying the condition 0 ^ x ^ 1 then. 18a and b). y] and min [x. Finally. 1] (Fig. However. y] = max [x. let us consider the algebra of least common multiples and greatest common divisors whose elements are all 43 . г/] and x <8> у = min [x. \1 17 Now let us consider the algebra of maxima and minima whose elements are numbers x such that 0 ^ x ^ 1 for which the Boolean addition © and the Boolean multiplication <Э are defined as x Ф у = max [x. the rules x +. Therefore. 17). for instance. y] = min [x.

7 = 30. Let us put — N m— — m for this algebra. in the case N = 210 we have 6 ф 21 = 16. 18 the possible divisors of a positive integer N for which the Boolean addition ® and the Boolean multiplication 0 are defined as m. и) where [m. 42 = 5. © n = [m.х© у=х®у 01 О У хфу О * * i О * О У л®у 1/ 01—О л О У fa) х®у=л©у * 1 О У О—II * ( b) Fig. 6 = 35. 30 = 7. 6 ® 2 1 = 3 5 ® 1 0 = = (35. 21 = 10. 211 = 42. 10] = 70 and 3 = 70 . 5 = 42. 10 = 21. 105 = 2. 6 © 21 = 35 © 10 = = [35. n] and m ® n = (те. 210 = 1 It is clear that in the general case of an arbitrary number N we have i = N and N = 1 Besides. 21) = 3. 70 = 3. 2 = 105. it is evident that - N = m m = -rr-— N/rn The De Morgan rules also hold here: rn © n = m 0 n and m 0 n= m © n For instance. 3 = 70. 14 = 15. 10) = 5 and 42 = 5 and also 6 0 21 = (6. 15 = 14. For instance. 35 = 6. in the case N=210 ered earlier we have consid- 1 = 2 1 0 . ri) is their greatest common divisor. n] is the least common multiple of the numbers m and n and (m.

В and С it remains valid if we simply denote the elements A. of the De Morgan rules). by virtue of the equalities 0 = 1 and 7=0 the transformed (dual) equality involves I instead of О and О instead of I.Let the reader prove the De Morgan rules l'or the general case of an arbitrary A' (in this connection also see Exercise 5* on page 52). this yields the equality A + AC + ВС = (A + В) (A + C) which is the dual of the original equality. 45 . in other words. В and С. we have A(A + C) (B + C) = [A(A + C)](B + C) = =-A(A + C) + B + C = I + A + C + BC = Â+ÂC + BC and AB + AC = ÂBÂC = (Â + B)(Â+C) Thus. the equality A (A + С) (В + С) = AB + AC which we have already mentioned. for instance. We see that the principle of duality is a consequence of the properties of the "bar" operation (and first of all. in the passage to the dual equality we must interchange О and / . by virtue of the De Morgan rules. В and С of the Boolean algebra by the letters А. we finally obtain Â + ÂC + BC = (À + B) {Â+ С) Since the last equality is fulfilled for any А . * * \'ow suppose that we have an arbitrary relation holding in any Boolean algebra. It should be however borne in mind that if the original equality involves the "special" elements О and / then. The application of the "bar " operation to both members of this equality results in A{A + C) (B + C) = AB + AC However.

by their 46 .* For instance. It is obvious that if. on applying the "bar" operation to both sides of the equality A (A + I) (B + 0) = AB (see Exercise 5 on page 23) we obtain A(A + I) (B-f-O) = ÂB Now. Namely. for instance. if we replace in the resultant equality the elements А. since A {A +1) (B + 0) = A {A +1) + ~B + 0 =r-A + and Z+7+ + 5 + 0 = 1 + 4 7 + 5 0 = Л + Ж? + £ / lB=A we can also write the + B relation Â=ÂO + BI = Â+B The last relation (note that A and В are arbitrary in it) is equivalent to the relation A + AO + BI = A + В which can be obtained from the original equality by interchanging addition and multiplication and interchanging simultaneously 0 and I. В. С etc. up till now we have spoken of those "Boolean equalities" which only involve the operations of addition and multiplication. a given formula involves an element A then the application of the "bar" operation to both members of the formula results in the transformation of A into the element A = A. * * Now we also note that the proof of the principle of duality we have presented allows us to extend immediately its statement. Finally. It turns out t h a t the relation A = A makes it possible to extend the duality principle to equalities involving the "bar" operation as well.

19). t h a t the relation I D for ele47 . Similarly. to explain this fact we must present one more notion playing an extremely important role in the theory of Boolean algebras. a in the algebra of numbers). In the algebra of sets the relation A zd В means that the set A contains the set В as its part (see Fig. of course. The inclusion relation is denoted by the symbol ZD (orcz). then instead of A we must again write A. similarly. Exercise 2 (d) below) is the evident relation A {A + B) = AB. if A is the set of the pupils in your class who have no bad marks and В is the set of the excellent pupils then. С = С etc. A zd B. В = В. It should also be taken into account t h a t when two sets A and В coincide it is also correct to write A zd В because in this case as well the set В is entirely contained in the set A. We see. It turns out that the principle of duality has even a wider range of application because it applies not only to Boolean equalities but also to "Boolean inequalities". However. For instance.complements A = А. if A 2 is the set of the even numbers and A e is the set of the integers divisible by 6 then obviously A2 zd Ae. for two elements A and В there may exist the inclusion relation A zd В or. For instance. Every Boolean algebra involves the equality relation between elements of the algebra (an equality A = В simply means that A and В are one and the same element of the Boolean algebra) and it also involves one more i m p o r t a n t relation (the inclusion relation) between elements whose role is analogous to that of the relation "greater than" (or "less than") in the algebra of numbers. It follows that when we pass u from a given formula to its dual the bar" operation goes into itself. B e A The last two relations have one and the same meaning (note that the form of these relations resembles that of the relations fif>& and b < . the De Morgan formulas Â+B = AB and ~ÂB = Â+B are the duals of each other and. which is the same. the dual of the equality A + A B = A + В (cf.

then A zd С (see Fig. 21). F u r t h e r . when we deal with numbers the relations a b and b ^ с imply t h a t a ^ c. Finally (this fact is particularly important). Any two (real) numbers a and b are comparable in the sense t h a t at least one of the relations a ^ b and 48 . then A = В For numbers we know the similar fact t h a t the relations a ^ b and b ^ a i m p l y a — b. Now we indicate an essential distinction between these relations. It is clear t h a t if А zd В and В zd C. il A zd В and В zd A. if A zd В then A czB (see Fig. 19 Fig. Up till now we have compared the properties of the relation zd for sets w i t h the properties of the relation ^ for numbers to stress the similarity between (he relations.Fig. 20). Similarly. 20 ments of a Boolean algebra is closer to the relation ^ ("greater than or equal to") t h a n to the relation > ("greater than") used in the algebra of numbers. For instance. since the set of the pupils in your class is wider t h a n the set of the excellent pupils it follows t h a t the set of the pupils having at least one bad m a r k is contained in the set of the pupils who are not excellent pupils.

A (a) S (b) Fig. 19). in the general case. Since A zd A for a n y A the last two equalities can be considered a generalization of the idempotent laws A f A — A and A A — A. 22 I n case b o t h relations hold simultaneously the numbers a and b are s i m p l y equal. 22). Finally. for two arbitrary sets A and В neither of the relations л гэ В and В zd A holds (see Fig.b a must be fulfilled for them 1 ). 23). A ZD О and t h a t there always hold (for any A and В) the inclusion relations /1 + В zd Fig. it is evident t h a t if A zd В then A + В = A and AB = В (see Fig. 21 A and А В cz A (Fig. . In contrast to it. We also note t h a t for any element A of an algebra of sets we have I ZD A.

Let the reader check that the relation ZD defined in the above indicated way in each of the algebras we have considered possesses all the enumerated properties of the relation ZD in the algebra of sets. 1 :z> p . t h a t is neither of the relations p ZD q and q ZD p holds for them). We shall 1 ) I n this Boolean algebra for any two elements x and y at least one of the relations x г:> y and y z: x holds. Finally. /) Z 3 0 and q ZD 0 (the elements p and q of this algebra are incomparable. For the "algebra of two numbers" (Example 1 on page 25) this relation is specified by the condition A+B 1 ZD 0 В For the "algebra of four numbers" (Example 2 on page 27) the relation ZD is specified by the conditions F i g . * Now it appears natural to apply the term "Boolean inequality" to any formula whose left-hand and right-hand members are connected by the relation ZD (or cz). 23 IZDO. 1 Z D с/. in this case we have 42 ZD 6 while the numbers 42 and 35 are incomparable in this algebra (that is neither of the relations 42 ZD 35 and 42 cz 35 takes place). 50 . in the "algebra of least common multiples and greatest common divisors" (Example 4 on page 31) the relation m ZD n means that the number n is a divisor of the number m.>: we assume that two elements x and у of this algebra are connected by the relation x ZD у if the number x is not less than the number у (for instance. we have 1/2 ZD 1/3 and 1 ZD 1 in this case) 1 ). For the "algebra of maxima and minima" (Example 3 on page 28) the relation гз coincides with the relation . for instance.* Let us discuss the meaning of the relation ZD for the other Boolean algebras known to us.

of a Boolean algebra entering into the inequalities in question. such are the inequalities I ^ > A . from the relation (A + В) {A + C){A (see Exercise 8 have + / ) => ABC (b) on page 53) it follows that we always AB + AC + AO с A + В + С * To prove the general principle of duality it suffices to apply the "bar" operation to both members of the original inequality. For instance. I are also interchanged (provided that О or / or both enter into the enequality) then. A ^ > 0 . For instance. The principle of duality states t h a t if addition and multiplication are interchanged in such an inequality and if the elements О and. taking into account that 1 = 0 . A + B ^ > zd A and A zd AB considered above. В. on changing the sign of inequality to the opposite (that is on replacing the relation ID by the relation с or vice versa) we again arrive at a correct inequality (that is at an inequality which is fulfilled for all the values of the elements of the Boolean algebra which enter into it). В and С there holds AB + AC + AO czÂ+B +C . we conclude that for any the inequality +C А. . we obtain RTB)(A + C)(A + I) = (A + B) (A + C) + = A + B + A + C +1f I+T= Г ÂB + AC + AO Similarly. it follows that the inequality (A + B)(A+ C) (A +1) cz ABC is also valid. . For instance. By the De Morgan rules. C. . since the inequality (A В) X X {A — С) (A + I) zd ABC holds and since we have the rule "if A zd В then A cz B.speak of only those inequalities which hold for all the possible values of the elements A. Tbc^I+B rims.

• • . Prove t h a t in this case the "algebra of least common multiples and greatest common divisors" whose elements are the divisors of the number N (see Example 4 on page 31) reduces to the "algebra of the subsets of the universal set / = p2. A + B ) { A + B) = A (b) AB-\-{A (c) ÀÏÏC ÂB AC = О + B){I-\-B) = A+ B (d*) A + AB = A + B 3. We thus arrive at the inequality AB + AC + AO a A + В + С which is dual to the original inequality scribed above. (a) Let N = ргр2 . 5*. .b. 2. Ph a r e pairwise different.. Check that in the "algebra of four numbers" (Example 2 on page 27) there holds De Morgan's second rule ab = a -). . . ph}". Write down the dual equalities for all the equalities whose proof is discussed in Exercises 1-10 on page 23. Prove that in this case the "algebra of least common multiples and greatest common divisors" whose elements are the divisors of the number N reduces to the "algebra of maxima and minima" defined in the set consisting of the numbers 0 .. Show that in this "algebra 52 . 1 . since А. .. Prove the following identities of algebra of the sets: (a) ( . 2. В and С are arbitrary here we can simply denote them as fi and С respectively. ph where all prime numbers Pi. Proceeding from this fact show that in this "algebra of least common multiples and greatest common divisors" all the laws of a Boolean algebra hold including the De Morgan rules..Now. p2. * in the sense de- * Exercises 1. . (b) and (c) by means of the principles of duality? 4. W h a t equalities are obtained from the equalities in Exercises 2 (a). A. . . (b) Let TV = pA where p is a prime number and A is a positive integer.

10. also prove these inequalities directly without resorting to the duality principle. Prove the following inequalities of the set algebra: (a) A + B + C =>(A + B)(A + C) (b) (A + B)(A + C)(A + I)=> ABC (c) (A + B)(B + C)(C + A)=>ABC (d) A + Bz=> ÂB + AB 9. 8. 6*. 0 ^ ah Ah (cf. . / Л and m = pa2'. What is the decomposition of the number m = N/m into prime factors? Use the formula obtained for this decomposition to prove the De Morgan rules in the general case of an arbitrary "algebra of least common multiples and greatest common divisors". (b) the algebra of four elements. Prove that if a Boolean inequality involves the "bar" operation then is also valid the inequality obtained from the original one by interchanging the Boolean addition and the Boolean multiplication and by interchanging simultaneously the element О and the element / while the "bar" pperation is retained at each place it occupies in the original ^ e q u a l i t y and the sign of the inequality is changed to the opposite. (c) Let N = p f ' pi' . / Л where O ^ a ^ ^ j . . . Use this principle to form a new inequality from tne inequality in Exercise 8 (d). For which of the Boolean algebras known to you do the equalities A + I=I and AA = 0 hold and for which do they not hold? 7. 0 < . Check that all the properties of the relation zэ hold for the following algebras: (a) the algebra of two elements (see Example 1 on page 25). 53 . Exercise 6 on page 36). (c) the "algebra of maxima and minima". Write down the inequalities obtained from the inequalities in Exercises 8 (a)-(c) using the duality principle.of least common multiples and greatest common divisors" all the laws of a Boolean algebra hold including the De Morgan rules. . (d) the "algebra of least common multiples and greatest common divisors".

Let some sets A and В be such t h a t A zd В. John. X . . . the sets we have mentioned can be written as A = {Peter. 4. However. 3. it becomes completely inapplicable when the set in question is infinite (we cannot enumerate an infinite number of Llie elements of the set!). 12*. Let us discuss the methods for specifying the sets which are the elements of this algebra. t h a t is by enumerating all the elements of the set. even in those cases when a set can be defined by tabulation and the tabulation is quite simple it may nevertheless happen that the enumeration itself does not indicate why these elements are collected toj form the set. 3. For instance. 2. subtraction.. Besides. Mary} В = {1. (b) AB. :} (in the last expression the signs of the operations symbolize the operations themselves) 1 ). 5} and C= {+. (c) A + B\ (d) AB 4. for instance. Tom. Verify all the properties of the relation ZD for (a) the "algebra of maxima and minima". Propositional Algebra Let us come back to the Boolean algebra of sets which plays the most important role in the present book.11. In mathematics the elements of a set which is defined by tabulation are usually written in curly brackets. It is obvious t h a t the simplest method is to specify a given set by tabulation. Tom and Mary" or the "set of the numbers: 1. John. 4" or the "set of the four operations of arithmetic: addition. Sets and Propositions. we can consider the "set of the pupils: Peter. (b) the "algebra of least common multiples and greatest common divisors". multiplication and division". 2. this method of representing a set is highly inconvenient in case there are very many elements in the set. Simplify the following expressions: (a) A + B. x 54 ) Also see Exercise 5 (a) on page 52.

etc. 24) 2 ). such a proposition m a y simply reduce to the enumeration of the elements of a given set. "he is a chess-player". if we are interested in the sets whose elements are some (or all) pupils in your class then such propositions can be "he is an excellent pupil". the collection of the pupils in your class or the collection of the integral numbers) and then state a proposition which is true for all the elements of a set under consideration and only for these elements. The set A of all those elements of the universal set I in question (for instance. or Tom.Therefore another method which specifies the sets implicitly by description is more widely used. "his name is George" and the like. see Fig. there is a "two-way connection" between sets and propositions: every set is described by a proposition (in particular. The descriptive method of representing sets connects the sets with propositions which are studied in m a t h e m a t i c a l logic. for instance.) which satisfy the condition mentioned as the characteristic property in a given proposition a is called the truth set of this proposition 1 ) (for instance. Namely. the set of the numbers. the essence of the method is t h a t we fix a collection of the objects we are interested in (for instance. moreover. "the name of the pupil is Peter or John. 55 . 2 ) We shall denote propositions by small letters and the t r u t h sets corresponding to the propositions by the same capital letters. we can consider the "set of all excellent pupils in your class" (it m a y turn out that the set A mentioned above coincides with this set of excellent pupils) or the "set of all integers x such t h a t 0 ^ x ^ 5" (this set exactly coincides with the set В mentioned above) or the "set of all animals in a zoo". For instance. Thus. infinite sets can be defined by description only. the set of the pupils. For instance. The descriptive method for specifying sets is quite applicable for the definition of infinite sets such as the "set of all integers" or the "set of all triangles with area equal to 1". When a set is defined by description we indicate a property characterizing all the elements of the set. — Tr. i t would be more precise to use the term prepositional function or open sentence (or open statement) and to speak of the t r u t h set of the given propositional function but we shall s i m p l y speak of propositions throughout the present translation. as has been mentioned. or Mary") x ) According to the terminology of modern mathematical logic.

