Philosophy

by

Bertrand Russell

Originally published by

George Allen & Unwin, Ltd., London. May 1o1o.

Online Corrected Edition version 1.o (February ¸, zo1o),

based on the “second edition” (second printing) of April 1ozo,

incorporating additional corrections, marked in green.

ii

[Russell’s blurb from the original dustcover:]

This book is intended for those who have no previous acquaintance

with the topics of which it treats, and no more knowledge of mathe-

matics than can be acquired at a primary school or even at Eton. It

sets forth in elementary form the logical deﬁnition of number, the

analysis of the notion of order, the modern doctrine of the inﬁnite,

and the theory of descriptions and classes as symbolic ﬁctions. The

more controversial and uncertain aspects of the subject are subordi-

nated to those which can by now be regarded as acquired scientiﬁc

knowledge. These are explained without the use of symbols, but in

such a way as to give readers a general understanding of the methods

and purposes of mathematical logic, which, it is hoped, will be of

interest not only to those who wish to proceed to a more serious study

of the subject, but also to that wider circle who feel a desire to know

the bearings of this important modern science.

CONTENTS

Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

Editor’s Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi

I. The Series of Natural Numbers . . . . . . . . . . . . . 1

II. Deﬁnition of Number . . . . . . . . . . . . . . . . . . . 8

III. Finitude and Mathematical Induction . . . . . . . . . 1¸

IV. The Deﬁnition of Order . . . . . . . . . . . . . . . . . zz

V. Kinds of Relations . . . . . . . . . . . . . . . . . . . . . jz

VI. Similarity of Relations . . . . . . . . . . . . . . . . . . ju

VII. Rational, Real, and Complex Numbers . . . . . . . . . a¬

VIII. Inﬁnite Cardinal Numbers . . . . . . . . . . . . . . . . ¸8

IX. Inﬁnite Series and Ordinals . . . . . . . . . . . . . . . 6¬

X. Limits and Continuity . . . . . . . . . . . . . . . . . . ¬j

XI. Limits and Continuity of Functions . . . . . . . . . . . 8o

XII. Selections and the Multiplicative Axiom . . . . . . . . 88

XIII. The Axiomof Inﬁnity and Logical Types . . . . . . . . uu

XIV. Incompatibility and the Theory of Deduction . . . . . 1o8

XV. Propositional Functions . . . . . . . . . . . . . . . . . 116

XVI. Descriptions . . . . . . . . . . . . . . . . . . . . . . . . 1z¸

XVII. Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . 1j¸

XVIII. Mathematics and Logic . . . . . . . . . . . . . . . . . . 1a¸

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1¸a

Appendix: Changes to Online Edition . . . . . . . . . . . . . . 1¸¬

iii

PREFACE

v This book is intended essentially as an “Introduction,” and does not

aim at giving an exhaustive discussion of the problems with which

it deals. It seemed desirable to set forth certain results, hitherto

only available to those who have mastered logical symbolism, in a

form oﬀering the minimum of diﬃculty to the beginner. The utmost

endeavour has been made to avoid dogmatism on such questions

as are still open to serious doubt, and this endeavour has to some

extent dominated the choice of topics considered. The beginnings of

mathematical logic are less deﬁnitely known than its later portions,

but are of at least equal philosophical interest. Much of what is set

forth in the following chapters is not properly to be called “philoso-

phy,” though the matters concerned were included in philosophy so

long as no satisfactory science of them existed. The nature of inﬁnity

and continuity, for example, belonged in former days to philosophy,

but belongs now to mathematics. Mathematical philosophy, in the

strict sense, cannot, perhaps, be held to include such deﬁnite scien-

tiﬁc results as have been obtained in this region; the philosophy of

mathematics will naturally be expected to deal with questions on the

frontier of knowledge, as to which comparative certainty is not yet

attained. But speculation on such questions is hardly likely to be

fruitful unless the more scientiﬁc parts of the principles of mathe-

matics are known. A book dealing with those parts may, therefore,

claim to be an introduction to mathematical philosophy, though it

can hardly claim, except where it steps outside its province, to be

actually dealing with a part of philosophy. It does deal, | vi however,

with a body of knowledge which, to those who accept it, appears to

invalidate much traditional philosophy, and even a good deal of what

is current in the present day. In this way, as well as by its bearing on

still unsolved problems, mathematical logic is relevant to philosophy.

For this reason, as well as on account of the intrinsic importance of

the subject, some purpose may be served by a succinct account of

the main results of mathematical logic in a form requiring neither

a knowledge of mathematics nor an aptitude for mathematical sym-

iv

Preface v

bolism. Here, however, as elsewhere, the method is more important

than the results, from the point of view of further research; and the

method cannot well be explained within the framework of such a

book as the following. It is to be hoped that some readers may be

suﬃciently interested to advance to a study of the method by which

mathematical logic can be made helpful in investigating the tradi-

tional problems of philosophy. But that is a topic with which the

following pages have not attempted to deal.

BERTRAND RUSSELL.

EDITOR’S NOTE

vii [The note below was written by J. H. Muirhead, LL.D., editor of the

Library of Philosophy series in which Introduction to Mathematical

Philosophy was originally published.]

Those who, relying on the distinction between Mathematical Phi-

losophy and the Philosophy of Mathematics, think that this book

is out of place in the present Library, may be referred to what the

author himself says on this head in the Preface. It is not necessary

to agree with what he there suggests as to the readjustment of the

ﬁeld of philosophy by the transference from it to mathematics of such

problems as those of class, continuity, inﬁnity, in order to perceive the

bearing of the deﬁnitions and discussions that follow on the work of

“traditional philosophy.” If philosophers cannot consent to relegate

the criticism of these categories to any of the special sciences, it is

essential, at any rate, that they should know the precise meaning that

the science of mathematics, in which these concepts play so large a

part, assigns to them. If, on the other hand, there be mathematicians

to whom these deﬁnitions and discussions seem to be an elabora-

tion and complication of the simple, it may be well to remind them

from the side of philosophy that here, as elsewhere, apparent simplic-

ity may conceal a complexity which it is the business of somebody,

whether philosopher or mathematician, or, like the author of this

volume, both in one, to unravel.

vi

CHAPTER I

THE SERIES OF NATURAL NUMBERS

1 Mathematics is a study which, when we start from its most familiar

portions, may be pursued in either of two opposite directions. The

more familiar direction is constructive, towards gradually increas-

ing complexity: from integers to fractions, real numbers, complex

numbers; from addition and multiplication to diﬀerentiation and in-

tegration, and on to higher mathematics. The other direction, which

is less familiar, proceeds, by analysing, to greater and greater abstract-

ness and logical simplicity; instead of asking what can be deﬁned and

deduced from what is assumed to begin with, we ask instead what

more general ideas and principles can be found, in terms of which

what was our starting-point can be deﬁned or deduced. It is the

fact of pursuing this opposite direction that characterises mathemat-

ical philosophy as opposed to ordinary mathematics. But it should

be understood that the distinction is one, not in the subject matter,

but in the state of mind of the investigator. Early Greek geometers,

passing from the empirical rules of Egyptian land-surveying to the

general propositions by which those rules were found to be justiﬁable,

and thence to Euclid’s axioms and postulates, were engaged in math-

ematical philosophy, according to the above deﬁnition; but when

once the axioms and postulates had been reached, their deductive

employment, as we ﬁnd it in Euclid, belonged to mathematics in the

| z ordinary sense. The distinction between mathematics and mathe-

matical philosophy is one which depends upon the interest inspiring

the research, and upon the stage which the research has reached; not

upon the propositions with which the research is concerned.

We may state the same distinction in another way. The most obvi-

ous and easy things in mathematics are not those that come logically

at the beginning; they are things that, fromthe point of viewof logical

deduction, come somewhere in the middle. Just as the easiest bodies

to see are those that are neither very near nor very far, neither very

small nor very great, so the easiest conceptions to grasp are those that

are neither very complex nor very simple (using “simple” in a logical

sense). And as we need two sorts of instruments, the telescope and

1

Chap. I. The Series of Natural Numbers z

the microscope, for the enlargement of our visual powers, so we need

two sorts of instruments for the enlargement of our logical powers,

one to take us forward to the higher mathematics, the other to take us

backward to the logical foundations of the things that we are inclined

to take for granted in mathematics. We shall ﬁnd that by analysing

our ordinary mathematical notions we acquire fresh insight, new

powers, and the means of reaching whole new mathematical sub-

jects by adopting fresh lines of advance after our backward journey.

It is the purpose of this book to explain mathematical philosophy

simply and untechnically, without enlarging upon those portions

which are so doubtful or diﬃcult that an elementary treatment is

scarcely possible. A full treatment will be found in Principia Mathe-

matica;

1

the treatment in the present volume is intended merely as

an introduction.

To the average educated person of the present day, the obvious

starting-point of mathematics would be the series of whole numbers,

1, z, j, ±, . . . etc. |

j Probably only a person with some mathematical knowledge would

think of beginning with o instead of with 1, but we will presume this

degree of knowledge; we will take as our starting-point the series:

o, 1, z, j, . . . n, n +1, . . .

and it is this series that we shall mean when we speak of the “series

of natural numbers.”

It is only at a high stage of civilisation that we could take this series

as our starting-point. It must have required many ages to discover

that a brace of pheasants and a couple of days were both instances of

the number z: the degree of abstraction involved is far fromeasy. And

the discovery that 1 is a number must have been diﬃcult. As for o, it

is a very recent addition; the Greeks and Romans had no such digit.

If we had been embarking upon mathematical philosophy in earlier

days, we should have had to start with something less abstract than

the series of natural numbers, which we should reach as a stage on

our backward journey. When the logical foundations of mathematics

have grown more familiar, we shall be able to start further back, at

what is now a late stage in our analysis. But for the moment the

natural numbers seem to represent what is easiest and most familiar

in mathematics.