24 and to every proposition there corresponds a definite set which is the t r u t h set of this proposition. W h e n two given propositions a and b (for instance. Another highly i m p o r t a n t condition is t h a t by a proposition we shall only mean a statement about which it makes sense to say that it is either true or false when it is applied to a definite element of the given universal set. t h a t such statements as "the person has two heads and sixteen arms" or " 2 x 3 = 6" are propositions (the second of these sentences is even completely independent of the choice of the universal set I ) while such sentences as the exclamation "Be careful!" or "Oh!" are not considered propositions. "he is an excellent pupil" and "he has only the highest marks" or "the number x is odd" and "the division of the number x by 2 gives 1 in the remainder") are equivalent we shall write a = b 56 . it should also be borne in mind t h a t such sentences like "two hours is a long time" or "the e x a m i n a t i o n in m a t h e matics is a highly unpleasant procedure" are not considered propositions either because t h e y are quite subjective and t h e i r t r u t h or f a l s i t y depends on a number of circumstances and on the character of the person who states these sentences. Therefore any two propositions a and b to which one and the same t r u t h set corresponds will be identified and will be considered equivalent ("equal"). This means. It is also import a n t t h a t for any collection of propositions (even for propositions concerning objects of different kinds) it is always possible to indicate a certain universal set / corresponding to all the propositions in question and c o n t a i n i n g all the objects mentioned in these propositions. Finally. for instance. W h e n considering propositions we are only interested in the sets t h e y describe.a-"the figure is quadrangular" b-"the figure is triangular" Fig.

Alice} then a -f b is the proposition "the pupil can play chess or 'iie pupil can play draughts" (or. the notation A U В for the sum of two sets A and B). that is the propositions which are always true irrespective of which element of the set I is considered. We shall denote such propositions by the letter o.All the necessarily true propositions. by the sum of two propositions a and b we shall mean a proposition whose truth set coincides with the sum of the truth set A of the proposition a and the truth set В of the proposition b. Harry. Mary. 14 57 . the sum a + b of two propositions a and b is simply the proposition "a or b" where the word "or" means that at least one of the propositions a and b (or both propositions) is true. John. Namely. Let us agree to denote all necessarily true propositions by the letter i. are also regarded as equivalent to one another. Mary. Helen} while the proposition b asserts that "the pupil can play draughts" and its truth set is В = {Peter. Tom. For instance. will also be regarded as equivalent. "the height of the pupil exceeds 4m" and "the number x is greater than 3 and less than 2". Similarly. Since the sum of two sets is nothing but the union of all the elements contained in both sets. briefly. "this pupil can fly like a bird". "this pupil is a boy or a girl". Examples of necessarily true propositions are " 2 x 3 = 6". "the pupil can l ) In mathematical logic the sum of two propositions a and b usually called the disjunction of these propositions and is denoted by the symbol a V b (cf. The connection between the sets and the propositions makes it possible to define some algebraic operations on propositions similar to those introduced earlier for the algebra of sets. "the height of the pupil does not exceed 3m" and the like.f b1). Bob. Tom. all the necessarily false (that is contradictory) propositions which are never true. examples of necessarily false propositions are " 2 x 2 = 6". George. Ann. that is the propositions whose truth sets are empty. Let us agree to denote this new proposition by the symbol a . if the proposition a states "the pupil is a chess-player" and if among the pupils in your class the truth set corresponding to this proposition is A = {Peter.

25). briefly. George. Similarly.b is A + В — {Peter. Mary. then the proposition ab asserts t h a t "the pupil can play chess and the pupil can play draughts" (or. Alice} If the universal set is the set of the geometrical figures shown in Fig. Tom. the n o t a t i o n A f| В for the p r o d u c t of t w o s e t s A and B). "the pupil can play chess and draughts"). H a r r y . The truth set corresponding to this proposition а -f. 24 and if the propositions с and d assert t h a t "the figure is round" and "the figure is shaded" respectively then the proposition с + d asserts that "the figure is round or shaded" (see Fig. the product ab of the propositions a and b is the proposition "a and b" where the word "and" means t h a t both propositions a and b are true. Ann.с-"the figure is round" d-"the figure is shaded' C+D- c + d-"the figure is round OR shaded" cd-"the round AND figure is shaded" Pig. Helen. . if the propositions a and b concerning the pupils in your class are the same as above. the truth set corresponding to this I n m a t h e m a t i c a l logic the p r o d u c t of t w o p r o p o s i t i o n s a and b is more often c a l l e d the conjunction of these p r o p o s i t i o n s and is denoted b y the s y m b o l a/\b (cf. by the product ab of two propositions a and b with truth sets A and В we shall mean a proposition whose truth set coincides with the product AB of the sets A and 5 1 ). Since the product of two sets A and В is nothing b u t their intersection (that is their common part) containing those and only those elements of the universal set I which are contained in both sets A and B. For instance. John. 25 play chess or draughts"). Bob.

The connection between the sets and the propositions makes it possible to extend to the propositions all the rules of the algebra of sets: a b = b + a and ab = ba t h e c o m m u t a t i v e l a w s of a l g e b r a of p r o p o s i t i o n s (a + 6) + c = a + (& + e) and (ab)c-^a(bc) t h e a s s o c i a t i v e l a w s of a l g e b r a of p r o p o s i t i o n s (a-f b) с — ac-f be and ab-{-с = ( a c ) (b-{. 24 mean that "the figure is round" and "the figure is shaded" respectively. in M W o s h a l l also w r i t e t h e r u l e s we h a v e e n u m e r a t e d i n t h e w h i c h t h e y are u s u a l l y g i v e n i n m a t h e m a t i c a l l o g i c : a\/b = (a\/h)\Jc = (a\/b)/\c a V" : a Vо ~ a\/i = b\J a a\/(b\/c) (aAc)V(b/\c) " a i form a/\b = b Д a (аДЬ)Дс = аД(6Дс) (аЛЬ)\/с = (aVc)/\(b\Jc) аДa . the proposition "the pupil has only the highest marks or the pupil has two heads" is equivalent to the proposition "the pupil has only highest marks" while the proposition "the pupil can swim and the pupil is not yet 200 years old" is equivalent to the proposition "the pupil can swim" 1 ). Mary} If two propositions с and d concerning the set of the geometrical figures shown in Fig.c) t h e d i s t r i b u t i v e l a w s of a l g e b r a of p r o p o s i t i o n s a+ a= a and aa = a t h e i d e m p o t e n t l a w s of a l g e b r a of p r o p o s i t i o n s Besides.о For instance.a аД i .proposition is AB . Tom.{Peter. ai = a and a + i = i. if i is a necessarily true proposition and о is a necessarily false proposition then we always have (that is for any proposition a) the relations a о — a. 25).a a/\o = о 59 . ao ----. then the proposition cd asserts t h a t "the figure is round and shaded" (see Fig.

the truth set of the proposition à contains those and only those elements of the universal set I which are not contained in the set A. By virtue of the second distributive law of the set algebra. For instance. for instance. В and С are the truth sets of the propositions a. Namely. In other words. that is the elements which are not contained in the truth set of the proposition a.* To demonstrate how the rules ol the algebra of propositions are derived from the rules of the algebra of sets let us consider. the "bar" operation of the algebra of sets can also be extended to the algebra of propositions. it is evident that the truth set of the compound proposition ab + с which means "the proposition "a and b" or the proposition с is t r u e " is the set AB + С where А.) * * Like the operations of addition and multiplication of sets. we have AB + С = (A + С) {В + C) Thus. which means that the propositions ab + с and (a + c) (b-\-c) are equivalent. Similarly. "the pupil can play draughts" and "the pupil can swim" respectively. b and с respectively. Since the truth set of the sum of two propositions is the union of the truth sets of these propositions and since the truth set of the product of two propositions is the intersection of the truth sets of the given propositions. b and с are "the pupil can play chess". if the . by a should be meant the proposition whose truth set is the set A where A is the truth set of the proposition a. the truth set of the (compound) proposition (a + c) (b + c) is the set (A + С) {B + C). the truth sets of the propositions ab + с and (a + с) X X (b + c) coincide. (Also see page 20 where we indicated that the propositions "the pupil can play chess and draughts or can swim" and "the pupil can play chess or can swim and also can play draughts or can swim" have one and the same sense. that is ab + с — (a + c) (b + c) where the propositions a. the derivation of the second distributive law.

Generally. the proposition à has the sense "not a". see Fig. there is no need to verify them since they simply follow from the corresponding rules of the set algebra 1 ). since the t r u t h sets of the p r o p o s i t i o n s a + b and аь are А + В and А В where A a n d В are the t r u t h sets of the p r o positions a and b r e s p e c t i v e l y and since . according t o the d e f i n i t i o n of the equivalence ( e q u a l i t y ) of p r o p o s i o n s . we have. " i t is false that 2 X 2 is equal to 5" or "it is false that the pupil has two heads") is always a necessarily true proposition while the negation of a necessarily true proposition (for instance. the "bar" operation of the propositional algebra is the operation of forming the negation (denial) à of the proposition a. The proposition a can be formed from a by prefixing "it is false that".b = ab. If the universal set / consists of the geometrical figures shown in Fig. 26). "it is false that the pupil is not yet 120 years old") is always necessarily false. simply. All the other laws can also be readily checked (let the reader check them). f'y the way.b "IT IS FALSE THAT the figure is triangular" Fig. hence. 24 and the proposition b asserts that "the figure is triangular" then the proposition b means "it is false that the figure is triangular" (that is.4 + H = A B . the e q u a l i t y ч -j. 61 . 26 proposition a asserts that "the pupil has bad marks" then the proposition a means "the pupil has no bad marks". "the figure is not triangular". Now let us enumerate the rules of the algebra of propositions r e h t e d to the operation of forming negation: a— a a-\-a — i o= i and and aa = o i — o and аЬ — а-\-Ъ Indeed. a-\-b = ab ' ) For instance. the negation of a necessarily false proposition (for instance.

(c) ab. (c) "an elephant. b "he is dark" and let с mean "he can swim". (b) ab. State the propositions (a) a + b and ab and (b) ab 62 and a+ b .be and (b) ab + с and (a + c) (b + c) 5. (h) ab 4. is an insect". (g) ab. Explain the meaning of the following propositions: (a) a + b.Exercises 1. W h a t is the meaning of the proposition a in all these cases? Is the proposition à necessarily true? Is it necessarily false? 3. 2. Let the proposition a assert t h a t : (a) " 2 X 2 = = 4". Let a and b be the propositions "this pupil is interested in mathematics" and "this pupil sings well" respectively. (f) ab. Let the proposition a mean "the pupil can play chess" and let the proposition b be "the pupil can play draughts". State the following propositions: (a) ab. (b) a + 6. (d) "he can fly". (d) a-\~T>. (b) "the pupil is a boy". Give three examples of necessarily true propositions and two examples of necessarily false propositions. (c) a + b. Explain the meaning of the propositions (a) (a + b) с and ac -f. Let a be the proposition "he is an excellent pupil". (e) a + b W h a t are the t r u t h sets of these propositions? 6. Let the propositions a and b mean "the given positive integer is even" and "the given positive integer is a prime number" respectively. (e) a+Ъ. (d) ab.

t r a n s l a t i n g books from one language into another. "Laws of Thought". in everyday life few people t h i n k of these rules as m a t h e m a t i c a l laws of thought b u t even children freely use them. t h a t is all the people know (although they m a y not be aware of it) t h a t propositions ab and ba have the same meaning. We can also explain why nowadays George Boole's approach to the m a t h e m a t i c a l i n t e r p r e t a t i o n of the laws of logic as certain "algebraic rules" has become an object of intense interest. Indeed. which must be followed by the thinking" machines constructed by the man. as is known. economic p l a n n i n g and finding necessary data in scientific literature. Rules for Deduction Now we can explain why George Boole called his work (where the "unusual algebra" considered in the present book was constructed) "Laws of Thought". for instance. As long as logical operations were performed only by people who used them in the process of thinking quite i n t u i t i v e l y there was no need to formulate the logical rules rigorously. solving mathematical problems. the "bar" operation has the sense of the negation and the laws of the propositional algebra describe the basic rules for the logical operations which all the people follow in the process of thinking. t h a t the "laws of thought". which is the same. nobody doubts t h a t to say "he is a good runner and a good jumper" is just the saine as to say "he is a good jumper and a good runner". 4 reduce to nothing but the logical (propositional) connectives "or" and "and" respectively. are equivalent. In recent decades the situation lias changed and nowadays we want to make electronic computers perform the f u n c t i o n s which in the past could only be performed by people.5. modern electronic computers can even play chess! It is obvious that in order ! o compile the necessary programs for the computers it necessary to state rigorously the "rules of the game". traffic schedulling. Of course. People can -ч1 the rules of logic i n t u i t i v e l y but for a computer these les must be M a t e d in a c l e a r manner usina the only 03 . such functions as production process control. The m a t t e r is t h a t tlie algebra of propositions is closely related to the rules of the process of t h i n k i n g because the sum and the product of propositions defined in Sec. or.