But though familiar, they are not understood. Very few people are

prepared with a deﬁnition of what is meant by “number,” or “o,” or

“1.” It is not very diﬃcult to see that, starting from o, any other of the

1

Cambridge University Press, vol. i., 1o1o; vol. ii., 1o1z; vol. iii., 1o1j. By

Whitehead and Russell.

Chap. I. The Series of Natural Numbers j

natural numbers can be reached by repeated additions of 1, but we

shall have to deﬁne what we mean by “adding 1,” and what we mean

by “repeated.” These questions are by no means easy. It was believed

until recently that some, at least, of these ﬁrst notions of arithmetic

must be accepted as too simple and primitive to be deﬁned. Since all

terms that are deﬁned are deﬁned by means of other terms, it is clear

that human knowledge must always be content to accept some terms

as intelligible without deﬁnition, in order | ± to have a starting-point

for its deﬁnitions. It is not clear that there must be terms which

are incapable of deﬁnition: it is possible that, however far back we

go in deﬁning, we always might go further still. On the other hand,

it is also possible that, when analysis has been pushed far enough,

we can reach terms that really are simple, and therefore logically

incapable of the sort of deﬁnition that consists in analysing. This is a

question which it is not necessary for us to decide; for our purposes

it is suﬃcient to observe that, since human powers are ﬁnite, the

deﬁnitions known to us must always begin somewhere, with terms

undeﬁned for the moment, though perhaps not permanently.

All traditional pure mathematics, including analytical geometry,

may be regarded as consisting wholly of propositions about the natu-

ral numbers. That is to say, the terms which occur can be deﬁned by

means of the natural numbers, and the propositions can be deduced

from the properties of the natural numbers—with the addition, in

each case, of the ideas and propositions of pure logic.

That all traditional pure mathematics can be derived from the

natural numbers is a fairly recent discovery, though it had long been

suspected. Pythagoras, who believed that not only mathematics, but

everything else could be deduced from numbers, was the discoverer

of the most serious obstacle in the way of what is called the “arith-

metising” of mathematics. It was Pythagoras who discovered the

existence of incommensurables, and, in particular, the incommen-

surability of the side of a square and the diagonal. If the length of

the side is 1 inch, the number of inches in the diagonal is the square

root of z, which appeared not to be a number at all. The problem

thus raised was solved only in our own day, and was only solved

completely by the help of the reduction of arithmetic to logic, which

will be explained in following chapters. For the present, we shall take

for granted the arithmetisation of mathematics, though this was a

feat of the very greatest importance. |

Having ¸ reduced all traditional pure mathematics to the theory

of the natural numbers, the next step in logical analysis was to re-

duce this theory itself to the smallest set of premisses and undeﬁned

terms from which it could be derived. This work was accomplished

by Peano. He showed that the entire theory of the natural numbers

could be derived from three primitive ideas and ﬁve primitive propo-

Chap. I. The Series of Natural Numbers ±

sitions in addition to those of pure logic. These three ideas and ﬁve

propositions thus became, as it were, hostages for the whole of tra-

ditional pure mathematics. If they could be deﬁned and proved in

terms of others, so could all pure mathematics. Their logical “weight,”

if one may use such an expression, is equal to that of the whole series

of sciences that have been deduced from the theory of the natural

numbers; the truth of this whole series is assured if the truth of the

ﬁve primitive propositions is guaranteed, provided, of course, that

there is nothing erroneous in the purely logical apparatus which is

also involved. The work of analysing mathematics is extraordinarily

facilitated by this work of Peano’s.

The three primitive ideas in Peano’s arithmetic are:

o, number, successor.

By “successor” he means the next number in the natural order. That

is to say, the successor of o is 1, the successor of 1 is z, and so on.

By “number” he means, in this connection, the class of the natural

numbers.

z

He is not assuming that we know all the members of this

class, but only that we know what we mean when we say that this or

that is a number, just as we know what we mean when we say “Jones

is a man,” though we do not know all men individually.

The ﬁve primitive propositions which Peano assumes are:

(1) o is a number.

(z) The successor of any number is a number.

(j) No two numbers have the same successor. |

(±) 6 o is not the successor of any number.

(¸) Any property which belongs to o, and also to the successor of

every number which has the property, belongs to all num-

bers.

The last of these is the principle of mathematical induction. We shall

have much to say concerning mathematical induction in the sequel;

for the present, we are concerned with it only as it occurs in Peano’s

analysis of arithmetic.

Let us consider brieﬂy the kind of way in which the theory of the

natural numbers results from these three ideas and ﬁve propositions.

To begin with, we deﬁne 1 as “the successor of o,” z as “the successor

of 1,” and so on. We can obviously go on as long as we like with

these deﬁnitions, since, in virtue of (z), every number that we reach

will have a successor, and, in virtue of (j), this cannot be any of the

numbers already deﬁned, because, if it were, two diﬀerent numbers

would have the same successor; and in virtue of (±) none of the

z

We shall use “number” in this sense in the present chapter. Afterwards the

word will be used in a more general sense.

Chap. I. The Series of Natural Numbers ¸

numbers we reach in the series of successors can be o. Thus the series

of successors gives us an endless series of continually new numbers.

In virtue of (¸) all numbers come in this series, which begins with

o and travels on through successive successors: for (a) o belongs to

this series, and (b) if a number n belongs to it, so does its successor,

whence, by mathematical induction, every number belongs to the

series.

Suppose we wish to deﬁne the sum of two numbers. Taking any

number m, we deﬁne m+o as m, and m+(n +1) as the successor of

m+ n. In virtue of (¸) this gives a deﬁnition of the sum of m and

n, whatever number n may be. Similarly we can deﬁne the product

of any two numbers. The reader can easily convince himself that

any ordinary elementary proposition of arithmetic can be proved by

means of our ﬁve premisses, and if he has any diﬃculty he can ﬁnd

the proof in Peano.

It is time now to turn to the considerations which make it nec-

essary to advance beyond the standpoint of Peano, who | ; represents

the last perfection of the “arithmetisation” of mathematics, to that

of Frege, who ﬁrst succeeded in “logicising” mathematics, i.e. in re-

ducing to logic the arithmetical notions which his predecessors had

shown to be suﬃcient for mathematics. We shall not, in this chapter,

actually give Frege’s deﬁnition of number and of particular numbers,

but we shall give some of the reasons why Peano’s treatment is less

ﬁnal than it appears to be.

In the ﬁrst place, Peano’s three primitive ideas—namely, “o,”

“number,” and “successor”—are capable of an inﬁnite number of

diﬀerent interpretations, all of which will satisfy the ﬁve primitive

propositions. We will give some examples.

(1) Let “o” be taken to mean 1oo, and let “number” be taken to

mean the numbers from 1oo onward in the series of natural numbers.

Then all our primitive propositions are satisﬁed, even the fourth, for,

though 1oo is the successor of oo, oo is not a “number” in the sense

which we are now giving to the word “number.” It is obvious that any

number may be substituted for 1oo in this example.

(z) Let “o” have its usual meaning, but let “number” mean what

we usually call “even numbers,” and let the “successor” of a number

be what results from adding two to it. Then “1” will stand for the

number two, “z” will stand for the number four, and so on; the series

of “numbers” now will be

o, two, four, six, eight . . .

All Peano’s ﬁve premisses are satisﬁed still.

(j) Let “o” mean the number one, let “number” mean the set

1,

1

z

,

1

±

,

1

8

,

1

16

, . . .

Chap. I. The Series of Natural Numbers 6

and let “successor” mean “half.” Then all Peano’s ﬁve axioms will be

true of this set.

It is clear that such examples might be multiplied indeﬁnitely. In

fact, given any series

x

o

, x

1

, x

z

, x

j

, . . . x

n

, . . . |

8 which is endless, contains no repetitions, has a beginning, and has no

terms that cannot be reached from the beginning in a ﬁnite number

of steps, we have a set of terms verifying Peano’s axioms. This is easily

seen, though the formal proof is somewhat long. Let “o” mean x

o

, let

“number” mean the whole set of terms, and let the “successor” of x

n

mean x

n+1

. Then

(1) “o is a number,” i.e. x

o

is a member of the set.

(z) “The successor of any number is a number,” i.e. taking any

term x

n

in the set, x

n+1

is also in the set.

(j) “No two numbers have the same successor,” i.e. if x

m

and x

n

are two diﬀerent members of the set, x

m+1

and x

n+1

are diﬀerent; this

results from the fact that (by hypothesis) there are no repetitions in

the set.

(±) “o is not the successor of any number,” i.e. no term in the set

comes before x

o

.

(¸) This becomes: Any property which belongs to x

o

, and belongs

to x

n+1

provided it belongs to x

n

, belongs to all the x’s.

This follows from the corresponding property for numbers.

A series of the form

x

o

, x

1

, x

z

, . . . x

n

, . . .

in which there is a ﬁrst term, a successor to each term (so that there is

no last term), no repetitions, and every term can be reached from the

start in a ﬁnite number of steps, is called a progression. Progressions

are of great importance in the principles of mathematics. As we have

just seen, every progression veriﬁes Peano’s ﬁve axioms. It can be

proved, conversely, that every series which veriﬁes Peano’s ﬁve axioms

is a progression. Hence these ﬁve axioms may be used to deﬁne the

class of progressions: “progressions” are “those series which verify

these ﬁve axioms.” Any progression may be taken as the basis of pure

mathematics: we may give the name “o” to its ﬁrst term, the name

“number” to the whole set of its terms, and the name “successor” to

the next in the progression. The progression need not be composed

of numbers: it may be | o composed of points in space, or moments of

time, or any other terms of which there is an inﬁnite supply. Each

diﬀerent progression will give rise to a diﬀerent interpretation of all

the propositions of traditional pure mathematics; all these possible

interpretations will be equally true.

Chap. I. The Series of Natural Numbers ;

In Peano’s system there is nothing to enable us to distinguish

between these diﬀerent interpretations of his primitive ideas. It is

assumed that we know what is meant by “o,” and that we shall not

suppose that this symbol means 1oo or Cleopatra’s Needle or any of

the other things that it might mean.