64 . (The law of excluded middle and the law of contradiction are referred to as the laws of complement or the complementation laws. Now let us come back to the "laws of thought" themselves. The rule expressed by the relation aa = o is called the law (principle) of contradiction. if a pupil has no bad marks then the proposition "the pupil has bad marks" is of course false when applied to this pupil. this law asserts t h a t the propositions a and 'a. is always true. The most interesting logical rules are connected with the logical operation of negation. The rule expressed by the formula a= a is called the law of double negation (or the law of double denial). that is the language of mathematics 1 ). It asserts t h a t the double negation of a propoWe must warn the reader that the elementary algebra of propositions to w h i c h the present book is devoted does not provide sufficient means for constructing modern electronic computers and for posing complex m a t h e m a t i c a l problems in the f o r m i n w h i c h they can be "inserted"' i n t o computers. t h a t is "a or not a".) For instance. For this purpose a more i n t r i c a t e mathem a t i c a l and logical apparatus must be developed which is not considered here. that is a and "not a". For instance. if a whole number n is even then the proposition "the number n is odd" is of course false for this number. can never be true simultaneously and therefore the product of these propositions is always false. For instance. even w i t h o u t h a v i n g any information 011 the tallest pupil in your school we can definitely assert t h a t this pupil is "either an excellent pupil or not an excellent pupil" or t h a t this pupil "either can play chess or cannot play chess" etc. the rule a+ a= i expresses the so-called law (principle) of excluded middle: it means t h a t either a is true (here a is an a r b i t r a r y proposition) or the proposition à is true and therefore the proposition a-f a."language" which a m a t h e m a t i c a l machine can"»nderstand". some of t h e m have special names in logic.

the verbal s t a t e m e n t of lliese rules is a little more complicated (in S h is connection see Exercise 1 below). which is the same. the "two-way connection" between sets and propositions allows us to extend easily the relation из of the Igebra of sets (the inclusion relation) to the algebra of .'impositions. that is logical rules which I he people follow when deducing new inferences from those which are already known to be true. the proposition a is a conse• ' nee of the proposition b: the meaning of what has been 'I is thai the truth set A of the proposition. However. All the other rules of the proposilional algebra such as the d i s t r i b u t i v e laws (a + /)) с — ас -f be and or the idempotent. Namely. Similarly. as In the negation of 1he latter proposition. Л particularly i m p o r t a n t role is played by the r e l a t i o n s which can be extended from the algebra of sets to the mathematical logic (prepositional calculus). a ab + с — (a + c) (b + c) laws a —a and aa — a ne definite "laws of thought". Up till now we have not considered Ibis relation in connection with propositions nul only discussed it in connection with the algebra of sets. the negation of the proposition "the given integer is even" is the proposition "the given integer is odd".sition is equivalent to the original proposition itself. For instance. a contains the 65 . The De Morgan rules a-\-b — ab ab-=a-\-b and for the propositions are also very i m p o r t a n t . let us agree t h a t the relation a n> b written for two propositions a and b should be understood n the sense that the proposition a follows from the propo<iion b or. it asserts t h a t "the given integer is not odd" and hence it is equivalent to the original proposition asserting t h a t the integer is even. the negation of the proposition "the pupil has no bad marks" means thai "the pupil has bad^marks" and the double negation of the former proposition slates that "it is falsi' llial the pupil has bad marks" and is therefore equivalent to the original proposition asserting t h a t the pupil has no bad marks.

.truth set В of the proposition b. . These properties can be stated as the following rules for I f a zd b we also say that the proposition 6 is a sufficient condition for the proposition a (for instance. the above relation a zd b means t h a t M \ А гэ В \ For instance. In the deduction processes (for instance. 27 no bad marks" is a consequence of the proposition b asserting t h a t "the pupil has only the highest marks". 18. 18. 10. the set \ Ae - {G. see Fig. 66 . 20. W h e n showing t h a t the relation a ZD b takes place we deduce the conclusion a from the condition b. 27) follows a conclusion a of the theorem (for instance. 8. 16. A2 = {2. for a p u p i l to have only the highest marks i t is of course necessary that this p u p i l should have no bad marks). for instance. the conclusion t h a t " M P 2 -f N P 2 = = MN2". In o t h e r words. 12. in this case the relation а b is equivalent to the Pythagoras theorem). since the set В of excellent pupils in your class is obviously contained in the \ JJ set A of all pupils h a v i n g no bad marks. in the proof of a theorem) we always use the basic properties of the relation (sometimes without being aware of it). for a p u p i l to have no bad marks i t is of course sufficient that this p u p i l should have only the highest marks) while the proposition a is said to be a necessary condition for the proposition b (for instance. The process of establishing the fact t h a t two propositions a and b are connected by the relation a I D b is called deduction. . the proposition a staling t h a t "the pupil has Fig. 12. the proof of a m a t h e m a t i c a l theorem reduces to deduction. 4. as a rule.} of the even whole numbers. 6. Similarly. from the condition t h a t "the angle P of a triangle MNP is a right angle". . therefore the proposition "the number is even" (it is meant t h a t the universal set I we deal with consists of all the positive integers) follows from the proposition "the number is divisible by 6" 1 ).} of the whole numbers divisible by 6 is contained in the set. . In a m a t h e m a t i c a l proof it is usually required to show t h a t from a condition b of the theorem (for instance. . 14. In everyday life and in science we often deal with deduction.

that is to prove that the proposition A follows from the proposition b.'/ a ZD b and b ZD a then a = b" is sometimes stated as "if b is a necessary and sufficient condition for a then the propositions a and b are equivalent" (from this ['oint of view we have assumed the propositions a and b are considered ']nal in this case). Let it be required to prove that if an integer n which is greater than 3 is a prime number (the proposition b) then n has the form 6/c ± 1 (where к is an integer). we know that "in a parallelogram the opposite angles are congruent" (the proposition с). 67 . that is to show that the proposition "not b" (that is the proposition b) follows from the proposition "not a" (that is from the proposition a). that is a = b. -) In this case we even have a ZD b and b ZD a. Let us dwell in more detail on the application of the rule asserting that if a ZD b then b ZD a. that is when The second of these rules asserting that ". It often turns out that it is easier to prove the fact that if a is false then b cannot be true. This rule serves as the basis for the so-called proofs by contradiction (Latin reductio ad absurdum proofs). Let it be required to prove that the relation A ZD b holds. we know that if the diagonals of a quadrilateral bisect each other (the proposition b) the quadrilateral is a parallelogram (the proposition a). Thus. On the other hand. Let us consider an example of a proof by contradiction.deduction1): a zd a if a zd b and b zd a then a - if a zd b and and b zd С then a zd с i b zd a zd a and a zd о a ZD ah for any for any B a a and b if a zd h then 6 :r> a For instance. we have a zd b2) and czd a and therefore с ZD b In other words. if the diagonals of a quadrilateral bisect each other then its opposite angles are congruent.

therefore we shall try to resort to a proof by contradiction. 3.n is divided by 6 the remainder is + 1 or —1 (the proposition a). Give the verbal statement of the De Morgan rules of the algebra of propositions: a -f b — ab and ab --. since by v i r t u e of the l a w of double negation we have a b . if an integer n > 3 gives 2 or 4 in the remainder when it is divided by 6. then this integer is even and therefore it cannot be a prime number. 2. b follows from a (or. it follows that the assumption that и is true means that when n is divided by 6 we obtain 0 in the remainder (that is n is divisible by 6) or 2. An integer divisible by 6 cannot be a prime number. Give examples demonstrating the following rules: (a) the law of excluded middle. 1 ) The f o l l o w i n g argument is more precise: from the relation b ZD a we have proved i t follows t h a t (a) ZD (ft). then n is divisible by 3 and cannot be a prime number either. or 4.a + b. 5. or to 5 (the last case is equivalent to the one when the remainder is equal to —1). a and . or to 3. or to 4. To this end let us suppose that the proposition à is true. Simplify the expressions of the sum a + b and of the product ab of the two propositions a and b. or to 2. It is rather difficult to prove this fact directly without using the rule if a ZD b then b ZD a. (c) the law of double negation. there must 08 be a ZD b. Exercises 1. (b) the law of contradiction. Thus. b ZD a) whence we conclude that a ZD b which is what we intended to prove 1 ). or 3. symbolically. Give an example of a proof by contradiction and write it in the symbolic form. Since the remainder obtained when an arbitrary integer n is divided by 6 can be equal to 0 (in this case n is divisible by 6) or to 1. if the division of n by 6 gives 3 in the remainder.b. 4. that is the number n (which is an integer greater than 3) cannot be represented in the form 6 f r ± l . Let a ZD b. t h a t is a ZD b\ now. Give one example to demonstrate each of the properties of the relationiD for propositions enumerated on page 67.

a mother (M). Implication The rules of propositional algebra can be applied to the solution of logical problems whose conditions form a collection of propositions using which we must establish the truth or the falsity of some other propositions. it is also known that if the son goes swimming his sister D t goes with him. Dx and D 2 respectively. "the mother went swimming on Sunday morning". a son {S) and two daughters (Z^ and D2) spends its vacation at the sea shore. + S) (MDZ i. it is known that last Sunday only one of the daughters went swimming.t л here the letter t denotes a necessarily true proposition. A family consisting of a father (F).DZ f D. As usual. The second daughter J92 g ° e s swimming then and only then when her mother does and it is known that at least one of the parents goes swimming every morning. the negations of these propositions will be denoted by the same symbols supplied with the bar. Further Examples of Application of Rules for Deduction. They often swim early in the morning and it is known that when the father goes swimming then the mother and the son always go swimming together with him. "the son went swimming on Sunday morning". 'the first daughter went swimming on Sunday morning" and "the second daughter went swimming on Sunday morning" by the same symbols F. The question is: who of the members of the family went swimming last Sunday morning? Let us denote the propositions "the father went swimming on Sunday morning". 'u multiplying all these equalities we obtain the relation *\rS+T) (SD.MD2) (F + M) (D.Я 2 ) i 69 . S. In the symbolic form the conditions of the problem are written thus: (1) FMS + F = i (2) SDt + S=i (3) MD2]MD2 (4) F + M = i i (5) DiDz + DtDz . M. Finally. Below is an example of this kind.6.

i which means that only the mother M and the second daughter D 2 went swimming on Sunday morning. we finally obtain FM'SD. pn with the aid of the basic operations + .. Let us open the parentheses in the expression on the lefthand side using the first distributive law and also the commutative. * The solution of the problem we have presented is based on the algebraic transformations by means of which we have simplified a rather complex expression {FMS + F) (SDi + S) (MD2 + MD2) (F + M) (ВД + 5 Д ) and have brought it to the form FMSDJ)2. This procedure turns out to be useful in many other cases and therefore we shall dwell on it in more detail. + S)~- I FMDJ)2 FMSDiD2 Thus. We shall prove that if the compound (composite) proposition f is not necessarily false (that is / is not equivalent to o) then it can be reduced to the form i 70 114' •• • n . the associative. pn) composed of propositions pu p2. • and the "bar" operation of the algebra of propositions.which is equivalent to the system of equalities (l)-(5) because the product of propositions is true if and only if all the multiplicand propositions are true. I D{D2) .О together with the relations A + + О — A and AO — 0. . .FMSDtD2 4 ÏMDJJ2) (SI). To simplify the transformation we change the order of the factors: (FM S + F) (F + M) = F MS -f FM {F M S + FM) (MDZ + MD2) = FMSD2 + FMDZ (FMSD2 (FMSDiD2 f FMD2) (ОД. . . . .. Suppose that we are given an arbitrary algebraic expression / = / (р1ч p2. the idempotent laws and the law of contradiction AA .D2 .

pn) so that the sign of negation (the bar) stands only above some (or all) constituent (prime) propositions px. First of all. see page 22). Further. 71 . Further. . using the first distributive law we can open the parentheses in all those cases when they mean that to obtain the expression / it is necessary to multiply by one another sums of prime propositions p j and their negations p j or some more complex combinations of the propositions. if a term A of the sum we have obtained contains. n) in each term of the sum denotes either pj or pj and all the terms of the sum are pairwise different. / ) „ or their negations. . 2. . . Besides. p2. In this way we can bring the sum / to the form in which all the summands contain as factors all the propositions p^ p2. . if two composite propositions f j and /2 are equal (equivalent) they have the same forms of type (*) and if they are different (not equal to each other) their forms of type (*) are also different. pn but not above their combinations (their sums or products). . using the De Morgan formulas we can transform the composite proposition f (pu p2.. The proof of the assertion we have stated is quite simple. . Form (*) of a composite proposition / is called its additive normal form1). . .. neither the proposition p t nor its negation p x we can replace it by the equivalent expression A{Pi + Pt)=--Api + Api which is a sum of two terms each of which contains the factor pi or pi.. .inhere the symbol p] (/ = 1. for instance. If all these operations result 1 ) In mathematical logic form (*) of a composite proposition is more often referred to as its disjunctive normal form. . . if a term in the sum contains one and the same proposition ph (or pk) several times as a factor then we can retain only one such factor and if the sum / contains several identical terms we can also retain only one of them (we remind the reader that the Boolean algebra is an "algebra without exponents and coefficients". If a term in this sum contains both factors p j and p j it can be simply dropped (because pp = о for any /?). On opening all the parentheses of this kind we reduce the compound proposition / to an "additive" form written as a sum of a number of terms each of which is a product of prime propositions p} and their negations. .

Wednesday and Friday. because if. r. if forms (*) of two propositions are different these propositions cannot be equal (equivalent). etc. 72 . as to the proposition / 2 . For if. for instance. and let there be not more than three lessons every educational day. p3 and /?4 are true while the proposition p 2 is false. On the other hand. it cannot be true in the latter case. q. one lesson in chemistry. two lessons in physics.4 and form (*) of a proposition fa involves the summand PiPlPsPi while form (*) of a proposition / 2 does not involve such term the proposition Д is true if the propositions p.in the disappearance of all the terms in the sum / we shall have / о If otherwise. One example of this kind was considered above. one lesson in history and one lesson in English. n . Obviously. Monday. It is required that during a week the pupils should have ihree lessons in mathematics. Let us consider a simplified curriculum in which there are only three educational days a week.. the given composite proposition / is reduced to its additive (disjunctive) normal form (*). It is also required that the time-table should satisfy the following conditions. it is evident that if two compound propositions /1 and / 2 reduce to one and the same form (*) they must be equal. namely. for instance. Finally. a proposition p is contained in the expression of fa but is not contained in the expression of / 2 we can write / 2 in the form Î2(P + P) which involves the prime proposition p. Below is one more problem whose mathematical content is close to that of the "swimming problem" solved above. It also turns out that the additive (disjunctive) normal form of a compound proposition is very useful in many other cases. we can always assume that any two given composite propositions fa and / 2 contain the same prime propositions p. The last property can be used for verifying whether two composite propositions fa and / 2 are equivalent or different. We have thus proved that to every proposition / there corresponds its uniquely determined additive normal form.

(7) In this curriculum there are only 3 + 2 + 1 + 1 + 4 . he wants to have the first lesson on Friday. Chj. = л / з л л й ы л м г * ./: 73 . he does not want his lesson to precede the English lesson. then to solve the problem we must stablish the truth or the falsity of the 54 propositions 11 j. on the contrary. that the /th lesson is devoted to mathematics. English or is the time of cisure.Ph9 (Pht Ь Phj + Ph-7) PhriPhHl>h(i ^ i (3) и . . physics. (6) It is naturally required that every educational day Ihe pupils should have not more than one lesson in every subject. (5) The English teacher insists that his lesson should be the last and. besides.! . besides. (3) The history teacher can only teach on Monday and un Wednesday. on Wednesday his lesson must not he the first while. respectively. How can the time-table satisfying all these conditions he worked out? Let us index the 9 possible lessons in succession by the numbers from 1 to 9. 2.(1) The mathematics teacher insists that his lessons should never he the last and that at least twice a week they should he the first. The conditions of the problem can now be written 4 the following system of relations: (1) /. 9) meaning. Phj. besides.Ph?Phe. i M f 7 + (2) / 2 . (2) The physics teacher does riot want his lessons to he the last either.Hj. . he cannot teach on Friday. The last requirement to be satisfied is hat the pupils should have time of leisure either instead of the last lesson on Friday or instead of the first lesson on Monday. history.1 = 8 lessons a week while the total number of possible •essons is 3 X 3 = 9 and hence once a week the pupils have only two lessons. chemistry. . he wants to have the first or the second lesson on Monday or the second lesson on Wednesday..Ej andLy(where / = 1. at least once a week he wants his lesson to he the first.. (4) The chemistry teacher insists that his lessons should not be on Friday and that the day he has his lesson the pupils should have no lesson in physics./•: // .( Я 4 + F f z + H 5 ) ( / / j //.