This point, that “o” and “number” and “successor” cannot be

deﬁned by means of Peano’s ﬁve axioms, but must be independently

understood, is important. We want our numbers not merely to verify

mathematical formulæ, but to apply in the right way to common

objects. We want to have ten ﬁngers and two eyes and one nose. A

system in which “1” meant 1oo, and “z” meant 1o1, and so on, might

be all right for pure mathematics, but would not suit daily life. We

want “o” and “number” and “successor” to have meanings which

will give us the right allowance of ﬁngers and eyes and noses. We

have already some knowledge (though not suﬃciently articulate or

analytic) of what we mean by “1” and “z” and so on, and our use of

numbers in arithmetic must conform to this knowledge. We cannot

secure that this shall be the case by Peano’s method; all that we can

do, if we adopt his method, is to say “we know what we mean by

‘o’ and ‘number’ and ‘successor,’ though we cannot explain what we

mean in terms of other simpler concepts.” It is quite legitimate to say

this when we must, and at some point we all must; but it is the object

of mathematical philosophy to put oﬀ saying it as long as possible.

By the logical theory of arithmetic we are able to put it oﬀ for a very

long time.

It might be suggested that, instead of setting up “o” and “number”

and “successor” as terms of which we know the meaning although

we cannot deﬁne them, we might let them | 1o stand for any three terms

that verify Peano’s ﬁve axioms. They will then no longer be terms

which have a meaning that is deﬁnite though undeﬁned: they will

be “variables,” terms concerning which we make certain hypotheses,

namely, those stated in the ﬁve axioms, but which are otherwise

undetermined. If we adopt this plan, our theorems will not be proved

concerning an ascertained set of terms called “the natural numbers,”

but concerning all sets of terms having certain properties. Such a

procedure is not fallacious; indeed for certain purposes it represents a

valuable generalisation. But from two points of view it fails to give an

adequate basis for arithmetic. In the ﬁrst place, it does not enable us

to know whether there are any sets of terms verifying Peano’s axioms;

it does not even give the faintest suggestion of any way of discovering

whether there are such sets. In the second place, as already observed,

we want our numbers to be such as can be used for counting common

objects, and this requires that our numbers should have a deﬁnite

meaning, not merely that they should have certain formal properties.

This deﬁnite meaning is deﬁned by the logical theory of arithmetic.

CHAPTER II

DEFINITION OF NUMBER

11 The question “What is a number?” is one which has been often asked,

but has only been correctly answered in our own time. The answer

was given by Frege in 188±, in his Grundlagen der Arithmetik.

1

Al-

though this book is quite short, not diﬃcult, and of the very highest

importance, it attracted almost no attention, and the deﬁnition of

number which it contains remained practically unknown until it was

rediscovered by the present author in 1oo1.

In seeking a deﬁnition of number, the ﬁrst thing to be clear about

is what we may call the grammar of our inquiry. Many philosophers,

when attempting to deﬁne number, are really setting to work to

deﬁne plurality, which is quite a diﬀerent thing. Number is what is

characteristic of numbers, as man is what is characteristic of men. A

plurality is not an instance of number, but of some particular number.

A trio of men, for example, is an instance of the number j, and the

number j is an instance of number; but the trio is not an instance

of number. This point may seem elementary and scarcely worth

mentioning; yet it has proved too subtle for the philosophers, with

few exceptions.

A particular number is not identical with any collection of terms

having that number: the number j is not identical with | 1z the trio

consisting of Brown, Jones, and Robinson. The number j is something

which all trios have in common, and which distinguishes them from

other collections. A number is something that characterises certain

collections, namely, those that have that number.

Instead of speaking of a “collection,” we shall as a rule speak of

a “class,” or sometimes a “set.” Other words used in mathematics

for the same thing are “aggregate” and “manifold.” We shall have

much to say later on about classes. For the present, we will say as

little as possible. But there are some remarks that must be made

immediately.

A class or collection may be deﬁned in two ways that at ﬁrst sight

seem quite distinct. We may enumerate its members, as when we say,

1

The same answer is given more fully and with more development in his

Grundgesetze der Arithmetik, vol. i., 18oj.

8

Chap. II. Deﬁnition of Number o

“The collection I mean is Brown, Jones, and Robinson.” Or we may

mention a deﬁning property, as when we speak of “mankind” or “the

inhabitants of London.” The deﬁnition which enumerates is called

a deﬁnition by “extension,” and the one which mentions a deﬁning

property is called a deﬁnition by “intension.” Of these two kinds of

deﬁnition, the one by intension is logically more fundamental. This

is shown by two considerations: (1) that the extensional deﬁnition

can always be reduced to an intensional one; (z) that the intensional

one often cannot even theoretically be reduced to the extensional one.

Each of these points needs a word of explanation.

(1) Brown, Jones, and Robinson all of them possess a certain

property which is possessed by nothing else in the whole universe,

namely, the property of being either Brown or Jones or Robinson.

This property can be used to give a deﬁnition by intension of the

class consisting of Brown and Jones and Robinson. Consider such a

formula as “x is Brown or x is Jones or x is Robinson.” This formula

will be true for just three x’s, namely, Brown and Jones and Robinson.

In this respect it resembles a cubic equation with its three roots. It

may be taken as assigning a property common to the members of

the class consisting of these three | 1j men, and peculiar to them. A

similar treatment can obviously be applied to any other class given

in extension.

(z) It is obvious that in practice we can often know a great deal

about a class without being able to enumerate its members. No one

man could actually enumerate all men, or even all the inhabitants

of London, yet a great deal is known about each of these classes.

This is enough to show that deﬁnition by extension is not necessary

to knowledge about a class. But when we come to consider inﬁnite

classes, we ﬁnd that enumeration is not even theoretically possible

for beings who only live for a ﬁnite time. We cannot enumerate all

the natural numbers: they are o, 1, z, j, and so on. At some point

we must content ourselves with “and so on.” We cannot enumerate

all fractions or all irrational numbers, or all of any other inﬁnite

collection. Thus our knowledge in regard to all such collections can

only be derived from a deﬁnition by intension.

These remarks are relevant, when we are seeking the deﬁnition

of number, in three diﬀerent ways. In the ﬁrst place, numbers them-

selves form an inﬁnite collection, and cannot therefore be deﬁned

by enumeration. In the second place, the collections having a given

number of terms themselves presumably form an inﬁnite collection:

it is to be presumed, for example, that there are an inﬁnite collection

of trios in the world, for if this were not the case the total number of

things in the world would be ﬁnite, which, though possible, seems

unlikely. In the third place, we wish to deﬁne “number” in such a

way that inﬁnite numbers may be possible; thus we must be able to

Chap. II. Deﬁnition of Number 1o

speak of the number of terms in an inﬁnite collection, and such a

collection must be deﬁned by intension, i.e. by a property common to

all its members and peculiar to them.

For many purposes, a class and a deﬁning characteristic of it are

practically interchangeable. The vital diﬀerence between the two

consists in the fact that there is only one class having a given set of

members, whereas there are always many diﬀerent characteristics

by which a given class may be deﬁned. Men | 1± may be deﬁned as

featherless bipeds, or as rational animals, or (more correctly) by the

traits by which Swift delineates the Yahoos. It is this fact that a

deﬁning characteristic is never unique which makes classes useful;

otherwise we could be content with the properties common and

peculiar to their members.

z

Any one of these properties can be used

in place of the class whenever uniqueness is not important.

Returning now to the deﬁnition of number, it is clear that number

is a way of bringing together certain collections, namely, those that

have a given number of terms. We can suppose all couples in one

bundle, all trios in another, and so on. In this way we obtain various

bundles of collections, each bundle consisting of all the collections

that have a certain number of terms. Each bundle is a class whose

members are collections, i.e. classes; thus each is a class of classes.

The bundle consisting of all couples, for example, is a class of classes:

each couple is a class with two members, and the whole bundle of

couples is a class with an inﬁnite number of members, each of which

is a class of two members.

How shall we decide whether two collections are to belong to the

same bundle? The answer that suggests itself is: “Find out how many

members each has, and put them in the same bundle if they have

the same number of members.” But this presupposes that we have

deﬁned numbers, and that we know how to discover how many terms

a collection has. We are so used to the operation of counting that

such a presupposition might easily pass unnoticed. In fact, however,

counting, though familiar, is logically a very complex operation;

moreover it is only available, as a means of discovering how many

terms a collection has, when the collection is ﬁnite. Our deﬁnition

of number must not assume in advance that all numbers are ﬁnite;

and we cannot in any case, without a vicious circle, | 1¸ use counting

to deﬁne numbers, because numbers are used in counting. We need,

therefore, some other method of deciding when two collections have

the same number of terms.

In actual fact, it is simpler logically to ﬁnd out whether two col-

lections have the same number of terms than it is to deﬁne what

z

As will be explained later, classes may be regarded as logical ﬁctions, manu-

factured out of deﬁning characteristics. But for the present it will simplify our

exposition to treat classes as if they were real.

Chap. II. Deﬁnition of Number 11

that number is. An illustration will make this clear. If there were

no polygamy or polyandry anywhere in the world, it is clear that the

number of husbands living at any moment would be exactly the same

as the number of wives. We do not need a census to assure us of this,

nor do we need to know what is the actual number of husbands and

of wives. We know the number must be the same in both collections,

because each husband has one wife and each wife has one husband.

The relation of husband and wife is what is called “one-one.”

A relation is said to be “one-one” when, if x has the relation in

question to y, no other term x

**has the same relation to y, and x does
**

not have the same relation to any term y

**other than y. When only the
**

ﬁrst of these two conditions is fulﬁlled, the relation is called “one-

many”; when only the second is fulﬁlled, it is called “many-one.” It

should be observed that the number 1 is not used in these deﬁnitions.