Pi. Friday: physics. + Ph7MHL9 .. (1) Monday. history. physics. 4. time of leisure. . Pn) can also be brought to its multiplicative normal form1) / = 11 (р[+р'г+ .. chemistry. • • -. Wednesday: mathematics. 74 proposition . The proof of this assertion is quite analogous to the argument which we used in proving that every compound proposition f 1 ) In mathematical logic form (**) of a composite is usually referred to as its conjunctive normal form. English-. history. Ch7ChHCh9 (Ph. mathematics.+Pn) (**) where the symbols p] have the same sense as in formula (*) and all the terms in the product are pairwise different.. Friday: physics. there are only two ways of working out a time-table satisfying all the requirements stated'. English-.Ch2-\-Ph. . . time of leisure". this form is also uniquely determined for a given proposition and characterizes completely the whole class of the propositions equal (equivalent) to that proposition.(4) /4 . .+MaM9) X Ph8Ph9) = i of relations is equivalent to one equality / = /1/2/3/4/5/6/7 - * The additive normal form of the proposition / is f - M\H2Ch9M^РкъЕePh7M8L9 + M-. physics. it should be noted that every compound proposition / = f (Pi.Ph^E^M . or (2) Monday: mathematics. mathematics. mathematics.+ (7) f 7 = L t + LB = i This system Ë7ËsË9 = i .Ch3 + Ph2Chi + Ph2Ch3+ + • • • 4. where the dots symbolize 54 — 9 = 45 multiplicands entering into each of the two terms of the sum / with the sign of negation (the bar).. Likewise. Wednesday: mathematics.PKCh-. .Ph9Chs) = i (5) /5 E3 + E6 + E9 + E2L3 + E5Le + ESL9) (6) fe = (MtM2 + MiM3+M2M3+MiMs+ X (PhiPhi -f PhtPh3 + ... Hence. . chemistry.

Z>2 M and M ZD D2 Therefore M = D2 and. the above argument can be stated thus: "by the conditions of the problem. for instance. For instance. since only one of the daughters went swimming that morning the father could not go swimming" and so on. Let us discuss the instructive comparison of the solution of the problem presented above with a solution which can be given by a pupil who is not familiar with elements of mathematical logic. * * Let us come back to the "swimming problem". the difference between the proofs is very slight.can be brought to form (*). Since only one of these propositions is true we conclude that the proposition F is true" and so on. Thus. d2=>F while from Dt ZD S and S zd F it follows that £>! => F Thus. Such a pupil would replace equalities (l)-(5) and their formal transformations presented above by a "non-formalized" argument (that is an argument which is based on "common sense" instead of the laws of logic stated in a rigorous way) such as the following: "if the father went swimming on Sunday morning then the mother and the son would go with him. Dx ZD S. the "formalization" of ordinary inferences which was demonstrated in the solution of the problem we presented above reduces simply to the exact enumeration of all the conditions used 75 . consequently. instead of the first distributive law we must use the second one. however. we have M ZD F and S zd F Besides. It is however quite evident that an argument of this kind is in fact also based on the rigorous laws of propositional algebra and that the so-called "common sense" exactly follows these laws. from proposition F follow propositions Dx and D2. but the first daughter would follow the son and the second daughter would go together with her mother.

lessons.. etc. if the propositions q and p are "Peter is an excellent pupiF and "an elephant is an insect". then p" or. "q implies For instance. the proposition q => p reads: "if q. then "the proposition q => p means: "if Peter is an excellent pupil. For instance. thus. the problem of working out a real time-table for an educational institution has a similar character because it is necessary to take into account many interrelated conditions such as the wishes and the possibilities of teachers and pupils or students. in particular. . (cf. .in the argument and to the introduction of mathematical symbols making it possible to write in a concise form both the given conditions and the course of the solution. the necessary alternation of subjects of different character and different difficulty. then . on "propositional calculus" to which Sees." or with the word "implies". The proposition q =>. A traffic controller deals with a similar problem when introducing a rational dispatching system and the like. .p is formed of the propositions q and p by connecting them with the expression "if . the programming of the work of a computer is based on the laws of mathematical logic and. lectures. * The relation p ZD q between two propositions p and q plays an important role and that is why it is advisable to consider one more binary operation of algebra of propositions which is connected with this relation. which is the same. which is the same. 5 and 6 of the present book are devoted. the problem mentioned on pages 72-75). . then an elephant is an insect" or. The solution of the "swimming problem" can easily be obtained by using an electronic computer because the rules of propositional algebra on which the solution given in this book is based can easily be inserted in the "memory" of the computer and the further course of the solution becomes automatical. respectively. This binary operation forms a new proposition called the implication of propositions p and q\ we shall write this operation as q p. laboratory work. Problems of this kind are rather often encountered in practice. At present many problems of this type are often solved on electronic computers. "from the fact that Peter is an excellent pupil it follows that an elephant is an i usee Г or "the fact that Peter is an excellent 76 .

if a pupil whose name is Peter has bad marks we shall consider the above proposition q =>. then all its sides are equal"). For instance. As to the relation p zd q. the implication p =>. Therefore. then p" (for instance. The relationship between the implication q p and its converse p =>. we shall consider an implication q =Ф. At the same time. "if all the sides of a quadrilateral are equal. generally. the proposition q => p may turn out to be true or false. The implication p =>. that is. As is well-known. By definition.q called the contrapositive of the implication q => p is equivalent to the latter for the relation p zd q holds when and only when the relation q zd p holds I'he relationship between the proposition q p (which 77 . in case the proposition q is false the proposition q =Ф. We should stress a peculiarity of the implication q p of two given propositions q and p: in contradistinction to the operations of forming the sum ("disjunction") p + q and the product ("conjunction") pq of propositions.q is quite similar to the relationship between the direct theorem "if q.p to be true when and only when p zd q. thus. as any other proposition. in the general case. the statements of a direct theorem and the corresponding converse theorem are not necessarily equivalent: one of them may turn out to be true while the other can be false. it connects only some pairs of propositions. the composite proposition q => p is equivalent to the proposition "the relation p zd q takes place".p which means "if Peter is an excellent pupil then an elephant is an insect" to be true.p is true for any proposition p (because a false proposition q implies any proposition p). the operation q p is non-commutative. It should be stressed that there is a great difference between the operation of forming an implication q p (this is one of the operations of propositional algebra) and the relation p zd q. then q" ("if the diagonals of a quadrilateral are mutually perpendicular. then its diagonals are mutually perpendicular") and the converse theorem "if p.q differs from the proposition q =>. The composite proposition q => p can be formed of any constituent (prime) propositions p and q.p. the fact that the relation p zd q holds is not a proposition but is a fact concerning these two propositions p and q. the proposition p =>.q is called the converse of the implication q=>p.pupil implies that an elephant is an insect".

78 . In other words. then all its sides are not equal". we shall denote it by the symbol q <=> p. "if the diagonals of a quadrilateral are not mutually perpendicular. if and only if q"x). then its diagonals are mutually perpendicular") and the converse of the theorem which is the inverse of the original theorem.means "if q. then it is false that all its sides are equal"). — Tr. Together with the implication q p. For instance. then p") and its contrapositive p q (that is the proposition "if p is false then q is also false') is quite similar to the relationship between a direct theorem (for instance. in mathematical logic is sometimes considered the so-called biconditional proposition formed of two propositions q and p. By the inverse of a theorem (implication) q p is meant the implication q =>. then b ZD a" which is one of the basic properties of the r e l a t i o n s implies that the relations q ZD p and p ZD q hold or do not hold simultaneously). it is not equivalent to the direct theorem (q => p) but is equivalent to the converse theorem (implication) p => q which means "if p. then its diagonals are not mutually perpendicular").p is the theorem which reads: "if it is false that all the sides of a quadrilateral are equal.p.p (in the case of the above example. and the converse of the latter is the implication p => q (to the example of the direct theorem we have given there corresponds the converse of the inverse asserting that "if the diagonals of a quadrilateral are not mutually perpendicular. the theorem p => q is equal to the direct theorem q p. then q" (because the property "if а ZD b. The biconditional proposition q <=> p reads: "p. q =>. As to the implication q p expressing the inverse of the theorem (implication) q =>. for the propositions p and q given as examples on page 76 the proposition q <=> p reads: "Peter is an excellent pupil if and only if an elephant is an insectThe last statement is a new proposition (although rather funny!): the operation of forming the biconditional proposition from two given propositions q and p is also a binary operation of propositional algebra which assigns to every pair q and p *) Instead of "if and only if" in English mathematical literature is often used the expression "iff" which is the abbreviation for the former. "if all the sides of a quadrilateral are equal.

then an elephant is an insect" is equivalent to the proposition "Peter is not an excellent pupil or Peter is an excellent pupil and an elephant is an insect". It is obvious that the implication q => p is true if and only if either the proposition q is false (a false hypothesis implies any conclusion) or the propositions q and p are simultaneously true (a true hypothesis implies any other true statement). the biconditional proposition is commutative: the propositions q <=s> p and p <=> q are equivalent for any two propositions q and p.a).of propositions a new proposition which we denote q <=> p. In contrast to the implication q =$* p of two propositions q and p. Similarly.p as Q =>• P and the truth set of the proposition q p as Q <=> P. the implication q =>. according to the last relation. It is quite clear that the relationship between the biconditional proposition q <=> p and the equivalence relation p — q is similar to the relationship between an implication p and the relation q гэ p: the proposition q p is true when and only when the equivalence p q takes place. a biconditional proposition q <=> p is t r u e l a n d o n l y if e i t h e r b o t h p r o p o s i t i o n s q a n d p a r e t r u e o r 7!) . that is in terms of the operations of addition of propositions. 28. that is we always have (p <=> q) = (q p) The notion of the "truth set" of a proposition (see page 55) makes it possible to extend the new operations q => p and q <=> p of algebra of propositions to algebra of sets. For instance. This means that the set Q =>• P is the union of the complement of the set Q and the intersection of the sets Q and P (see Fig. It follows that Q^P = Q + QP and consequently Thus. Let q and p be two arbitrary propositions and let Q and P be their truth sets respectively. the above proposition "if Peter is an excellent pupil. We shall denote the truth set of the proposition q =>. multiplication of propositions and the operation of forming the negation. .p formed of two propositions q and p can be defined in terms of the basic operations of propositional algebra.

h): Q <=> P .QP + QP It follows that the biconditional proposition q p formed of two propositions q and p can also lie expressed in terms of the operations of propositional algebra studied earlier. 28 both propositions q and p are false. 28. the set Q <=> P is the union of the intersection of the sets Q and P and the intersection of the sets Q and P (see Fig. Hence.P=>Q Ce) Fig. namely: q p-^qp + qp T h e f o r m u l a s we h a v e w r i t t e n r e a d i l y s h o w t h a t th« oper a t i o n < = > of forming a b i c o n d i t i o n a l p r o p o s i t i o n is a 80 .

the operation of forming an implication is highly important we do not include it into the list of the basic operations which form the foundation of the definition of a Boolean algebra.q of the implication can also be proved without resorting to the Venn diagram: p =>. it also turns out that the three original operations + .q = p + pq = p (q + q) + pq = = P?-!-p? + p? = p<? + (p f-p)q = pq + q ^ q + qp = q ^ p (here we have used the commutative. Similarly. However. For instance.and <=> in terms of the basic operations of addition and multiplication and the "bar" operation of the Boolean algebra. That is why. .q of the implication q=> p is equivalent to the latter: p=>q--= (P) + P(1 = P + P<7 = Я + P7 = 9 P because from Fig. P => Q of the converse p =>.a it is seen that the sets Q => P = Q + QP and P Q = P + PQ coincide. the contrapositive p =>. the associative and the distributive laws and also the identities pi = p and p + p = i).commutative operation of algebra of proposition? while an implication q => p is non-commutative: 4 P = 4P -.q of the implication q is shown). the proposition q =>. 28. On the other hand. 28c. the "Boolean multiplication" of 81 . although.p is equivalent to the converse p => q of the implication q => p: q ^ p = q + qp = p + pq = p=^q * The formulas q =>• p = q' -f qp and q <=> p = qp + qp express the operations =4. • and — are not independent: using the De Morgan formulas we can express each of the two operations -(.and • in terms of the other and the "bar" • peration. where the truth set.</p = p <=> <z but (see Fig. The equivalence between the implication q p of propositions q and p and the contrapositive p =>. for instance.

. . • and — can be expressed. the operation of multiplication AB and the operation of forming the complement A of any set A are defined as was done in Secs.a). C. One of the well-known operations of this kind is the socalled (Sheffer1)) stroke operation a | p (here a and p are elements of an arbitrary Boolean algebra) which is expressed in terms of the "Boolean multiplication" and the "bar" operation as a I p = ap In the case when the Boolean algebra in question is an algebra of sets whose elements А. it turns out that there is an operation defined in Boolean algebra in terms of which all the three operations + . . are some sets for which the operation of addition A -f.B. This makes it possible to reduce all the variety of operations used in Boolean algebras to a single operation and its various combinations. 1 and 2. 29. 29 propositions can be defined thus: pq = p + q Moreover. M. В.Fig. the Sheffer operation A | В reduces to forming the intersection of the complements of the sets A and В (see Fig. Sheffer. The Sheffer operation is obviously commutative. an American logician of the beginning of the 20th 82 century. that is « IP = P I« 3 ) IT. .

p. if we take the Sheffer operation a | p as the basic one if is possible to define а + P. from the basic properties of the operations of a Boolean algebra il follows that (a I p) I (a I P) = (Tp) (âfj . а р and a can also be defined in terms of the operation a j p. F.(a + p) (P + 7) (V + a) This operation assigns a new element ô — {а. а р and a as (a J P) | (a | P). P.a + p Ь(Р)1-оф and (ce (а) = а а — а Thus. p.) For an algebra of sets the element {ABC) = Л В -f. that ( а | а ) И Р 1 Р ) = Г м Г ^ ) + (Р + Р) . 2 9 6 ) .xercise 6 on page 23).ВС С A is the set coinciding with the union the painrise intersections of the sets Л . p defined as а|Р = а+Р (for algebra of sets the operation A j В reduces to forming the union of the complements It. у of a Boolean algebra (cf.а (а Ш J (a IP) = (« + Р) I (â f P) гР = а + Р â'P i а P - ap and ( а | а ) = а-(-а = а Therefore the operations а + p. the "bar" operation assigns a new element a to one element a and is an example of a "unary" operation. The definition of a Boolean algebra is sometimes stated using only one "ternary" operation {a. 7} to every triple of elements а . (a I a) I (P I P) and а | a respectively. MO) . can readily be seen of the sets A and B\ s e e F i g . Further. (The operations + and-which assign new elements to any pair of elements of a Boolean algebra are binary operations.for any a and (3. В and С (see F i g .[(a) + (F)I I W (a I a) I (PIP) Н^Ш(1) f (1)1 . 7 } defined as {a$y} = ap + p7 + ya . The role analogous to that of the Sheffer operation can also be played by another binary operation a j.

if we fix a "special element" i among the elements a.a The operation a can be de-fined with the help of the ternary operation {aPy} by means of the following condition similar to the idempotent law we have written: {aâp} = p (A) Since this condition is symmetric with respect to the elements a and a it obviously implies that a = a Further. for this operation there holds a law analogous to the idempotent laws for addition and multiplication: {aap} =. forming a set for which the ternary operation {офу} is defined and put i = о then it is also possible to define the basic (binary) operations of the Boolean algebra in terms of the ternary operation {аРу}: <x + P = {aPi} and ар = {сфо} (В) By virtue of (A) and the idempotent law. . 30 J T. . 7. p.and ai = (aio) — {an} . we shall also have a + 0 = (ctoi) =• {an} = a. this operation possesses a certain kind of distributivity: {aP{Yfie}}^{{aPY}Ô{aPe}} It also possesses a (weakened) associativity: {ap{apô}} = {{apv}pô} Finally. . whicli is the same.a a 1 — {au}-= i and ао = {аоо}^-=о S4 .or. The ternary operation { } is obviously commutative with respect to the interchange of any two of the elements entering into it: {офу} = {Pva} Fig. with the intersection of the pairwise unions of these sets.*|i ! {yap} !PaVf (#%! Further.

q and r are true".« -I. This obviously follows from the expression of the implication in terms of the other operations of a Boolean algebra: we have q =>.'IP 85 .Definitions (A) and (B) make it possible to state all the properties of the operations of the Boolean algebra so that the corresponding expressions involve the ternary operation {сфу} solely.q + 4P and 1(7 I 4) I {{P I P) I (7 I ?))] I 1(7 I 7) I ((p \ p) I (7 | ï))] p = pq + pq = r\r where Г -. P\P consequently. if the implication q=> p is true and the proposition q is true then the proposition p is also true. 7 P . Further.P) 7 In other words. Using algebraic symbols we can express this relationship as P => (7 =>. we have P + 7 = (P I 7) I (P I 7) pq = (p I p) I (<7 I q) and P = and.4P .[(p ! P) I (q I 7)1 [((p I p) I (p I P)) I ((7 I 7) I (7 I 7))] With the aid of the ternary operation {pqr} the sum ("disjunction") and the product ("conjunction") of two propositions p and q are expressed thus: P + q = { № } and * * pq = {^70} We have already discussed the relationship between the operations q=> p and the relation q zd p. In prepositional algebra the Sheffer operation p | q and the ternary operation {pgr} have the following meaning: p I q reads: "neither p nor q is true" (that is why in logic the Sheffer operation is sometimes referred to as the joint negation) and the proposition {pqr} reads: "at least two of the three propositions p.p ~ q + qp and consequently (7 P) 7 (7 + qp) 7 .

Indeed. Similarly. this can be written as an implication q =>• p). Paul reasons illogically (p). this can be regarded as an implication q => p). p is equivalent to r and r is false then q is also false. it means that if the implication q =s. we have q ZD (q => p) (p <=> r) r which means that if q implies a proposition p.('! =>. we have (ч=>р) (p — qp(rr) r ) r = С 7 + pq) (pr + pr) r - + qpr + pq(rr) + (pp)'ir- ° + я?r4 and.p) q The relation p zd (g =>. Consequently. Peter is mortal (p)". a typical syllogism is: "All men are mortal (that is if N is a man then N is mortal.whence it follows that P => IP . by Virtue of the formulas expressing the implication and the biconditional proposition.p is true and the proposition p is false then the proposition q is also false. Consequently. The logical statement expressed by the relation q=>(q=> p) p is also true. Peter is a man (q). Here is an example demonstrating the application of the logical rule q ZD (q p) p: "All mathematicians reason logically (that is if N is a mathematician then he reasons logically. The last relation also readily follows from the formula for the implication: we have (q p) p = (q + pq)p = чp + (pp) q = qp + o = qp and therefore q ZD {q => p) p = qp. q ZD q pr -= (q => p) (p <=> r) r 86 °-i-o^--qpr . consequently. for instance.p) q expresses the form of a logical statement known as the classical syllogism. Paul is not a mathematician (q)".

"opposite sides of a parallelogram are equal\ the opposite sides AB and CD of the given quadrilateral AHCD are equal. we have (. (or. more precisely. consequently A BCD is a parallelogram"). In contrast to the above. Ibis proposition is an implication q => p). follow from the rules of propositional algebra and may not take place. by non-mathematicians) do not follow from the rules for deduction and are therefore incorrect (it should be noted that an electronic computer which was "taught" the theory of Boolean algebras can never make ^uch a mistake!): "q implies p. Therefore it is false that all the sides of the quadrilateral A BCD are equal (q)".Here is an example of an argument following this rule: "If the sides of a quadrilateral are equal then the quadrilateral is a parallelogram. That is why the following two statements (which. are rather frequently used. particularly. unfortunately. a rhombus. consequently N does not speak well"). "lawyers speak well\ \ is not a lawyer. of course. The diagonals of the given quadrilateral A BCD do not bisect each other (r). therefore the proposition q is also 'rue" (for instance. .i nd u q implies p\ the proposition q is false\ therefore the pro•sition p is also false" (for instance. p is true. Ы . the following two relations may turn out to be false: ? =5 (q=> p) p and p=>(q=>p)q Indeed. A quadrilateral is a parallelogram if and only if its diagonals bisect each other (p <=> r)".q =>p) p^ (q + qp) p = qp -f qp = (q + q) p = ip = p and (q p) q = (q + qp) q = q + (qq) p = q -f op = q and the relation qZD p and the (equivalent) relation pzDq do not.

(b) 88 . 3.q. q and r themselves but not over their combinations: (a) p q\ (P + q) r(C) (p + q) (d) /><=>(/> + q). Find which of the following propositions are true and which are false: (a) p + q^ p. (c) (p =>p) <=s> p. (e) pq <=> qp. 4. (b) pqr + p + q + r. pq=>q. (c) (p + q) (q + 0 (r + p) to (1) form (*) (see page 70). (b) (c) (d) (p + q) [(/> + r) =*(</ + r)h pq <=> rs. Exercises 1. and to (2) form (**) (see page 74). Rewrite the following proposition in the form involving only the addition (disjunction) and the "bar" operation (negation): (a) q=>(p + q). 2.The examples we have considered (their number can easily be increased) demonstrate the role which mathematical rules of propositional algebra play even in everyday life. Write the negations of the following propositions in the form which only involves the sign of negation (the bar) over the propositions p. (d) (p => p) <=i> p. (b) pq q. Reduce the composite propositions (a) pq + p 4.

31) will be denoted by capital letters. Fig. t h a t is electric circuits with a number of switches each of which can be open or closed. Since a section of an electric circuit is only meant for conducting an electric current. it is these sections t h a t are elements of the peculiar algebra under consideration (earlier we used capital letters for denoting sets). 7. Express the implication proposition (<=>) in terms of (a) (b) the operation the operation (=>) and the biconditional \ (see page 83). any two sections containing the same switches and simultaneously permitting or not perm i t t i n g passage of current for the same states of all the switches (every switch can be in one of the two states: "open" or "closed") will be considered "equal" to each other. In other words. we shall consider any two sections which are similar in t h i s sense to be identical ("equal"). Separate sections of such a circuit (for instance. As the elements of this algebra we shall consider various switching circuits. {}. 32 8У . Propositions and Switching Circuits Here we shall discuss one more example of a Boolean algebra which m a y seem rather unexpected. see a section shown in Fig.5. 31 -AB A+B (b) (a) Fig.

33 (A+B)C AC+BC (a) Fig. let us agree that by the sum A + В of two sections A and В will be meant a circuit with these two given sections A and В in parallel connection and t h a t by the product A B will be meant a circuit section with the sections A and В in series connection.-/IВ С Л+В+С (b) (aj Fig. 34 AB+C (A+C)(B+C) С (V С (b) Fig.1В (С) = А (ВС) Du . 32.1 ВС .a and b where each of the sections A and В of the circuit contains only one switch. see Fig.. 35 Further. It is clear that the addition and the multiplication of the sections of an electric circuit are commutative: A + В = В + A These operations are also (A + В) + С - and A B = BA associative: Л + (В + С) - /1 + В + Г and . For instance.

The verification of the distributive laws [A i. 33a and 6 where the "triple sum" /1 + /? -f 6' of three switches and their "triple product" ABC are shown).a is equal to that in Fig.a is "equal to" the circuit in Fig.6). 37 91 . 34. However. i hose laws also hold here (it can readily be checked that (he switching circuit shown in Fig. i'he idempotent laws A + A = A and AA = A : so hold for these operations because when two switches which are in one and the same state (that is when they are .6 while the circuit in Fig. 37) and t h a t A + f = I I -o -o—o(a) 0 o- (b) Fig. 34. 36. 36 AI=A A+0=/l (a) (b) Fig.В) С = AB + BC and AB + C = = (A + С) (В + C) in this "algebra of switching circuits" is a little. as can be seen from Figs. more complicated. 35.I-co Fig.a) and Jet О denote an always-open switch (Fig. 34 and 35. 36. It is evident that A + О = A and AI = A and AO = О (see Fig.6). Finally. 35. let us agree that I denotes an always-closed switch (Fig.ipen or closed simultaneously) are in series connection or il parallel connection the resultant circuit section gives ilie same result as a single switch in that state.

41.Â В and ÂB = A + П are proved in a more intricate manner but they also hold in this algebra (for instance.В and A + В of the circuit satisfy the condi- . 39 (*) (ь) F i g . 40. 38). It is evident that X=A.a and h where the sections A -f. the roles of the "special" elements / and О of this Boolean algebra are played by the circuit sections equal to an always-closed and an always-open switches respectively. such a pair of switches can easily be constructed' (see Fig. Let us also agree to denote as A and A a pair of switches such that when the switch A is closed the switch A is necessarily open and vice versa.a and b). T^O and and 6=1 also A+ A= I and AA = 0 (see Fig. Thus. 40 (see Fig. 38 • . 39).-o— A F i g . The De Morgan rules Â+B . see Fig.V A+1=1 I (a) A0=0 (b) F i g .

In the first place. then the composite proposition d can be represented by the circuit in iig. b and c. The similarity between the "algebra of switching circuits" and the "algebra of propositions" is extremely valuable. "and" and the negation of a proposition respectively. let us consider the composite proposition d = abc + abc where a.- V л • - -О— в (b) Fig. this similarity makes it possible to model composite propositions by means of electric circuits. 41 в с r fi v c и ABC+ABC= =A(BC+BC) Fig. b and с are some "prime" propositions and the addition. For instance. 42 which corresponds to the combination D = ABC + ABC 93 . В tnd С with the given propositions a. the multiplication of propositions and the "bar" operations are understood in the ordinary sense as the ''igical connectives "or". 42 lion that when the section A В permits passage of current !he section A + В does not and vice versa).A8=A+B А+В=АВ а П Г\ -А+В вГ -АВ W —•О. Let us associate some switches А.

that is transforms 94 . Solution. The problem reduces to designing a combination С of the switches A and В (and perhaps A and B) such that the change of the state of any of the two switches changes the state of the whole circuit С to the opposite. В and С in these states permits the electric current to flow then D corresponds to the true proposition i (that is to the circuit I conducting electric current). The condition which must be satisfied is that when each of the switches is operated on the circuit must become open if it is closed before the operation and must become closed if it is open before the operation irrespective of the state of the other switch. If the circuit D with the switches А. 13 of the switches А. in this case the proposition d is true. 43). In the second place. the propositions a and b are true while the proposition с is false it suffices to close the switches A and В in the circuit D and to open the switch С (see Fig.A - '• с С Л V_ - B F i g . Here we shall give two examples to demonstrate what has been said. for instance. the similarity between the algebra of switching circuits and the algebra of propositions allows us to use the rules of logic for constructing switching circuits satisfying some given conditions (which can be rather complex). To verify whether the proposition d is true when. It is required to design an electric circuit for a bedroom with one electric lamp and with two switches one of which is by the door and the other by the bed-side. In case the circuit D does not permit passage of current (that is it is "equal" to the circuit 0) for the given states of the switches then the proposition d is equivalent to the false proposition о when a and b are true while с is false. In other words. В and C. Let us denote as A and В the switches in the circuit. * Example 1.

In other words. Further. 44). which hints that it is possible to represent the proposition с as a sum of two propositions one of which is true when a and b are true while the other is true when a and b are true (that is when a and b are false). 8 V "л /48 M B в Fig.the circuit permitting the passage of current into the one not permitting it and vice versa. and the same requirement refers to the proposition b. This description of the circuit involves the connective "or". we have 1o find a combination с of two propositions a and b such that the replacement of the true proposition a by the false proposition à or vice versa changes to the opposite the sense (the "truth" or the "falsity") of the whole proposition c. we finally obtain с = ab -). The condition stated is satisfied by a proposition с which is true when both propositions a and b are simultaneously true or simultaneously false and which is false in all the other cases (that is when one of the two propositions a and b is true while the other is false). since the descriptions of the summands of the sought-for sum involve the connective "and" we conclude that these summands are ab and ab Thus.ab It can readily be seen that this proposition с satisfies all the requirements stated above. 37 95 . passing back from the propositions to the switching circuits we see that the electric circuit С we are interested in can be expressed by the formula С = АВ + Л В The construction of such a circuit clearly involves no difficulties (Fig. Now.

F—the switch connected with the floor of the car. The electric circuit controlling the downward motion of the lift must be closed only when the car is on the second floor and. we shall also confine ourselves to the circuit controlling the downward motion of the lift1). The circuit also involves the following additional switches: a switch which is closed only when the car is on the second floor. besides. For the sake of simplicity. 96 . a switch connected with the door of the car (the inner door) which is closed when the inner door is closed and a switch connected with the floor of the car which is closed when a person is in the car and the weight of the person exerts pressure on the floor. there is no person in the car. when one of the two following conditions is fulfilled: (1) the outer doors on the first and on the second floors and the inner door (in the car) are closed. According to the conditions of the problem the sought-for circuit Cd controlling the descent of the lift must be closed (must conduct electric current) only if: (1) the switch S is closed and the switch Di is closed and the switch Dz is closed and the switch D is closed and the 1 ) The circuit controlling the upward motion of the lift can he designed in just the same way (see Exercise 6 on page 100). This circuit must involve two switches (push-buttons) one of which is in the car (the descent button) and the other is placed by the lift shaft door on the first floor (the call button). D{ and D2—the switches which are closed when the outer doors on the first and on the second floors respectively are closed. we shall assume thai there are only two floors'. It is required to design an electric circuit for controlling a lift. a person is in the car and pushes the descent button. D—an analogous switch connected with the door of the car. a person on the first floor presses the call button.Example 2. Solution. Let us denote the switches in the circuit as follows: S—the switch which is closed only when the car is on the second floor. Bd and Bc—the switches connected with the descent button in the car and with the call button by the lift shaft door on the first floor respectively. (2) both outer doors (of the lift shaft) are closed while the door of the car is closed or open. two switches connected with the outer (lift shaft) doors on the first and on the second floors which are closed when the doors are closed.

6) is equivalent to the possibility of designing any electric switching circuit using only one special component (we denote it 2 ) with two inputs and one output such that the output electric current can flow if and only if neither of the inputs is supplied with electric urrent. Such an element can easily be constructed (see rig. 45). to this end we can agree that to -420 97 . 45 switch F is closed and the switch Bd is closed or (2) the switch S is closed and the switch D{ is closed and the switch D2 is closed and the switch D is closed or open and the switch Bc is closed and the switch F is open. 4B). Taking into account that the logical operation "and" corresponds to the product of propositions (of switches) and the logical operation "or" corresponds to their sum we readily find Cd = SD^DFBa + SDtDz (D + D) BCF Using the equality D+ D= I and the property of the switch I {AI = A for any switch A) and also the commutative law for multiplication and the distributive law we can simplify the expression we have derived: Cd = SDtD2 (FDBd -f J'BC) Such a circuit can easily be constructed (see Fig. * We also note that the possibility of expressing all the operations of a Boolean algebra in terms of only one Sheffer operation (see Sec.F >y Î c^o Bc Fig. For instance.