In Christian countries, the relation of husband to wife is one-one;

in Mahometan countries it is one-many; in Tibet it is many-one. The

relation of father to son is one-many; that of son to father is many-one,

but that of eldest son to father is one-one. If n is any number, the

relation of n to n +1 is one-one; so is the relation of n to zn or to jn.

When we are considering only positive numbers, the relation of n to

n

z

is one-one; but when negative numbers are admitted, it becomes

two-one, since n and −n have the same square. These instances should

suﬃce to make clear the notions of one-one, one-many, and many-one

relations, which play a great part in the principles of mathematics,

not only in relation to the deﬁnition of numbers, but in many other

connections.

Two classes are said to be “similar” when there is a one-one |

16 relation which correlates the terms of the one class each with one

term of the other class, in the same manner in which the relation of

marriage correlates husbands with wives. A few preliminary deﬁni-

tions will help us to state this deﬁnition more precisely. The class of

those terms that have a given relation to something or other is called

the domain of that relation: thus fathers are the domain of the relation

of father to child, husbands are the domain of the relation of husband

to wife, wives are the domain of the relation of wife to husband,

and husbands and wives together are the domain of the relation of

marriage. The relation of wife to husband is called the converse of the

relation of husband to wife. Similarly less is the converse of greater,

later is the converse of earlier, and so on. Generally, the converse of a

given relation is that relation which holds between y and x whenever

the given relation holds between x and y. The converse domain of a

relation is the domain of its converse: thus the class of wives is the

converse domain of the relation of husband to wife. We may now

state our deﬁnition of similarity as follows:—

One class is said to be “similar” to another when there is a one-one

Chap. II. Deﬁnition of Number 1z

relation of which the one class is the domain, while the other is the converse

domain.

It is easy to prove (1) that every class is similar to itself, (z) that if

a class α is similar to a class β, then β is similar to α, (j) that if α is

similar to β and β to γ, then α is similar to γ. A relation is said to be

reﬂexive when it possesses the ﬁrst of these properties, symmetrical

when it possesses the second, and transitive when it possesses the

third. It is obvious that a relation which is symmetrical and transitive

must be reﬂexive throughout its domain. Relations which possess

these properties are an important kind, and it is worth while to note

that similarity is one of this kind of relations.

It is obvious to common sense that two ﬁnite classes have the

same number of terms if they are similar, but not otherwise. The act

of counting consists in establishing a one-one correlation | 1; between

the set of objects counted and the natural numbers (excluding o) that

are used up in the process. Accordingly common sense concludes

that there are as many objects in the set to be counted as there are

numbers up to the last number used in the counting. And we also

know that, so long as we conﬁne ourselves to ﬁnite numbers, there

are just n numbers from 1 up to n. Hence it follows that the last

number used in counting a collection is the number of terms in the

collection, provided the collection is ﬁnite. But this result, besides

being only applicable to ﬁnite collections, depends upon and assumes

the fact that two classes which are similar have the same number

of terms; for what we do when we count (say) 1o objects is to show

that the set of these objects is similar to the set of numbers 1 to 1o.

The notion of similarity is logically presupposed in the operation of

counting, and is logically simpler though less familiar. In counting,

it is necessary to take the objects counted in a certain order, as ﬁrst,

second, third, etc., but order is not of the essence of number: it is

an irrelevant addition, an unnecessary complication from the logical

point of view. The notion of similarity does not demand an order:

for example, we saw that the number of husbands is the same as the

number of wives, without having to establish an order of precedence

among them. The notion of similarity also does not require that the

classes which are similar should be ﬁnite. Take, for example, the

natural numbers (excluding o) on the one hand, and the fractions

which have 1 for their numerator on the other hand: it is obvious that

we can correlate z with 1/z, j with 1/j, and so on, thus proving that

the two classes are similar.

We may thus use the notion of “similarity” to decide when two

collections are to belong to the same bundle, in the sense in which

we were asking this question earlier in this chapter. We want to make

one bundle containing the class that has no members: this will be for

the number o. Then we want a bundle of all the classes that have one

Chap. II. Deﬁnition of Number 1j

member: this will be for the number 1. Then, for the number z, we

want a bundle consisting | 18 of all couples; then one of all trios; and so

on. Given any collection, we can deﬁne the bundle it is to belong to

as being the class of all those collections that are “similar” to it. It is

very easy to see that if (for example) a collection has three members,

the class of all those collections that are similar to it will be the class

of trios. And whatever number of terms a collection may have, those

collections that are “similar” to it will have the same number of terms.

We may take this as a deﬁnition of “having the same number of terms.”

It is obvious that it gives results conformable to usage so long as we

conﬁne ourselves to ﬁnite collections.

So far we have not suggested anything in the slightest degree

paradoxical. But when we come to the actual deﬁnition of numbers

we cannot avoid what must at ﬁrst sight seem a paradox, though

this impression will soon wear oﬀ. We naturally think that the class

of couples (for example) is something diﬀerent from the number z.

But there is no doubt about the class of couples: it is indubitable

and not diﬃcult to deﬁne, whereas the number z, in any other sense,

is a metaphysical entity about which we can never feel sure that it

exists or that we have tracked it down. It is therefore more prudent

to content ourselves with the class of couples, which we are sure of,

than to hunt for a problematical number z which must always remain

elusive. Accordingly we set up the following deﬁnition:—

The number of a class is the class of all those classes that are similar to

it.

Thus the number of a couple will be the class of all couples. In

fact, the class of all couples will be the number z, according to our

deﬁnition. At the expense of a little oddity, this deﬁnition secures

deﬁniteness and indubitableness; and it is not diﬃcult to prove that

numbers so deﬁned have all the properties that we expect numbers

to have.

We may now go on to deﬁne numbers in general as any one of the

bundles into which similarity collects classes. A number will be a set

of classes such as that any two are similar to each | 1o other, and none

outside the set are similar to any inside the set. In other words, a

number (in general) is any collection which is the number of one of

its members; or, more simply still:

A number is anything which is the number of some class.

Such a deﬁnition has a verbal appearance of being circular, but in

fact it is not. We deﬁne “the number of a given class” without using

the notion of number in general; therefore we may deﬁne number in

general in terms of “the number of a given class” without committing

any logical error.

Deﬁnitions of this sort are in fact very common. The class of

fathers, for example, would have to be deﬁned by ﬁrst deﬁning what

Chap. II. Deﬁnition of Number 1±

it is to be the father of somebody; then the class of fathers will be

all those who are somebody’s father. Similarly if we want to deﬁne

square numbers (say), we must ﬁrst deﬁne what we mean by saying

that one number is the square of another, and then deﬁne square

numbers as those that are the squares of other numbers. This kind

of procedure is very common, and it is important to realise that it is

legitimate and even often necessary.

We have now given a deﬁnition of numbers which will serve for

ﬁnite collections. It remains to be seen how it will serve for inﬁnite

collections. But ﬁrst we must decide what we mean by “ﬁnite” and

“inﬁnite,” which cannot be done within the limits of the present

chapter.

CHAPTER III

FINITUDE ANDMATHEMATICAL

INDUCTION

zo The series of natural numbers, as we saw in Chapter I., can all be

deﬁned if we know what we mean by the three terms “o,” “number,”

and “successor.” But we may go a step farther: we can deﬁne all the

natural numbers if we know what we mean by “o” and “successor.” It

will help us to understand the diﬀerence between ﬁnite and inﬁnite

to see how this can be done, and why the method by which it is done

cannot be extended beyond the ﬁnite. We will not yet consider how

“o” and “successor” are to be deﬁned: we will for the moment assume

that we know what these terms mean, and show how thence all other

natural numbers can be obtained.

It is easy to see that we can reach any assigned number, say jo,ooo.

We ﬁrst deﬁne “1” as “the successor of o,” then we deﬁne “z” as

“the successor of 1,” and so on. In the case of an assigned number,

such as jo,ooo, the proof that we can reach it by proceeding step by

step in this fashion may be made, if we have the patience, by actual

experiment: we can go on until we actually arrive at jo,ooo. But

although the method of experiment is available for each particular

natural number, it is not available for proving the general proposition

that all such numbers can be reached in this way, i.e. by proceeding

from o step by step from each number to its successor. Is there any

other way by which this can be proved?

Let us consider the question the other way round. What are the

numbers that can be reached, given the terms “o” and | z1 “successor”?

Is there any way by which we can deﬁne the whole class of such

numbers? We reach 1, as the successor of o; z, as the successor of

1; j, as the successor of z; and so on. It is this “and so on” that we

wish to replace by something less vague and indeﬁnite. We might be

tempted to say that “and so on” means that the process of proceeding

to the successor may be repeated any ﬁnite number of times; but

the problem upon which we are engaged is the problem of deﬁning

“ﬁnite number,” and therefore we must not use this notion in our

deﬁnition. Our deﬁnition must not assume that we know what a

ﬁnite number is.

1¸

Chap. III. Finitude and Mathematical Induction 16

The key to our problem lies in mathematical induction. It will be

remembered that, in Chapter I., this was the ﬁfth of the ﬁve primitive

propositions which we laid down about the natural numbers. It stated

that any property which belongs to o, and to the successor of any

number which has the property, belongs to all the natural numbers.

This was then presented as a principle, but we shall now adopt it as a

deﬁnition. It is not diﬃcult to see that the terms obeying it are the

same as the numbers that can be reached from o by successive steps

from next to next, but as the point is important we will set forth the

matter in some detail.

We shall do well to begin with some deﬁnitions, which will be

useful in other connections also.

A property is said to be “hereditary” in the natural-number series

if, whenever it belongs to a number n, it also belongs to n + 1, the

successor of n. Similarly a class is said to be “hereditary” if, whenever

n is a member of the class, so is n + 1. It is easy to see, though we

are not yet supposed to know, that to say a property is hereditary is

equivalent to saying that it belongs to all the natural numbers not less

than some one of them, e.g. it must belong to all that are not less than

1oo, or all that are not less than 1ooo, or it may be that it belongs to

all that are not less than o, i.e. to all without exception.