49 every section of the circuit there correspond two conductors through one of which the current flows permanently.иззъ (а) (b) (d) (с) Fig. An analogous role can also be played by a component M with three inputs and one output such that the output current can flow only when at least two of the three inputs of the element M are supplied with electric current (see (he scheme in 98 . 47. pages 82-85). 47 к (b) (e) (I) (c) (d) (g) (h) F i g . 48 (a) (b) Fig.a-c the sum A -f.В and the product AB of two circuits A and В are shown and also the scheme of the circuit A corresponding to the circuit A constructed with the aid of the "Sheffer component" S (cf. In Fig. 46 НУ A+B vn> m - AB A U В НУ a () (b) (c) Fig.

Fig. Draw switching circuits corresponding to the following composite propositions: (a) ( a + b ) J c + J ) (b) abc + ab + a (o) abc -f. В. (a) A committee consists of three members. В. Exercises 1.abc -f. Sketch switching circuits corresponding to the propositions (a + с) (b + с) {a + d) (b + d) and ab + cd and check the "equality" of these circuits. С and D (and also perhaps the switches А. 48. It is required that the lamp should give light only in the case when the majority votes for the proposal or when the numbers of the votes for the proposal and against it are equal and the chairman voles for the proposal. Design an electric circuit with . В. 4. 49a and b we see how the "addition" and the "multiplication" of two circuits A and В can be realized by means of the element M. Design an electric circuit showing the results of voting: every member of the committee presses a button when he agrees with the proposal under discussion. (b) Design an analogous circuit for a committee consisting of a chairman and five members. cf. In Fig. С and D are closed or none of the switches is closed. the bulb in the circuit must only be switched when the majority votes for the proposal. С and D are closed. Design an electric circuit E containing switches A.abc (d) {a + b) (à +b) + ab + ah 2. (b) the circuit E is closed only in the case when some but not all of the switches А. В. 5*. С and D) such that (a) the circuit E is closed only when all the switches A. pagej 83).> lamp making it possible to switch on and to switch off the light by means of 99 . 3*. the element M corresponds to the operation {ABC} — AB + ВС + С A = (A + В) (В + С) (C + A) of the Boolean algebra.

I1 I= 1 One of the most important applications lies in the of the so-called probability theory on which. 6. . unfortunately. * * 8. then |a+p|=|a|-f|p|. I о I = 0. For the conditions of Example 2 on page 96 design an electric circuit controlling the upward motion of the lift. however. 2 ) From the fact that a o = о and a + о = a for any of the Boolean algebra and from condition (2) it follows | a | = |a + o| = | a | + whence foundation we cannot element a that |o| Iо1 = 0 Thus. Normed Boolean Algebras The material of Sees. 1-6 of this small book does not exhaust the extensive theory of Boolean algebras. dwell in the present small book. the equality | i | — 1 specifying the "unit norm" is not very important either. using property (2) we can prove that | о | = 0 and. These numbers can be taken as the norms of the corresponding elements: I 0 I == 0. Example 1 on page 94): (h) n independent switches.) 100 . and (2) if ctp = o.(a) three independent switches (cf. In this section we shall discuss one notion which is directly related to the notion of a Boolean algebra and has very many applications 1 ). . conseq u e n t l y . у. and containing a "zero" element о and a "unit" element i is called a normed Boolean algebra if to every element a is assigned its "norm" ("absolute value") | a | which is a nonnegative number satisfying the following two conditions 2 ): (1) 0 < I a | < 1. Examples. . A Boolea n algebra consisting of elements oc. the conditions | a | ^ 0 and | i | > 0 arc essential.0 must not necessarily be included into the list of conditions defining a norm. p. | i | = 1. 1°. (Similarly. The "algebra of two elements" (see page 25) consists of two "numbers" 0 and 1. the equality [ о | =. etc. The notion of a Boolean algebra can be generalized in various ways.

in order to satisfy conditions (1) and (2). Condition (2) will also he fulfilled because the only pair of nonzero elements of this Boolean algebra whose product is equal to zero is the pair of the elements p and q. this Boolean algebra of four elements with the absolute values (norms) of the elements we have defined becomes a normed Boolean algebra.| 0 | + | 0 | 0-1=0 and I 0 + 1 I= 1 = I 0 I+ I1I condition (2) also holds. 3°. let J contain N elements. 2°. with this definition of the norm of an element the Boolean algebra of two elements 0 and 1 becomes a normed Boolean algebra. Now we shall consider an example which elucidates the essence of the notion of a normed Boolean algebra itself. and we have |р + д | = | 1 | = 1 = р + д = | р | + 1 < 7 | Thus. therefore. (Generally. for instance. since 0-0 = 0 and | 0 + 0 | = 0 . we shall also assume that the universal set (we denote it J here) is finite. for the condition \ J \ Л to be fulfilled the proportionality factor must obviously be 1/Ar so that | A | is the ratio if the number of elements contained in A to the number of 101 .) Hence.о | = | a \ — -= I a I + 0 = I a I + I о |. Further. Ip \ ~ p and \q\ = q Then condition (1) will be fulfilled.Then condition (1) oî the definition of the norm of an element is obviously fulfilled. For the "algebra of four elements" considered in Example 2 on page 27 we have pq ^ 0 and p + q = 1. Let the Boolean algebra under consideration be the algebra of sets which we considered in Sec. I1 I . we must put I 0 I = 0.1 and I p I+ Iq I = I1 I = 1 Let the "numbers" (elements) p and q entering into the definition of this Boolean algebra be two arbitrary positive numbers whose sum is equal to unity and let I1 | = I 0 I — 0. for any Boolean algebra in which condition (1) holds condition (2) also always holds if at least one of the elements a and |3 coincides with о because in this case ao = о and | a -j. 1. Let us define the norm of any subset A of the universal set J as a number proportional to the numbers к of elements contained in A.

if J is the set of all chessmen it often turns out that it is natural to consider different chessmen as having different "prices".-дР 102 . . . say.+ ±f=\A\ + \B\{AB = 0) Thus.+ • • • + tin where ilt i2.. . Suppose that difîerent elements au a2. . .. For instance. . . iN are some of the numbers 1.. 2.elements in the universal set / : Then condition (1) is fulfilled.. The above definition of the norm | A | of a subset A of the universal set J admits of further generalization. Let the "weights" ("prices") of different elements au a2. Condition (2) is also fulfilled and its meaning is quite clear: if two sets A and В do not intersect (that is AB = 0) then the number of elements contained in their sum can simply be obtained by adding together the number к of elements in the set A and the number I of elements in the set В whence it follows that \A + B\ = k+L = ±. its "price" is 1000 times that of a pawn. the Boolean algebra in question with the norm of the elements we have defined is a normed Boolean algebra. . . . ait. + tN = 1 \A | = |{a.. • • -. (If we put l t i . in a r c course pairwise different). . the rook "costs" 4 or 5 times as much as a pawn. .to -• . N chosen quite arbitrarily (h. . . . .. . .. the "price" of the queen is 8 or 9 times that of a pawn while the king "costs" much more than a pawn. i2. it is also convenient to choose "unit price" so that Now let us put h + h + . . . aN of the set J be equal to some nonnegative numbers tu t2. — t N =. . dij I = tu -Mi... tN respectively. When we try to "teach" an electronic computer to play chess we usually assume that a bishop or a knight "costs" approximately 3 times as much as a pawn. aN of the set J are assumed to have different "weights" (different "prices"). then the algebra of the subsets of the set J becomes a normed Boolean algebra.

then the new generalized deiinition of the norm | A | of a subset A reduces to the definition in the example above. let us assume that the Boolean algebra under consideration is the algebra of subsets of a set J. this weight is computed by means of the integral \A | = J / (M) da = j j f(x. * This example also admits of an important generalization.course. As the norm | A | of a subset A let us take the weight of the part A of the plate J. Let the specific weight of the material at a point M ---= (x. by the volume of the whole solid / ) . when necessary. almost nothing changes if the role of the universal set J is played not by a unit square but by some other geometrical figure with area S. under this definition. But now we choose. Of. if we take as / a three-dimensional solid then it is natural to define the norm of its subset A as the volume of A (divided. It is quite clear that. Condition (2) will also hold: in this case it simply means that if a figure С is split into non-intersecting parts A and В (that is AB — 0) then its area is equal to the sum of the areas of the figures A and J5. Suppose that the solid / is a thin plate of uniform thickness made of an arbitrary nonhomogeneous material. A A y)dxdy 103 . The algebra of subsets of the square J supplied with the norm thus defined becomes a normed Boolean algebra. The next example is in many respects analogous to the previous one. Similarly. that is as the "relative input" of the figure A into the whole area of J.We see that the conditions imposed on the norm have simple meaning in this case and they are similar to the conditions which define the notion of the area of a geometrical figure. y). y) (that is the weight per unit area) be determined by a (nonnegative!) function / ( M ) = / (x. In the latter case the norm of a figure A should be defined as its area divided by the number S. condition (1) of the general definition of the norm will be satisfied. for instance. As before. as the universal set / a unit square so that the various subsets A of the set J are some geometrical figures lying within the square J. By the norm (absolute value) of a set A we shall mean the area of the figure A.) 4°.

if the condition a 0 b — (a. Let us consider the Boolean algebra whose elements are the various divisors of a positive integer N for which the "sum" and the "product" of numbers are defined. Indeed. as their least common multiple and their greatest common divisor (see Example 4 on page 31). b) — 1 is fulfilled for some numbers a and b (the role of the element о of the Boolean algebra in question is played by the number 1!) means that the two given numbers a and b are mutually prime. more precisely. respectively. In this case we can define the norm | a | of a number a as the logarithm. the change of the base of l o g a r i t h m s from b to с s i m p l y reduces to the m u l t i p l i c a t i o n of a l l the l o g a r i t h m s by the constant factor log c b (by the modulus of the former system of l o g a r i t h m s to base b w i t h respect to the l a t t e r system of l o g a r i t h m s to base c): log c m = log c b -log. Further. it is obvious that condition (1) is fulfilled in this case. b] = ab *) The choice of the base of l o g a r i t h m s is not i m p o r t a n t here because the r a t i o log a/log N is independent of the base. 5°. as the ratio log a!log N because it is required that the norm of the number N (which plays the role of the element i of the Boolean algebra) should be equal to unity 1 ).where do is the (infinitesimal) element of area of the plate adjoining (or containing) the point M. In the same manner we can also construct a normed Boolean algebra whose elements are arbitrary domains contained in a given three-dimensional solid J assuming that the solid is made of a non-homogeneous material and that the norm of a domain lying within J is equal to its weight. that is j f(M)da= j [{/(*. in this case we have а ф b = [a. y) satisfying only one "normalization condition" written abpve) transforms the Boolean algebra of figures A into a normed Boolean algebra. "J y) dxdy = 1 It is easy to understand that the introduction of the norm defined in this way (with the aid of an arbitrary nonnegative function / (x. Indeed.. m Instead of the e q u a l i t y | a | = log a/log N wo can also w r i t e | a | = logjy a (because log j V a = l o g n a / l o g n N for any n). The "unit weight" should be chosen so that the weight of the whole plate J is equal to unity. of that number or. 104 .

we have a normed Boolean algebra in this example. i = 1 and let a =э b when a ^ b. 7). Further. does not permit passage of current. As can readily be seen. 6] and à — 1 — a (besides. it follows from the fact that here we have a ® b = 0 only when one of the elements a and b of the Boolean algebra coincides with 0. Hence. о = 0. As to condition (2). let the operalions be defined as a © b — max [a. Let us assume that the elements of the Boolean algebra are all real numbers x such that 0 ^ x 1. a ® b = min [a. the equality AB О means that at least one of the sections A and /?. b]. if. * * 6°. for instance. naturally. consequently. 105 . say A. let the norm | A | of a circuit section A be equal to 1 when this section permits passage of current for the given states of all switches it contains and let | A | = 0 when the current does not flow through the section A. condition (1) of the definition of a normed Boolean algebra is fulfilled because all the possible values of the norm are equal to 0 or 1. Then. see Example 3 on page 28). therefore when AB = О the circuit A + В permits passage of current when the other circuit (B) conflicts current and does not conduct current if otherwise.and. b = 0 then obviously а ф b — a and | а ф b | = | a \ — = I а I + 0 = I а I + I ft I7°. Another interesting example of a normed Boolean algebra can be obtained if we introduce a norm into the algebra of electric switching circuits (see Sec. I his Boolean algebra also becomes a normed algebra if we put I a I = a. By definition. Then condition (1) of the definition of the norm of an element of a Boolean algebra is obviously fulfilled. the norm of the section О which plays the role of the element о is equal to 0 (this circuit section never permits passage of current) and the norm of the section I which plays the role of the element i is equal I о 1. log (а ф b) — log (ab) = log a + log b that is | a © f t | = | a | + | f t | We have thus proved that condition (2) entering into the definition of a normed Boolean algebra is also fulfilled liere.

Let us introduce the norm | p | of a proposition p in the algebra of propositions by putting | p | = 1 when the proposition p is true and | p | = 0 when p is false. Let us consider an example which is very similar to the example of the normed Boolean algebra considered above. Hence. that is at least one of the two propositions p and q. in case pq = О the proposition p + q (that is the proposition "p or q") is true if and only if the other proposition (q) is true. All the operations of propositional algebra can be characterized by the indication of the truth values of the composite propositions obtained from the constituent (prime) propositions by means of these operations depending on the truth values of the constituent propositions. the truth table corresponding to the ope106 . the relation pq = 0 means that the proposition "p and g" is false. Here we also have 0 < | p | ^ 1 (or. 8°. say p. \. more precisely.whence it follows that for such circuits A and В we have \A+B I = \A I + \ B |. Further. Exercises 1-2 below). The sum (disjunction) p -f. is false.p I = 0 or I p I = 1) and I о I = 0. Therefore. the operations p + 4 and pq can be described by means of the following "truth tables": IpI l l 111 1 0 0 1 0 0 lp+«l \P4\ 1 1 0 1 0 1 0 0 Tn just Ihe same way we can compile the truth table corresponding to any other composite proposition (cf. The normalization condition for the elements of the algebra of propositions assigns to every proposition one of the two numbers 0 and 1 which is the truth value of the proposition. I i I = 1. Whence it readily follows that in this case | + <7 | = | jo | + | <7 ]• The most important normed Boolean algebras we have considered are of course those in Examples 7° and 8°.q of two propositions p and q is characterized by the condition that p + q is true if and only if at least one of the propositions p and q is true while the product (conjunction) pq of these same propositions is true if and only if both propositions p and q are true.

ration of forming negation is particularly simple: l/'l 1 0 • \~p\ 0 1 It is evident that such truth tables completely characterize the propositions to which they correspond. the norm of an electric circuit (see Example 7°) characterizes the conductivity of the circuit which is its only important characteristic: this measure of conductivity is equal to 1 or 0 depending on whether the current flows or doesnot flow through this circuit. in other words. then I a I > I p I The existence of the "difference" of two elements a and p for which a zd p also implies that for any two elements 107 . if o d P . Here we shall not dwell in more detail on the truth tables for propositions and the values characterizing the conductivity of switching circuits (for these questions see bibliography at the end of the book). It follows that M . First of all it follows that if а а ~ о and a + a = i then | a | + Ja| — j a . for any two elements a and p of a Boolean algebra for which the relation a zd p holds there exists an element £ (the "difference" between the elements a and P) such that a = p+ | and p|=o (see Exercise 3 below). From properties (1) and (2) of the norm (the absolute value) of an element of a Boolean algebra we can derive some further properties of the norm.= I P + Ê| = IPt + I Ê l > I P I or. Let us consider some other examples of the application of normed Boolean algebras to elementary mathematical problems.f a | = | i | = l Consequently. Analogously. we see that I а I = 1 — I a I for all a Further.