A property is said to be “inductive” when it is a hereditary |

zz property which belongs to o. Similarly a class is “inductive” when it

is a hereditary class of which o is a member.

Given a hereditary class of which o is a member, it follows that 1

is a member of it, because a hereditary class contains the successors of

its members, and 1 is the successor of o. Similarly, given a hereditary

class of which 1 is a member, it follows that z is a member of it; and so

on. Thus we can prove by a step-by-step procedure that any assigned

natural number, say jo,ooo, is a member of every inductive class.

We will deﬁne the “posterity” of a given natural number with re-

spect to the relation “immediate predecessor” (which is the converse

of “successor”) as all those terms that belong to every hereditary class

to which the given number belongs. It is again easy to see that the

posterity of a natural number consists of itself and all greater natural

numbers; but this also we do not yet oﬃcially know.

By the above deﬁnitions, the posterity of o will consist of those

terms which belong to every inductive class.

It is now not diﬃcult to make it obvious that the posterity of o is

the same set as those terms that can be reached from o by successive

steps from next to next. For, in the ﬁrst place, o belongs to both these

sets (in the sense in which we have deﬁned our terms); in the second

place, if n belongs to both sets, so does n + 1. It is to be observed

that we are dealing here with the kind of matter that does not admit

of precise proof, namely, the comparison of a relatively vague idea

Chap. III. Finitude and Mathematical Induction 1;

with a relatively precise one. The notion of “those terms that can

be reached from o by successive steps from next to next” is vague,

though it seems as if it conveyed a deﬁnite meaning; on the other

hand, “the posterity of o” is precise and explicit just where the other

idea is hazy. It may be taken as giving what we meant to mean when

we spoke of the terms that can be reached from o by successive steps.

We now lay down the following deﬁnition:—

The “natural numbers” are the posterity of o with respect to the |

zj relation “immediate predecessor” (which is the converse of “succes-

sor”).

We have thus arrived at a deﬁnition of one of Peano’s three primi-

tive ideas in terms of the other two. As a result of this deﬁnition, two

of his primitive propositions—namely, the one asserting that o is a

number and the one asserting mathematical induction—become un-

necessary, since they result fromthe deﬁnition. The one asserting that

the successor of a natural number is a natural number is only needed

in the weakened form “every natural number has a successor.”

We can, of course, easily deﬁne “o” and “successor” by means of

the deﬁnition of number in general which we arrived at in Chapter

II. The number o is the number of terms in a class which has no

members, i.e. in the class which is called the “null-class.” By the

general deﬁnition of number, the number of terms in the null-class is

the set of all classes similar to the null-class, i.e. (as is easily proved)

the set consisting of the null-class all alone, i.e. the class whose only

member is the null-class. (This is not identical with the null-class: it

has one member, namely, the null-class, whereas the null-class itself

has no members. A class which has one member is never identical

with that one member, as we shall explain when we come to the theory

of classes.) Thus we have the following purely logical deﬁnition:—

o is the class whose only member is the null-class.

It remains to deﬁne “successor.” Given any number n, let α be a

class which has n members, and let x be a term which is not a member

of α. Then the class consisting of α with x added on will have n +1

members. Thus we have the following deﬁnition:—

The successor of the number of terms in the class α is the number of

terms in the class consisting of α together with x, where x is any term not

belonging to the class.

Certain niceties are required to make this deﬁnition perfect, but

they need not concern us.

1

It will be remembered that we | z± have

already given (in Chapter II.) a logical deﬁnition of the number of

terms in a class, namely, we deﬁned it as the set of all classes that are

similar to the given class.

We have thus reduced Peano’s three primitive ideas to ideas of

logic: we have given deﬁnitions of them which make them deﬁnite,

1

See Principia Mathematica, vol. ii. ∗11o.

Chap. III. Finitude and Mathematical Induction 18

no longer capable of an inﬁnity of diﬀerent meanings, as they were

when they were only determinate to the extent of obeying Peano’s

ﬁve axioms. We have removed them from the fundamental apparatus

of terms that must be merely apprehended, and have thus increased

the deductive articulation of mathematics.

As regards the ﬁve primitive propositions, we have already suc-

ceeded in making two of them demonstrable by our deﬁnition of

“natural number.” How stands it with the remaining three? It is very

easy to prove that o is not the successor of any number, and that

the successor of any number is a number. But there is a diﬃculty

about the remaining primitive proposition, namely, “no two numbers

have the same successor.” The diﬃculty does not arise unless the

total number of individuals in the universe is ﬁnite; for given two

numbers m and n, neither of which is the total number of individuals

in the universe, it is easy to prove that we cannot have m+1 = n +1

unless we have m = n. But let us suppose that the total number of

individuals in the universe were (say) 1o; then there would be no

class of 11 individuals, and the number 11 would be the null-class.

So would the number 1z. Thus we should have 11 = 1z; therefore the

successor of 1o would be the same as the successor of 11, although

1o would not be the same as 11. Thus we should have two diﬀerent

numbers with the same successor. This failure of the third axiom

cannot arise, however, if the number of individuals in the world is

not ﬁnite. We shall return to this topic at a later stage.

z

Assuming that the number of individuals in the universe is not

ﬁnite, we have now succeeded not only in deﬁning Peano’s | z¸ three

primitive ideas, but in seeing how to prove his ﬁve primitive proposi-

tions, by means of primitive ideas and propositions belonging to logic.

It follows that all pure mathematics, in so far as it is deducible from

the theory of the natural numbers, is only a prolongation of logic.

The extension of this result to those modern branches of mathematics

which are not deducible fromthe theory of the natural numbers oﬀers

no diﬃculty of principle, as we have shown elsewhere.

j

The process of mathematical induction, by means of which we

deﬁned the natural numbers, is capable of generalisation. We deﬁned

the natural numbers as the “posterity” of o with respect to the rela-

tion of a number to its immediate successor. If we call this relation N,

any number m will have this relation to m+1. A property is “heredi-

tary with respect to N,” or simply “N-hereditary,” if, whenever the

property belongs to a number m, it also belongs to m+1, i.e. to the

number to which m has the relation N. And a number n will be said to

belong to the “posterity” of m with respect to the relation N if n has

z

See Chapter XIII.

j

For geometry, in so far as it is not purely analytical, see Principles of Mathematics,

part vi.; for rational dynamics, ibid., part vii.

Chap. III. Finitude and Mathematical Induction 1o

every N-hereditary property belonging to m. These deﬁnitions can

all be applied to any other relation just as well as to N. Thus if R is

any relation whatever, we can lay down the following deﬁnitions:

±

—

A property is called “R-hereditary” when, if it belongs to a term x,

and x has the relation R to y, then it belongs to y.

A class is R-hereditary when its deﬁning property is R-hereditary.

A term x is said to be an “R-ancestor” of the term y if y has every

R-hereditary property that x has, provided x is a term which has the

relation R to something or to which something has the relation R.

(This is only to exclude trivial cases.) |

The z6 “R-posterity” of x is all the terms of which x is an R-ancestor.

We have framed the above deﬁnitions so that if a term is the

ancestor of anything it is its own ancestor and belongs to its own

posterity. This is merely for convenience.

It will be observed that if we take for R the relation “parent,”

“ancestor” and “posterity” will have the usual meanings, except that

a person will be included among his own ancestors and posterity.

It is, of course, obvious at once that “ancestor” must be capable

of deﬁnition in terms of “parent,” but until Frege developed his

generalised theory of induction, no one could have deﬁned “ancestor”

precisely in terms of “parent.” A brief consideration of this point

will serve to show the importance of the theory. A person confronted

for the ﬁrst time with the problem of deﬁning “ancestor” in terms of

“parent” would naturally say that A is an ancestor of Z if, between

A and Z, there are a certain number of people, B, C, . . . , of whom B

is a child of A, each is a parent of the next, until the last, who is a

parent of Z. But this deﬁnition is not adequate unless we add that the

number of intermediate terms is to be ﬁnite. Take, for example, such

a series as the following:—

−1, −

1

z

, −

1

±

, −

1

8

, . . .

1

8

,

1

±

,

1

z

, 1.

Here we have ﬁrst a series of negative fractions with no end, and

then a series of positive fractions with no beginning. Shall we say

that, in this series, −1/8 is an ancestor of 1/8? It will be so according

to the beginner’s deﬁnition suggested above, but it will not be so

according to any deﬁnition which will give the kind of idea that

we wish to deﬁne. For this purpose, it is essential that the number

of intermediaries should be ﬁnite. But, as we saw, “ﬁnite” is to be

deﬁned by means of mathematical induction, and it is simpler to

deﬁne the ancestral relation generally at once than to deﬁne it ﬁrst

±

These deﬁnitions, and the generalised theory of induction, are due to Frege,

and were published so long ago as 18;o in his Begriﬀsschrift. In spite of the great

value of this work, I was, I believe, the ﬁrst person who ever read it—more than

twenty years after its publication.

Chap. III. Finitude and Mathematical Induction zo

only for the case of the relation of n to n +1, and then extend it to

other cases. Here, as constantly elsewhere, generality from the ﬁrst,

though it may | z; require more thought at the start, will be found in the

long run to economise thought and increase logical power.

The use of mathematical induction in demonstrations was, in the

past, something of a mystery. There seemed no reasonable doubt that

it was a valid method of proof, but no one quite knewwhy it was valid.

Some believed it to be really a case of induction, in the sense in which

that word is used in logic. Poincar´ e

¸

considered it to be a principle

of the utmost importance, by means of which an inﬁnite number of

syllogisms could be condensed into one argument. We now know

that all such views are mistaken, and that mathematical induction

is a deﬁnition, not a principle. There are some numbers to which it

can be applied, and there are others (as we shall see in Chapter VIII.)

to which it cannot be applied. We deﬁne the “natural numbers” as

those to which proofs by mathematical induction can be applied, i.e.

as those that possess all inductive properties. It follows that such

proofs can be applied to the natural numbers, not in virtue of any

mysterious intuition or axiom or principle, but as a purely verbal

proposition. If “quadrupeds” are deﬁned as animals having four legs,

it will follow that animals that have four legs are quadrupeds; and the

case of numbers that obey mathematical induction is exactly similar.

We shall use the phrase “inductive numbers” to mean the same

set as we have hitherto spoken of as the “natural numbers.” The

phrase “inductive numbers” is preferable as aﬀording a reminder that

the deﬁnition of this set of numbers is obtained from mathematical

induction.

Mathematical induction aﬀords, more than anything else, the

essential characteristic by which the ﬁnite is distinguished from the

inﬁnite. The principle of mathematical induction might be stated

popularly in some such form as “what can be inferred from next to

next can be inferred from ﬁrst to last.” This is true when the number

of intermediate steps between ﬁrst and last is ﬁnite, not otherwise.

Anyone who has ever | z8 watched a goods train beginning to move will

have noticed how the impulse is communicated with a jerk from each

truck to the next, until at last even the hindmost truck is in motion.

When the train is very long, it is a very long time before the last truck

moves. If the train were inﬁnitely long, there would be an inﬁnite

succession of jerks, and the time would never come when the whole

train would be in motion. Nevertheless, if there were a series of

trucks no longer than the series of inductive numbers (which, as we

shall see, is an instance of the smallest of inﬁnites), every truck would

begin to move sooner or later if the engine persevered, though there

¸

Science and Method, chap. iv.

Chap. III. Finitude and Mathematical Induction z1

would always be other trucks further back which had not yet begun

to move. This image will help to elucidate the argument from next

to next, and its connection with ﬁnitude. When we come to inﬁnite

numbers, where arguments from mathematical induction will be no

longer valid, the properties of such numbers will help to make clear,

by contrast, the almost unconscious use that is made of mathematical

induction where ﬁnite numbers are concerned.

CHAPTER IV

THE DEFINITION OF ORDER

zo We have now carried our analysis of the series of natural numbers to

the point where we have obtained logical deﬁnitions of the members

of this series, of the whole class of its members, and of the relation

of a number to its immediate successor. We must now consider the

serial character of the natural numbers in the order o, 1, z, j, . . . We

ordinarily think of the numbers as in this order, and it is an essential

part of the work of analysing our data to seek a deﬁnition of “order”

or “series” in logical terms.

The notion of order is one which has enormous importance in

mathematics. Not only the integers, but also rational fractions and

all real numbers have an order of magnitude, and this is essential

to most of their mathematical properties. The order of points on

a line is essential to geometry; so is the slightly more complicated

order of lines through a point in a plane, or of planes through a

line. Dimensions, in geometry, are a development of order. The

conception of a limit, which underlies all higher mathematics, is

a serial conception. There are parts of mathematics which do not

depend upon the notion of order, but they are very few in comparison

with the parts in which this notion is involved.

In seeking a deﬁnition of order, the ﬁrst thing to realise is that

no set of terms has just one order to the exclusion of others. A set of

terms has all the orders of which it is capable. Sometimes one order

is so much more familiar and natural to our | jo thoughts that we are

inclined to regard it as the order of that set of terms; but this is a

mistake. The natural numbers—or the “inductive” numbers, as we

shall also call them—occur to us most readily in order of magnitude;

but they are capable of an inﬁnite number of other arrangements.

We might, for example, consider ﬁrst all the odd numbers and then

all the even numbers; or ﬁrst 1, then all the even numbers, then

all the odd multiples of j, then all the multiples of ¸ but not of z

or j, then all the multiples of ; but not of z or j or ¸, and so on

through the whole series of primes. When we say that we “arrange”

the numbers in these various orders, that is an inaccurate expression:

zz

Chap. IV. The Deﬁnition of Order zj

what we really do is to turn our attention to certain relations between

the natural numbers, which themselves generate such-and-such an

arrangement. We can no more “arrange” the natural numbers than

we can the starry heavens; but just as we may notice among the ﬁxed

stars either their order of brightness or their distribution in the sky,

so there are various relations among numbers which may be observed,

and which give rise to various diﬀerent orders among numbers, all

equally legitimate. And what is true of numbers is equally true of

points on a line or of the moments of time: one order is more familiar,

but others are equally valid. We might, for example, take ﬁrst, on

a line, all the points that have integral co-ordinates, then all those

that have non-integral rational co-ordinates, then all those that have

algebraic non-rational co-ordinates, and so on, through any set of

complications we please. The resulting order will be one which the

points of the line certainly have, whether we choose to notice it or

not; the only thing that is arbitrary about the various orders of a set

of terms is our attention, for the terms themselves have always all the

orders of which they are capable.

One important result of this consideration is that we must not

look for the deﬁnition of order in the nature of the set of terms to be

ordered, since one set of terms has many orders. The order lies, not in

the class of terms, but in a relation among | j1 the members of the class,

in respect of which some appear as earlier and some as later. The fact

that a class may have many orders is due to the fact that there can be

many relations holding among the members of one single class. What

properties must a relation have in order to give rise to an order?

The essential characteristics of a relation which is to give rise to

order may be discovered by considering that in respect of such a

relation we must be able to say, of any two terms in the class which

is to be ordered, that one “precedes” and the other “follows.” Now,

in order that we may be able to use these words in the way in which

we should naturally understand them, we require that the ordering

relation should have three properties:—

(1) If x precedes y, y must not also precede x. This is an obvious

characteristic of the kind of relations that lead to series. If x is less

than y, y is not also less than x. If x is earlier in time than y, y is not

also earlier than x. If x is to the left of y, y is not to the left of x. On

the other hand, relations which do not give rise to series often do not

have this property. If x is a brother or sister of y, y is a brother or

sister of x. If x is of the same height as y, y is of the same height as

x. If x is of a diﬀerent height from y, y is of a diﬀerent height from

x. In all these cases, when the relation holds between x and y, it also

holds between y and x. But with serial relations such a thing cannot

happen. A relation having this ﬁrst property is called asymmetrical.

(z) If x precedes y and y precedes z, x must precede z. This may

be illustrated by the same instances as before: less, earlier, left of. But

Chap. IV. The Deﬁnition of Order z±

as instances of relations which do not have this property only two of

our previous three instances will serve. If x is brother or sister of y,

and y of z, x may not be brother or sister of z, since x and z may be

the same person. The same applies to diﬀerence of height, but not

to sameness of height, which has our second property but not our

ﬁrst. The relation “father,” on the other hand, has our ﬁrst property

but not | jz our second. A relation having our second property is called

transitive.

(j) Given any two terms of the class which is to be ordered, there

must be one which precedes and the other which follows. For exam-

ple, of any two integers, or fractions, or real numbers, one is smaller

and the other greater; but of any two complex numbers this is not

true. Of any two moments in time, one must be earlier than the other;

but of events, which may be simultaneous, this cannot be said. Of

two points on a line, one must be to the left of the other. A relation

having this third property is called connected.

When a relation possesses these three properties, it is of the sort

to give rise to an order among the terms between which it holds; and

wherever an order exists, some relation having these three properties

can be found generating it.

Before illustrating this thesis, we will introduce a few deﬁnitions.

(1) A relation is said to be an aliorelative,

1

or to be contained in

or imply diversity, if no term has this relation to itself. Thus, for

example, “greater,” “diﬀerent in size,” “brother,” “husband,” “father”

are aliorelatives; but “equal,” “born of the same parents,” “dear

friend” are not.

(z) The square of a relation is that relation which holds between

two terms x and z when there is an intermediate term y such that

the given relation holds between x and y and between y and z. Thus

“paternal grandfather” is the square of “father,” “greater by z” is the

square of “greater by 1,” and so on.

(j) The domain of a relation consists of all those terms that have

the relation to something or other, and the converse domain consists

of all those terms to which something or other has the relation. These

words have been already deﬁned, but are recalled here for the sake of

the following deﬁnition:—

(±) The ﬁeld of a relation consists of its domain and converse

domain together. |

(¸) jj One relation is said to contain or be implied by another if it

holds whenever the other holds.

It will be seen that an asymmetrical relation is the same thing as

a relation whose square is an aliorelative. It often happens that a

relation is an aliorelative without being asymmetrical, though an

1

This term is due to C. S. Peirce.

Chap. IV. The Deﬁnition of Order z¸

asymmetrical relation is always an aliorelative. For example, “spouse”

is an aliorelative, but is symmetrical, since if x is the spouse of y, y is

the spouse of x. But among transitive relations, all aliorelatives are

asymmetrical as well as vice versa.

From the deﬁnitions it will be seen that a transitive relation is one

which is implied by its square, or, as we also say, “contains” its square.

Thus “ancestor” is transitive, because an ancestor’s ancestor is an

ancestor; but “father” is not transitive, because a father’s father is

not a father. A transitive aliorelative is one which contains its square

and is contained in diversity; or, what comes to the same thing, one

whose square implies both it and diversity—because, when a relation

is transitive, asymmetry is equivalent to being an aliorelative.

A relation is connected when, given any two diﬀerent terms of its

ﬁeld, the relation holds between the ﬁrst and the second or between

the second and the ﬁrst (not excluding the possibility that both may

happen, though both cannot happen if the relation is asymmetrical).

It will be seen that the relation “ancestor,” for example, is an

aliorelative and transitive, but not connected; it is because it is not

connected that it does not suﬃce to arrange the human race in a

series.

The relation “less than or equal to,” among numbers, is transitive

and connected, but not asymmetrical or an aliorelative.

The relation “greater or less” among numbers is an aliorelative

and is connected, but is not transitive, for if x is greater or less than

y, and y is greater or less than z, it may happen that x and z are the

same number.

Thus the three properties of being (1) an aliorelative, (z) | j± transi-

tive, and (j) connected, are mutually independent, since a relation

may have any two without having the third.