50 a and p of a Boolean algebra we have a + p = a + (P - ap) where the element p — a p (the "difference" between p and aP) possesses the property that a . How many students can play neither chess nor draughts? 108 . besides.( P — aP) = о and. P . 50) and therefore \ A + В \ ~ \ A \ + \ В \ . Let there he a group of 22 students among whom 10 students are chess-players. 8 students can play draughts and 3 can play both. then the sum I A I -f.\ AB \ Example.Fig.I В I involves twice the area of the intersection AB (see Fig.a p + ( p .aP) where a p . by virtue of condition (2).( p — aP) — о whence. let | A | and | В | be the areas of two geometrical figures A and В (see Example 4 above).a p I On subtracting the second of the last equalities from the first one (these are number equalities) and transposing the term I p I to the right. we obtain Ia + p I- Ia I+ Ip - ap I and I P I = I «Р I + I P . we obtain the relation Ia + p I= I a I+ Ip I- I ap | (A) For instance. chess and draughts.

We have to determine the number of students in the set Ch-~Dr=Ch + Dr (see the corresponding De Morgan formula). page 102) and \Ch.^ = £2 Consequently. by virtue of formula (A).a 2 a 3 1) == ---1 a . It is clear that equality (A) is a generalization of property (2) of the norm. By virtue of (A). |+1 a 2 1 + | a 3 1 — | a 2 a 3 1 — (| a . a 2 | +1 a t a 3 1 — | a.I a .Dr]=^ whence.Let us denote the set of the students who can play chess hv the symbol Ch and the set of the students who can play draughts by the symbol Dr.g = 1 Thus. Let us consider three arbitrary elements a I ( a 2 and a 3 of a normed Boolean algebra. * We also note that equality (A) admits of a further generalization. there are 7 students in the group who cannot play either chess or draughts. Let us also define the norm in the algebra of the sets of the students in the group as it was done for a finite universal set in the Example 3°.a 2 a 3 1 109 . I Cfe • Dr I = I ChTDr] = 1 -1 С h + Dr | = 1 . we liave i ai + a 2 + a 3 1 = | a t + (а 2 +«з) | = | | +1 «2+a 3 1— I ^ ( « г + а з ) ! •= Ч a i I + (I a 21 + I a 31 — I I) — I «1«2 + ai<x31 = . We have (cf.a 3 1 — | a 2 a : ) | + | a. I + I сц I + I a 3 1 — I a j a 2 1 — | a. For оф = о it goes into property (2). From (A) it also follows that for any two elements a and |3 of a normed Boolean algebra we have Ia + M < Ia I+ I M (В) This property of the absolute values (norms) of the elements of a normed Boolean algebra is analogous to the wellknown property of the absolute values of numbers. we obtain \Ch + Dr\ = \Ch\ + \Dr\-\Ch.Dr\=--^ + l .

. We can similarly express the norm of a sum of four elements: I «1 + «2 + «3 + a 4 I = I «1 + («2 + «3 + «4) I == = I I + I a 2 -f a 3 + a 4 1 — | a . a 2 a 3 a 4 1 Generally.-ai2 . a 2 a 4 ] -f+ 1 a t a 3 a k | + 1 oc2a3a.. + а 2 + а п + . ih) under the summation sign indicates t h a t the summation extends over all the possible combinations of the pairwise different indices iu i2..because. | + | a2 | + . . (a 2 -f. ... i2. we can derive from formula (B) the relation I a! + a2 + . . 2. | — | a . a 3 1 -f1 a t a 4 1 — | a t a 2 a 3 1 — — I a4a2a4 ( — | a1a3a41 -f | aia2oi:jcci |) = = I a.. ij) I aitai2ai. . S in-1) + |a. + a „ | = S I «i I— 2 К. . if • •• in) of which the last term in the expression on the right-hand side does not involve the summation sign 2j). • • + proved by induction. In the same manner. + (-1Г2 (ii..1 + i + S (ii.. | ccn | (B') . 2. • • -. . ij.a 4 ) | = I a j I -I. . . 3. obviously ( a 1 a 2 ) . .. . n and the sum contains only one term (because ( i i . . ik each of which can be equal to 1.. ai(| J + + (-l)"-1|a1a2a3 . а >. an| (A') where the symbol (ij. . + an | < which 110 can readily be | a.( a 1 a 3 ) = a ^ a g . .a 3 4. I -f I a 2 I + I a 3 1 + I a41 — I a t a 2 1 — | aja 3 1 — | a t a 4 1 — — I a 2 a 3 1 — I a 2 a 4 1 — | a 3 a 4 1 + | а { а 2 а г | -f | a .b e t the reader prove formula (A') by induction. . . и.(I a 2 1 + I a 3 1 + I a 4 1 — I a 2 a 3 1 — | a 2 a 4 1—| a 3 a 4 1 + + I a 2 a 3 a 4 1) — | сцс^ -f a ^ + a t a 4 1 = ™ I «11 4-1 «21 + I аз i +1 a41 — I a2a31 — | a2a4 j — | a3a41 + + I a 2 a 3 a 4 1— (| a t a 2 1 + | a .. ii I — . Iа. in the case under consideration we have к = 1.

89 and consequently the percentage of the people who speak none of the three languages is equal to 1 ) I <7/У I I e + f + g | = l .0. In particular. One Hundred Problems in Elementary Mathe•itirs. -) The idea and the solution of this problem are adopted from the •"">k: H.l' a Boolean algebra there holds the relation a.11% Thus. til . By virtue of formula (A').0. 15% of the people speak English and French.02 + 0.0. we use the fact that efg = e + / + g. Example 1. Let us consider the algebra of the sets of the people in the group as a normed Boolean algebra in which the norm is introduced as in the first Example 3°.0.\ e g \ - I fg I + \efg I .2 . . a 2 . It is known that in a group of people 60% of the people speak English. for instance. 20% of the people speak German.\ e f \ . 5% of the people speak English and German-.f g | = l .2 this relation goes into the De Morgan formula a j a 2 = a. speaking French and speaking German as e.01 . f and g respectively.15 . Oxford.0. Then. What is the percentage of the people who can speak none of the three languages? Solution. There are 25 pupils in a class among whom 17 pupils are cyclists. 1963. 2 ). Let us denote the sets of the people speaking English. .05 .. •OLh = a l {-a 2 + . a 2 . 11% of the people speak none of the three languages. 2%) of the people speak French and German. au o. None of the pupils is good at all the three kinds 1 ) In this solution we use the following fact which can easily be proved by induction: for any number к of elements a x .Below are some examples demonstrating various applications of formula (A').0.|e + / .6 + 0.3 + 0.. we have \e + f + g \ = \ e \ + \ f \ + \ g \ . 30% of the people speak French-. Steinhaus. ef is the set of the people speaking both English and French. 13 pupils are swimmers and 8 pupils are skiers. 1 % of the people speak all the three languages.89 = 0 . Example 2. . + a ft l o r к —. • .i-a2 . Pergamon Press... 1 1 .

swimmers and skiers by the symbols Cy. the swimmers and the skiers have satisfactory marks in mathematics (let us agree that the progress of the pupils is appraised with the three marks "good". finally. It is obvious that and. swim and ski. I S Sk\ = \Sk\ and and it is required to determine the values of norms | G | and I Sw Sk |. How many pupils in the class have good marks in mathematics and how many swimmers can skil Solution. consequently. "satisfactory" and "baa!") and it is known that 6 pupils in the class have bad marks in mathematics. Then the conditions of the problem can be written in the form (1) | C „ | = g . Si and S2 the following sets of the pupils having satisfactory marks in mathematics: the set of the pupils who cannot cycle. Let us denote the sets of the cyclists.g 112 (a) .of sport. S denote the set of the pupils having satisfactory marks and. Let us denote by S0. I S 5 ю | = | 5 ю | . let В be the set of the pupils having bad marks in mathematics. The cyclists. Sw and Sk respectively. I + I S 2 I IGl + ISol + ISil + l ^ l . = I = | (2) I Cy Sw Sk | = 0 (3) IS Cy\ — \Cy\. Then we can write Consequently I 5 I = I S n I + I S. let G denote the set of the pupils who have only good marks in mathematics. the set of the pupils who are good at one of these kinds of sport and the set of the pupils who are good at two of the kinds of sport (remember that none of the pupils is good at all the three kinds of sport!). Further.

note t h a t the condition | S x | = 0 means t h a t each of the pupils who is good at least at one of the kinds of sport is also good at one more kind of sport. This yields 2 1 6 1 + 2150 1 + 1 5 x 1 = 0 Since the sum of three nonnegative numbers can only be equal to zero when each of the numbers is equal to zero we thus obtain I G I = 0. I S0 I = 0 and I Si I = 0 Further. ' c 2 . Taking this into account we arrive at the following system of equations: g = = Sw + Cy Sk\ = \Cy Sw\ + \Cy Sk I i § = | S u .| + | S w Sk I A = I S к I = I Cy Sk + Sw Sk\ = \Cy S к | + | Sw Sk | from which we find \Cy \Cy Sk\=^ and \Sw Sk\--^ Thus. „ . 8 25 25 25 . 13 . | = ]C!/ Sw + Sw Sk\ = \Cy Su.§ - (b) Now lot us duplicate equality (a) and subtract equality (b) from the result. the number of the pupils having good marks in mathematics is equal to zero and the number of swimmers who can <ki is equal to 2. 38 ' ~ 25 . It follows t h a t ]Si| + 2 | S a | = . -420 U 3 .But we have jS. I-f | 5 j | = | 5 Cy + S Sw + S Sk | = = \S Cj/| + | S + —15 Cy Sw\- — I 5 Cy Sk\ — \S Sw Sk | = = |Cty| + |Sï/'| + |Sfc| — С S Су 5Jfc| —15 = y Sw I — Sw Sk | = 17 . ' 2I (here we have again used the fact that none of the pupils is good at all the three kinds of sport).

0 . Yaglom and N. h. Prove that there are at least two patches which overlap so that their common part has an area not less than 1/5. h.h. ). At and A s . Moscow. Solution. Chentsov. O n adding together all the inequalities thus obtained we get \h\At\>2 i=l 2 1 4 a. h Ь) + 2 (г. N.}) 1 .3 2 14. 3. M. W e know that | At I ^ y where i 1. \ Al3 etc. 5. their pairwise intersections will be denoted A 12 . 2. Geometrical Estimates and Problems in Com binatorial Geometry.j. 1974. To this end we use the same formula (A') to obtain | 4 > | 4 + 4 + 4 + 4 l = = 2 | 4 | . S) 1 4 1 - S ('. — whence i . h) 14мЫ4з4 Я | where the summation indices i. Let us denote the patches (which are regarded as subsets of this coat J having unit area) as Au A2. "Nauka". 4.Example 3 A coat of area 1 has 5 patches the area of each of which is not less than 1/2. I. Shklyarsky. 5 (see Example 4° on page 103). | Â4 | and | A5 |. In just the same way we can write analogous inequalities for | A2 |. 4. A-j.j.2 j) 1 4 1 + S (i. + j) S (i.l) \Aijhl\~5\Al23l& 1 ) This example admits of an extensive generalization (see problems 59 and 60 in the book by D.2 l 4 + i s (г. j.2 | 4 Л i=2 (г.h) + 4 + 2 (i. h) 14*I- 2 (i. it is necessary to estimate the quantities | Ai} |. к run over the values 2.(i. in Russian).1 (i. 3. i) \Aijhl\-\Ami5\^0 (1) From the inequality we have obtained we now eliminate the term 2 М ш I the following terms. I) I 4fti I + I 4 з 4 5 I 14*1 + i. | Ая |. j. According to formula (A')'we can write 1= |/|>|Л1 + Л 2 4 A3 + Ak + A b \ = = i 2= t 1 4 . j. 114 .

j) whence we derive the 2 |Л г .4 S \ЛШ1\ (i.|>22МН-3>2(54)-3 = 2 (i.ft. И 12451. ft.J. i. j) 0. inequality . ) V Z 10. 115 . И 1 2 3 5 |. }) (i. j.2 2 Ий-1 + 3 S \Aijh I (i. we conclude that at ' least one of these terms is not less than 2 : 10 = 1 5 which is what we intended to prove.h) :=1 . inequality (3) can also be rewritten as 4 г (г. I) + + ~\Al2345|>0 (3) It is clear that each of the five quantities Hi234|. The inequality obtained in this way does not contain the term with ^ \Aijh\- 1-12ИН+! 2 l^l-s 2 i (i. ft. M) + 5\АШ16\^0 (2) Let us multiply inequality (2) by 1/3 and add the result to inequality (1). 1.whence ^ M i 1 . И13451 and И 23451 not less than \A 123451 and therefore \Aim\>b\Aiwa\ 2 (i.I) and 2 \AUM I — | " l ^1234 5 |>Ml2345|>0 (». since the number of the terms entering in the sum V^ | A t j I is (ij) equal to ( .i. 0 Consequently. J) i Now.

ah be arbitrary positive integers.. ^ (b) (P + j ) (p +1). (c) pq + r 3*. . c. Let a 2 . . b. c. ah. c.. a 3 = 30. аг. Apply the formula you obtain to the case к = 4. How can we determine the number ф (N) of the positive integers which are less than N and relatively prime to N? (The function ф (. Prove that if a and p are elements of a Boolean algebra such that A ID p then there exists the "difference" E of the elements a and p. 6. b. . p2. Compile the truth tables for the following operations: (a) the implication (b) the biconditional proposition (c) the Sheffer operation |. Let N --. .d — min [a. d] — min [a./V) of the positive integral argument N is called Euler's function. d] = a + b + с -f. c. Prove that max [a. that is an element of the Boolean algebra such that p + I = a and P£ = o.Exercises \ . с] + min [a. . d] + -f min [a. 4*.) 5*. pn are the prime factors of the integer. (d) the operation { } (see page 83). .p\'p%' • • • Pnn be a positive integer where Pi. b. аг =• 10. ah] of the numbers аг. the operation \ (see page 83). Compile the truth tables for the following compound propositions: (a) pq + p -f q. . a 4 = 45. d-] + min [a. d\ . d] -f + min [Ь. 2. . It is required to find the least common multiple [ a ^ . a 2 = 12. — min [c. . b. Let the numbers themselves and the greatest common divisors of any combinations of these numbers be known. b] — — min [a. с and d be 4 arbitrary numbers. b. . Let a. c] — . . .

page 25. ) As was mentioned. In a Boolean algebra a unary operation (the "bar" operation) is also defined which associates a new •lement a with one element a of the Boolean algebra. у.« ï + Pï (4a) оф + Т = ( а + у) (P + у) the distributive laws Rules concerning the elements о and i (5) a f o (5a) aaoi == oa (6) a . addition (multiplication) is a binary operation. it associates a new element a + (a(5) with every two elements a and P of the Boolean algebra.ф + 7 ) (2a) (оф) Y = a ((Jy) the associative laws (3) a -f а = а (3a) a a = а the idempotent laws Rules connecting addition and multiplication (4) (a + P ) Y .Appendix Definition of a Boolean Algebra A Boolean algebra is an arbitrary set of elements a. (5. It is also required that two "special" elements о and i should exist and that the following rules should hold: Rules for addition Rules for multiplication (1) ct + p = p + ct (la) оф = ра the commutative laws (2) ( а + р ) + 7 = а + . • • -. 2 117 . for which two operations called addition and multiplication are defined which associate with any two elements a and {5 their sum a + and their product aji1) and for which the "bar" operation is defined which associates with any element a a new element a2).f i == ia (6a) Rules concerning the "bar" operation (7) а = а (8) ô = i (8a) î = o 1 ) Cf.