We now lay down the following deﬁnition:—

A relation is serial when it is an aliorelative, transitive, and con-

nected; or, what is equivalent, when it is asymmetrical, transitive,

and connected.

A series is the same thing as a serial relation.

It might have been thought that a series should be the ﬁeld of a

serial relation, not the serial relation itself. But this would be an error.

For example,

1, z, j; 1, j, z; z, j, 1; z, 1, j; j, 1, z; j, z, 1

are six diﬀerent series which all have the same ﬁeld. If the ﬁeld were

the series, there could only be one series with a given ﬁeld. What

distinguishes the above six series is simply the diﬀerent ordering

relations in the six cases. Given the ordering relation, the ﬁeld and

the order are both determinate. Thus the ordering relation may be

taken to be the series, but the ﬁeld cannot be so taken.

Chap. IV. The Deﬁnition of Order z6

Given any serial relation, say P, we shall say that, in respect of this

relation, x “precedes” y if x has the relation P to y, which we shall

write “xPy” for short. The three characteristics which P must have in

order to be serial are:

(1) We must never have xPx, i.e. no term must precede itself.

(z) P

z

must imply P, i.e. if x precedes y and y precedes z, x must

precede z.

(j) If x and y are two diﬀerent terms in the ﬁeld of P, we shall have

xPy or yPx, i.e. one of the two must precede the other.

The reader can easily convince himself that, where these three prop-

erties are found in an ordering relation, the characteristics we expect

of series will also be found, and vice versa. We are therefore justiﬁed

in taking the above as a deﬁnition of order | j¸ or series. And it will be

observed that the deﬁnition is eﬀected in purely logical terms.

Although a transitive asymmetrical connected relation always

exists wherever there is a series, it is not always the relation which

would most naturally be regarded as generating the series. The

natural-number series may serve as an illustration. The relation

we assumed in considering the natural numbers was the relation of

immediate succession, i.e. the relation between consecutive integers.

This relation is asymmetrical, but not transitive or connected. We can,

however, derive from it, by the method of mathematical induction,

the “ancestral” relation which we considered in the preceding chap-

ter. This relation will be the same as “less than or equal to” among

inductive integers. For purposes of generating the series of natural

numbers, we want the relation “less than,” excluding “equal to.” This

is the relation of m to n when m is an ancestor of n but not identical

with n, or (what comes to the same thing) when the successor of m is

an ancestor of n in the sense in which a number is its own ancestor.

That is to say, we shall lay down the following deﬁnition:—

An inductive number m is said to be less than another number n

when n possesses every hereditary property possessed by the succes-

sor of m.

It is easy to see, and not diﬃcult to prove, that the relation “less

than,” so deﬁned, is asymmetrical, transitive, and connected, and has

the inductive numbers for its ﬁeld. Thus by means of this relation the

inductive numbers acquire an order in the sense in which we deﬁned

the term “order,” and this order is the so-called “natural” order, or

order of magnitude.

The generation of series by means of relations more or less resem-

bling that of n to n +1 is very common. The series of the Kings of

England, for example, is generated by relations of each to his suc-

cessor. This is probably the easiest way, where it is applicable, of

conceiving the generation of a series. In this method we pass on from

Chap. IV. The Deﬁnition of Order z;

each term to the next, as long as there | j6 is a next, or back to the one

before, as long as there is one before. This method always requires

the generalised form of mathematical induction in order to enable

us to deﬁne “earlier” and “later” in a series so generated. On the

analogy of “proper fractions,” let us give the name “proper posterity

of x with respect to R” to the class of those terms that belong to the

R-posterity of some term to which x has the relation R, in the sense

which we gave before to “posterity,” which includes a term in its own

posterity. Reverting to the fundamental deﬁnitions, we ﬁnd that the

“proper posterity” may be deﬁned as follows:—

The “proper posterity” of x with respect to R consists of all terms

that possess every R-hereditary property possessed by every term to

which x has the relation R.

It is to be observed that this deﬁnition has to be so framed as to

be applicable not only when there is only one term to which x has

the relation R, but also in cases (as e.g. that of father and child) where

there may be many terms to which x has the relation R. We deﬁne

further:

A term x is a “proper ancestor” of y with respect to R if y belongs

to the proper posterity of x with respect to R.

We shall speak for short of “R-posterity” and “R-ancestors” when

these terms seem more convenient.

Reverting now to the generation of series by the relation R be-

tween consecutive terms, we see that, if this method is to be possible,

the relation “proper R-ancestor” must be an aliorelative, transitive,

and connected. Under what circumstances will this occur? It will

always be transitive: no matter what sort of relation R may be, “R-

ancestor” and “proper R-ancestor” are always both transitive. But

it is only under certain circumstances that it will be an aliorelative

or connected. Consider, for example, the relation to one’s left-hand

neighbour at a round dinner-table at which there are twelve people.

If we call this relation R, the proper R-posterity of a person consists

of all who can be reached by going round the table from right to

left. This includes everybody at the table, including the person him-

self, since | j; twelve steps bring us back to our starting-point. Thus

in such a case, though the relation “proper R-ancestor” is connected,

and though R itself is an aliorelative, we do not get a series because

“proper R-ancestor” is not an aliorelative. It is for this reason that we

cannot say that one person comes before another with respect to the

relation “right of” or to its ancestral derivative.

The above was an instance in which the ancestral relation was

connected but not contained in diversity. An instance where it is

contained in diversity but not connected is derived from the ordinary

sense of the word “ancestor.” If x is a proper ancestor of y, x and y

cannot be the same person; but it is not true that of any two persons

one must be an ancestor of the other.

Chap. IV. The Deﬁnition of Order z8

The question of the circumstances under which series can be

generated by ancestral relations derived from relations of consecu-

tiveness is often important. Some of the most important cases are

the following: Let R be a many-one relation, and let us conﬁne our

attention to the posterity of some term x. When so conﬁned, the

relation “proper R-ancestor” must be connected; therefore all that

remains to ensure its being serial is that it shall be contained in di-

versity. This is a generalisation of the instance of the dinner-table.

Another generalisation consists in taking R to be a one-one relation,

and including the ancestry of x as well as the posterity. Here again,

the one condition required to secure the generation of a series is that

the relation “proper R-ancestor” shall be contained in diversity.

The generation of order by means of relations of consecutiveness,

though important in its own sphere, is less general than the method

which uses a transitive relation to deﬁne the order. It often happens

in a series that there are an inﬁnite number of intermediate terms

between any two that may be selected, however near together these

may be. Take, for instance, fractions in order of magnitude. Between

any two fractions there are others—for example, the arithmetic mean

of the two. Consequently there is no such thing as a pair of consec-

utive fractions. If we depended | j8 upon consecutiveness for deﬁning

order, we should not be able to deﬁne the order of magnitude among

fractions. But in fact the relations of greater and less among fractions

do not demand generation from relations of consecutiveness, and the

relations of greater and less among fractions have the three charac-

teristics which we need for deﬁning serial relations. In all such cases

the order must be deﬁned by means of a transitive relation, since only

such a relation is able to leap over an inﬁnite number of intermedi-

ate terms. The method of consecutiveness, like that of counting for

discovering the number of a collection, is appropriate to the ﬁnite;

it may even be extended to certain inﬁnite series, namely, those in

which, though the total number of terms is inﬁnite, the number of

terms between any two is always ﬁnite; but it must not be regarded

as general. Not only so, but care must be taken to eradicate from the

imagination all habits of thought resulting from supposing it general.

If this is not done, series in which there are no consecutive terms will

remain diﬃcult and puzzling. And such series are of vital importance

for the understanding of continuity, space, time, and motion.

There are many ways in which series may be generated, but all de-

pend upon the ﬁnding or construction of an asymmetrical transitive

connected relation. Some of these ways have considerable importance.

We may take as illustrative the generation of series by means of a

three-term relation which we may call “between.” This method is

very useful in geometry, and may serve as an introduction to relations

having more than two terms; it is best introduced in connection with

elementary geometry.

Chap. IV. The Deﬁnition of Order zo

Given any three points on a straight line in ordinary space, there

must be one of them which is between the other two. This will not

be the case with the points on a circle or any other closed curve,

because, given any three points on a circle, we can travel from any

one to any other without passing through the third. In fact, the notion

“between” is characteristic of open series—or series in the strict sense—

as opposed to what may be called | jo “cyclic” series, where, as with

people at the dinner-table, a suﬃcient journey brings us back to

our starting-point. This notion of “between” may be chosen as the

fundamental notion of ordinary geometry; but for the present we

will only consider its application to a single straight line and to the

ordering of the points on a straight line.

z

Taking any two points a, b,

the line (ab) consists of three parts (besides a and b themselves):

(1) Points between a and b.

(z) Points x such that a is between x and b.

(j) Points y such that b is between y and a.

Thus the line (ab) can be deﬁned in terms of the relation “between.”

In order that this relation “between” may arrange the points of

the line in an order from left to right, we need certain assumptions,

namely, the following:—

(1) If anything is between a and b, a and b are not identical.

(z) Anything between a and b is also between b and a.

(j) Anything between a and b is not identical with a (nor, conse-

quently, with b, in virtue of (z)).

(±) If x is between a and b, anything between a and x is also

between a and b.

(¸) If x is between a and b, and b is between x and y, then b is

between a and y.

(6) If x and y are between a and b, then either x and y are identical,

or x is between a and y, or x is between y and b.

(;) If b is between a and x and also between a and y, then either x

and y are identical, or x is between b and y, or y is between b and x.

These seven properties are obviously veriﬁed in the case of points

on a straight line in ordinary space. Any three-term relation which

veriﬁes them gives rise to series, as may be seen from the following

deﬁnitions. For the sake of deﬁniteness, let us assume | ±o that a is to

the left of b. Then the points of the line (ab) are (1) those between

which and b, a lies—these we will call to the left of a; (z) a itself; (j)

those between a and b; (±) b itself; (¸) those between which and a lies

b—these we will call to the right of b. We may now deﬁne generally

that of two points x, y, on the line (ab), we shall say that x is “to the

left of” y in any of the following cases:—

z

Cf. Rivista di Matematica, iv. pp. ¸¸ﬀ.; Principles of Mathematics, p. jo± (§j;¸).