а + р t h e De M o r g a n f o r m u l a s It is unnecessary to include into the definition of a Boolean algebra the requirement that the inclusion relation ID should exist for some pairs of elements a and P. namely: a zd a if a :r> p and p id a then a = p if a zd p and p ID y then a ID y i ID a a + ^D a if a zd p and and a ID о a zd оф then p zd a (let the reader derive them). books 11]. it is even unnecessary to require in the definition of a Boolean algebra that one of the two operations of addition and multiplication should exist. The above definition of a Boolean algebra is highly "noneconomical" in the sense that many of the properties we have enumerated can be derived from some other properties and it is therefore unnecessary to require that the former should be fulfilled. that is it suffices to require that only one of these operations and also the "bar" operation should exist. for instance. if the operation of addition and the "bar" operation exist. For instance. most of the books and scientific papers 118 . the matter is that the inclusion relation a ID p can be defined by means of any of the two conditions a + (3 — a and aft p from each of which all the properties of the relation ID can be derived. [2] and [5] (see the bibliography). in contradistinction to the definition of a Boolean algebra we use in this book. we can define the multiplication using the corresponding De Morgan rule: ap = cc + p However. Moreover. We also note that. On this question see. if only the operations of addition and multiplication exist (but the "bar" operation is not defined) they do 'not specify a Boolean algebra.Rules connecting the "bar" operation with addition and multiplication (9) a + p = a p (9а) « р .

However. . then this algebra coincides with the "algebra of two numbers" (page 25).add two more conditions connecting the "bar" operation with addition and multiplication. for this approach to the notion of a "Boolean algebra" Examples 3 and 4 in Sec. namely (10) a + cc = i and (10a) a a = o (cf. page 53). With these two additional conditions included the "algebra of least common multiples and greatest common divisors" (Example 4 on page 31) is a Boolean algebra when the number N decomposes only into pairwise different prime factors. If we include these conditions in the definition then the "algebra of maxima and minima" (see Example 3 on page 28) is a Boolean algebra in the single case when the set of numbers in question contains only two elements: 1 and 0. 2 can naturally be regarded as examples of "incomplete" (or "generalized") Boolean algebras because properties (10) and (10a) in which they differ from the "complete" Boolean algebras (for which these conditions hold) are not so very important.

[{A + В + С) (В + С + D)} (С + D + А) = = [AD + (В + С)} (С + D + А) = l(AD + В) + С] х X [(А + D) + С] = (AD + В) (А + D) + С = AD + + AD + АВ + BD + С = АВ + AD + BD + С Section 2 3 ./ . bft] • • Ph . 6. (A + B) (A + C) (B + D) (C + D) = {(B + A) x (С + Л)М(5 + D) (С + Я)] = (ВС + А) {ВС + D) = X = (A + ВС) (D + ВС) = AD + ВС second distributive law). [m.Answers and Hints Section 1 1. (here w e use the 2. (А + В) (В + С) (С + А) = ABC + АВ + АС + + ВС = ABC + ABI + АС + ВС = АВ (С + / ) + + АС + ВС = ABI + АС + ВС = АВ + ВС + С A (see the identity proved on page 23).5 = AB. n) = } 120 P P 1 I Q I Q I I I I I 0 P Q 1 0 О О О 0 P О P 0 0 О P QQ Q I max [a h . A (A + В) = AA + AB = A + AB --= 4 / + = = Л (I + B) = AI = 5. О P Q I n]=p (m. b k ] • Pk min [aft.A CD + ВС + BD = AC + ВС + BD 10. [(A + В) (В + C)} (C + D) = (AC + B)(C + D) = = AC -f. Л (Л + / ) (Я + 0) = Л . (a) + 0 1 0 1 0 1 I I О I and 0 0 0 1 О I (b) + О P Q I О P Q I and 0 P Q 1 6. 7.

Section 3

1. AB -f AC -f BD + CD = (A + D) (B + C)

(see

Exercise 1, Sec. 1); A + AB =- Л (see Exercise 2, Sec. 1);

AB + ВО + AI = A (see Exercise 9, Sec. 1); ABC +

+ BCD -f CD A = (A + B) (A + D) (B + D) С (see

Exercise 10, Sec. 1).

2. (a) {A + B) (A + B) = A A + AB_ + В A + BB = A+ AB + BA + 0 = A+BA+BA=A

+

+ {В + В) A = A + IA_ = A + A = A.

(b) AB + (A + В) (A + B)_= AB_+ AA + AB +

+ BA + BB = AB + О + AB + В A + О = AB + _

+ AB +_BA = (AB + AB) 4- (AB + AB) = A(B+B)

+

+ (A +2) В = AI 4- IB

+ B._

(c) ABC AB AC = (Â + B + C)(Â + B)ÂC =

= 1(4 +

(A + В) AC = A (A + B) AC +_

4- В (A -f В) AC + (Л_4- В) A (CC) = (ЛЛ) ( 4 + B J C +

4- [B (AA) С + (BB) AC] + (A + В)А0 = 0(А +В) С +

+ [BOC + OAC} + 0 = 0.

(d) A 4- В = A + IB = A +_(A + А) В = A +_

4- AB 4- AB = (AI + AB) + AB = A(I + B) + AB =

= AI 4- AB = A 4- AB.

3. Apply the "bar" operation to both members of the equality in question and use the fact that A = A.

4 .AB 4- А В = A (see Exercise 2 (a)}; (A + B) (AB 4+ AB) = AB (see Exercise 2 (b)); A (A

B) — AB (see

Exercise 2 (d)).

6. (a) To every divisor m of the number N there corresponds subset of the set I = {рг, p2, . . ., pu} of the prime

factors of the number N consisting of those prime factors

of N which are simultaneously prime factors of m. If A

and В are the subsets of the set I corresponding to two numbers m and n then to the numbers m ф n = [m, n],

m (g> n == (m, n) and m = N/m there correspond the sets

A 4- B, A B and A respectively.

(b) If m, = pa and n — ph then m ф n = [m, n] =

= рта* f». ь1, m ® n = (m, n) — pminfa, b] a n ( j m = дг/ ш ^

=

pA~a.

121

(с) m - Nlm

. . .

7. These equalities do not hold for the "algebra of maxima

and minima" (except for the case when the algebra consists

of only two numbers) and for the "algebra of least common

multiples and greatest common divisors" (except for the

case when the number N decomposes into pairwise different prime factors; cf. Exercise 6 (a)).

8. (a) (A + B) (A + C)

A + AC + AB + ВС

=

- AI

- AI

+

+

{A

AC + AB + ВС = A (I + С + В) + ВС

=

B C ^ A + BCczA+BŒLA

+ B + C.

+ В) (A + С) (A + / ) = (A + В) (A + С) I

(b)

= (A + B) (A -f С) = A + ВС id A zd ABC

cise (a)).

(c)

{A +

В) (В +

С) (С +

A)

=

AB

+

ВС +

(cf.

CA

=

Exer=>

**zd AB zd ABC (see Exercise 6, Sec. 1).
**

(d) Since A zd AB and В zd AB we have A + В zd

zd

a b

+

a b .

**9. ABC с- AB + AC (see Exercise 8(a); AB + AC +
**

AO cz A + В + С (see

Exercise 8(b)); AB + ВС +

+ CAœ:A+B

+ C (see Exercise 8(c)).

10. AB a{Â+

B) (A + B).

12. (a) A; (b) В; (с) I; (d) 0 .

Section 4

5. (a) "the number is even and prime"; the truth set is

{2}; (b) "the number is odd or prime"; the truth set is

{1, 2, 3, 5, 7, 9, 11, 13, 15, 17, . . .}; it differs from the

set of all odd numbers in the fact that it includes the number 2; (c) "the number is odd and prime"; the truth set is

{3, 5, 7, 11, 13, 17, 19, . . .} and it differs from the set

of all prime numbers in the fact that it excludes the number 2; (d) "the number is even and not prime"; the t r u t h

set is {4, 6, 8, 10, 12, 14, 16, . . .} and it differs from

the set of all even numbers in the fact that it excludes the

number 2; (e) "the number is odd or not prime"; the truth

set is {1, 3, 4, 5, 6, 7, 8, 9, 10, 11, . . .} and coincides

with the set of all positive integers with the exception of

the number 2.

Section 5

5. a + b — a; ab — b.

122

Section 6

**1. (a) pq + pq_+ pq_+ pq^ (b) pqr + pqr + pqr + pqr.
**

2. (d) pqrs -f pr + ps -f qr + qs.

3. Propositions (b), (d) and (e) are true (for all p and q).

Section 7

1. (a) See Fig. 51; (b) see Fig. 52.

3. (a) See Fig. 53a, (b) see Fig. 53,6.

4. (a) D = AB + AC + ВС (А, В and С are the push

buttons which the members of the committee press).

(b) G = А (ВС + BD + BE + BF + CD + CE +

4- CF + DE + DF + EF) + BCDE + BCDF +

+ BCEF + BDEF + CDEF (A is the push button which

the chairman presses and B, C, D, E and F are the push

buttons for the members of the committee).

5. (a) D = ABC + ABC + ABC + ABC.

- г г ^ г

L^V-I

В

L_aSr>_](/l+B)(C+D)

0

Fig. 51

A

Д y

в

SV

с

A(BC+B)+A

F i g . 52

0^

0„

E=(A + B + C + DJ( A + B+C+D)

E=ABCD+ABCD

(bj

Fig.

123

. n). This formula can be obtained by applying (/!') to Example 3° in which the roles of the elements of a Boolean algebra are played by the sets of natural numbers not exceeding N and divisible by pt (where i — 1. .. Apply formula (A') to the Boolean algebra in Example 5°. 2. 5 . .Section 8 3. . £ = ар.

E.J. Sets. G. New York. L.. Applied Boolean Algebras. to the Foundations Halt.. 7. An Introduction the Fundamental Concepts of Mathematics. Princeton (N. Prentice-Hall.J. 1948. Newson. . 9. M.. . 8. . Invitation a la recherches opêrationelle.Bibliography 1. Van Nostrand. Mathematics and Logic for Digital Devices. R . Englewood Cliffs (N. J. S t a l l . Francisco. C. This is an extensive text-book for junior students (non-mathematicians). All the questions treated in the present book are discussed in detail. Freeman. V. R . Logic and Axiomatic Theories. C u l b e r t s o n . K a u f m a n n . H . . This book does not require any prerequisites except the knowledge of elementary mathematics but the reader of the book is supposed to have some experience in studying mathematical literature. J. G. С. 1957. Introduction to Finite Mathematics. however much attention is paid to problems concerning computers.. Oxford University Press. In this book meant for senior p u p i l s some questions related to Boolean algebras are also considered. . 125 . The book starts with a detailed discussion of the questions considered in the present book and contains many problems. Macmillan.. Thompson. Mathématiques nouvelles. Paris. H . A . F a u r e . One of the chapters of this book requiring no special prerequisites is devoted to Boolean algebras. This book is meant for m a t h e m a t i c s teachers. Rein- hold. 1961. all the questions considered in the present book are discussed and there are many problems for self-training. I. Paris. Kemeny. E . R . ..). Faure. San 1974.. Reinhart and and Winston. New York. Eves. Courant. 5. T. K a u f m a n n . T. Snell. 1960. L.. J .. R . F. 1963. t h i s part is devoted to Boolean algebras. Symbolic Logic and Intelligent Machines. 4. 3. 1962. Dunod. New York.. Denis-Papin. R . Robbins. What is Mathematics?. .)-Toron to-New York-London. 1957. New York. 6. Dunod. B e r k e l e y . This book is in many respects similar to [3] but Boolean algebras are considered in it in less detail. Hohn. The first part of this book is a reference book meant for engineers and students studying mathematics. 1961. A. 2.

125 Sheffer. F. N. 8 Pythagoras 66 De Morgan. J. 125 Venn. 116 Eves. G. A. G. N. V. 114 Courant. 5 Hohn. 5 Chentsov.Name Index Berkeley. 41 Robbins. J. R. 125 Stall. 125 Newson. M. H. T. S. L. 1. F. 125 Faure. 125 Euler. H. L. 125 Gindikin. R. J. D. 13. Л. L. 33 Kenieny. H. I l l Thompson. I. 82 Shklyarsky. 125 Kaufmann. H. J. 125 Boole. 114 Snell. 125 Kizner. E. С. 114 . 1. 125 Niven. D. 13 Yaglojn. G. 125 Culbertson. R. G. 125 Steinhaus. M. E. R. C.

15. 84 of double negation 64 of dualization (De Morgan theorem) 40 / of excluded middle 68 idempotent 21. 13. 64 of contradiction 68 distributive 10. 19 Least common multiple 31 Multiplication of complex n u m bers 9 Multiplicative (conjunctive) normal form of a proposition 74 Multiplicative inverse (reciprocal) of a number 11 Necessary condition 66 Necessary and sufficient condition 67 Negation (denial) 61 double 64 joint 85 Norm (absolute value) 100 Number complex 7. 19.Subject Index Addition of complex numbers 9 Additive (disjunctive) normal form of a proposition 71 Additive inverse of a number 10 Algebra of propositions 54#. 65 Law (s) associative 10. 60 Algebra of sets 12 Boolean algebra 33 definition of 117 ff normed 100 Condition necessary 66 necessary and sufficient 67 sufficient 66 Conjunction 58 Conjunctive (multiplicative) normal form of a proposition 74 Deduction 66 De Morgan rules 40 f Denial (negation) of a proposition 61 Disjunction 57 Disjunctive (additive) normal form of a proposition 71 Dual formulas (relations) 38 Euler's circles 13 diagram 13 function 116 Greatest common divisor 31 Iff ("if and only if") 78 Implication 76 converse 77 contrapositive of 77 inverse of 78 Inclusion relation 47. 13. 9 irrational 7. 84 complementation 40. 9 pure imaginary 7 rational 9 real 7 signed 9 Operation binary 83 (Sheffer) stroke 82 t e r n a r y 83 u n a r y 83 Principle of duality 38 P r o b a b i l i t y theory 100 Proof by contradiction {redactin ad absurdum proof) 67 127 . 91 intersection 17. 15. 84 c o m m u t a t i v e 9f.

We would also be pleased to receive any other suggestions you may wish to make.Proposition (s) 55 biconditional 78 equivalence of 56. Moscow 1-110. 67 necessarily false 57 necessarily true 57 product (conjunction) of 58 sum (disjunction) of 57 Pythagorean theorem 66 product (intersection) sum (union) of 12 universal 19 Sufficient condition 66 Switch always-closed 91 always-open 91 Switching circuit(s) 89fi product of 90 sum of 90 Syllogism 86 Reciprocal of (multiplicative inverse) a number 10 / Self-dual 38 Set(s) 12 complement of 40 element of 12 empty (or void or null) Truth set 55 Truth table 106 Truth value 106 16 TO THE Venn diagram 13 READER Mir Publishers would be grateful for your comments on the content. GSP Pervy Rizhsky Pereulok. 129820. translation and design of this book. 2 Printed in the Union of Soviet Socialist Republics . Our address is: USSR.

.

There are also many exercises with answers and hints to some of them placed at the end of the book. in particular. The book is intended for senior pupils and can also be of use for adult readers interested in mathematics. Mir Publishers Moscow . Some applications of this algebra to the automation of mathematical proofs are discussed. it deals with algebra of propositions.This book is devoted to fundamentals of the theory of Boolean algebras which play an important role in mathematical logic and in the development of electronic computers and cybernetics. The book includes the definition of the notion of a Boolean algebra and many examples of such algebras. The solution of the problems in the exercises facilitates the understanding of the material of the book.