Chap. IV. The Deﬁnition of Order jo

(1) When x and y are both to the left of a, and y is between x and

a;

(z) When x is to the left of a, and y is a or b or between a and b or

to the right of b;

(j) When x is a, and y is between a and b or is b or is to the right

of b;

(±) When x and y are both between a and b, and y is between x

and b;

(¸) When x is between a and b, and y is b or to the right of b;

(6) When x is b and y is to the right of b;

(;) When x and y are both to the right of b and x is between b and

y.

It will be found that, from the seven properties which we have

assigned to the relation “between,” it can be deduced that the relation

“to the left of,” as above deﬁned, is a serial relation as we deﬁned

that term. It is important to notice that nothing in the deﬁnitions or

the argument depends upon our meaning by “between” the actual

relation of that name which occurs in empirical space: any three-term

relation having the above seven purely formal properties will serve

the purpose of the argument equally well.

Cyclic order, such as that of the points on a circle, cannot be

generated by means of three-term relations of “between.” We need a

relation of four terms, which may be called “separation of couples.”

The point may be illustrated by considering a journey round the

world. One may go from England to New Zealand by way of Suez

or by way of San Francisco; we cannot | ±1 say deﬁnitely that either of

these two places is “between” England and New Zealand. But if a

man chooses that route to go round the world, whichever way round

he goes, his times in England and New Zealand are separated from

each other by his times in Suez and San Francisco, and conversely.

Generalising, if we take any four points on a circle, we can separate

them into two couples, say a and b and x and y, such that, in order

to get from a to b one must pass through either x or y, and in order

to get from x to y one must pass through either a or b. Under these

circumstances we say that the couple (a, b) are “separated” by the

couple (x, y). Out of this relation a cyclic order can be generated, in

a way resembling that in which we generated an open order from

“between,” but somewhat more complicated.

j

The purpose of the latter half of this chapter has been to suggest

the subject which one may call “generation of serial relations.” When

such relations have been deﬁned, the generation of them from other

relations possessing only some of the properties required for series

becomes very important, especially in the philosophy of geometry

j

Cf. Principles of Mathematics, p. zo¸ (§1o±), and references there given.

Chap. IV. The Deﬁnition of Order j1

and physics. But we cannot, within the limits of the present volume,

do more than make the reader aware that such a subject exists.

CHAPTER V

KINDS OF RELATIONS

±z A great part of the philosophy of mathematics is concerned with

relations, and many diﬀerent kinds of relations have diﬀerent kinds of

uses. It often happens that a property which belongs to all relations

is only important as regards relations of certain sorts; in these cases

the reader will not see the bearing of the proposition asserting such a

property unless he has in mind the sorts of relations for which it is

useful. For reasons of this description, as well as from the intrinsic

interest of the subject, it is well to have in our minds a rough list of

the more mathematically serviceable varieties of relations.

We dealt in the preceding chapter with a supremely important

class, namely, serial relations. Each of the three properties which

we combined in deﬁning series—namely, asymmetry, transitiveness,

and connexity—has its own importance. We will begin by saying

something on each of these three.

Asymmetry, i.e. the property of being incompatible with the con-

verse, is a characteristic of the very greatest interest and importance.

In order to develop its functions, we will consider various examples.

The relation husband is asymmetrical, and so is the relation wife; i.e. if

a is husband of b, b cannot be husband of a, and similarly in the case

of wife. On the other hand, the relation “spouse” is symmetrical: if a

is spouse of b, then b is spouse of a. Suppose now we are given the

relation spouse, and we wish to derive the relation husband. Husband

is the same as male spouse or spouse of a female; thus the relation

husband can | ±j be derived from spouse either by limiting the domain to

males or by limiting the converse domain to females. We see fromthis

instance that, when a symmetrical relation is given, it is sometimes

possible, without the help of any further relation, to separate it into

two asymmetrical relations. But the cases where this is possible are

rare and exceptional: they are cases where there are two mutually

exclusive classes, say α and β, such that whenever the relation holds

between two terms, one of the terms is a member of α and the other

is a member of β—as, in the case of spouse, one term of the relation

belongs to the class of males and one to the class of females. In such a

jz

Chap. V. Kinds of Relations jj

case, the relation with its domain conﬁned to α will be asymmetrical,

and so will the relation with its domain conﬁned to β. But such cases

are not of the sort that occur when we are dealing with series of more

than two terms; for in a series, all terms, except the ﬁrst and last (if

these exist), belong both to the domain and to the converse domain

of the generating relation, so that a relation like husband, where the

domain and converse domain do not overlap, is excluded.

The question how to construct relations having some useful prop-

erty by means of operations upon relations which only have rudi-

ments of the property is one of considerable importance. Transitive-

ness and connexity are easily constructed in many cases where the

originally given relation does not possess them: for example, if R

is any relation whatever, the ancestral relation derived from R by

generalised induction is transitive; and if R is a many-one relation,

the ancestral relation will be connected if conﬁned to the posterity

of a given term. But asymmetry is a much more diﬃcult property to

secure by construction. The method by which we derived husband

from spouse is, as we have seen, not available in the most important

cases, such as greater, before, to the right of, where domain and con-

verse domain overlap. In all these cases, we can of course obtain a

symmetrical relation by adding together the given relation and its

converse, but we cannot pass back from this symmetrical relation to

the original asymmetrical relation except by the help of some asym-

metrical | ±± relation. Take, for example, the relation greater: the relation

greater or less—i.e. unequal—is symmetrical, but there is nothing in

this relation to show that it is the sum of two asymmetrical relations.

Take such a relation as “diﬀering in shape.” This is not the sum of

an asymmetrical relation and its converse, since shapes do not form

a single series; but there is nothing to show that it diﬀers from “dif-

fering in magnitude” if we did not already know that magnitudes

have relations of greater and less. This illustrates the fundamental

character of asymmetry as a property of relations.

From the point of view of the classiﬁcation of relations, being

asymmetrical is a much more important characteristic than implying

diversity. Asymmetrical relations imply diversity, but the converse

is not the case. “Unequal,” for example, implies diversity, but is

symmetrical. Broadly speaking, we may say that, if we wished as

far as possible to dispense with relational propositions and replace

them by such as ascribed predicates to subjects, we could succeed in

this so long as we conﬁned ourselves to symmetrical relations: those

that do not imply diversity, if they are transitive, may be regarded as

asserting a common predicate, while those that do imply diversity

may be regarded as asserting incompatible predicates. For example,

consider the relation of similarity between classes, by means of which

we deﬁned numbers. This relation is symmetrical and transitive and

Chap. V. Kinds of Relations j±

does not imply diversity. It would be possible, though less simple

than the procedure we adopted, to regard the number of a collection

as a predicate of the collection: then two similar classes will be two

that have the same numerical predicate, while two that are not similar

will be two that have diﬀerent numerical predicates. Such a method

of replacing relations by predicates is formally possible (though often

very inconvenient) so long as the relations concerned are symmetrical;

but it is formally impossible when the relations are asymmetrical,

because both sameness and diﬀerence of predicates are symmetrical.

Asymmetrical relations are, we may | ±¸ say, the most characteristically

relational of relations, and the most important to the philosopher

who wishes to study the ultimate logical nature of relations.

Another class of relations that is of the greatest use is the class of

one-many relations, i.e. relations which at most one term can have

to a given term. Such are father, mother, husband (except in Tibet),

square of, sine of, and so on. But parent, square root, and so on,

are not one-many. It is possible, formally, to replace all relations by

one-many relations by means of a device. Take (say) the relation less

among the inductive numbers. Given any number n greater than 1,

there will not be only one number having the relation less to n, but we

can form the whole class of numbers that are less than n. This is one

class, and its relation to n is not shared by any other class. We may

call the class of numbers that are less than n the “proper ancestry”

of n, in the sense in which we spoke of ancestry and posterity in

connection with mathematical induction. Then “proper ancestry” is

a one-many relation (one-many will always be used so as to include

one-one), since each number determines a single class of numbers as

constituting its proper ancestry. Thus the relation less than can be

replaced by being a member of the proper ancestry of. In this way a one-

many relation in which the one is a class, together with membership

of this class, can always formally replace a relation which is not one-

many. Peano, who for some reason always instinctively conceives of a

relation as one-many, deals in this way with those that are naturally

not so. Reduction to one-many relations by this method, however,

though possible as a matter of form, does not represent a technical

simpliﬁcation, and there is every reason to think that it does not

represent a philosophical analysis, if only because classes must be

regarded as “logical ﬁctions.” We shall therefore continue to regard

one-many relations as a special kind of relations.

One-many relations are involved in all phrases of the form “the

so-and-so of such-and-such.” “The King of England,” | ±6 “the wife of

Socrates,” “the father of John Stuart Mill,” and so on, all describe

some person by means of a one-many relation to a given term. A

person cannot have more than one father, therefore “the father of

John Stuart Mill” described some one person, even if we did not know

Chap. V. Kinds of Relations j¸

whom. There is much to say on the subject of descriptions, but for the

present it is relations that we are concerned with, and descriptions

are only relevant as exemplifying the uses of one-many relations. It

should be observed that all mathematical functions result from one-

many relations: the logarithm of x, the cosine of x, etc., are, like the

father of x, terms described by means of a one-many relation (loga-

rithm, cosine, etc.) to a given term(x). The notion of function need not

be conﬁned to numbers, or to the uses to which mathematicians have

accustomed us; it can be extended to all cases of one-many relations,

and “the father of x” is just as legitimately a function of which x is

the argument as is “the logarithm of x.” Functions in this sense are

descriptive functions. As we shall see later, there are functions of a

still more general and more fundamental sort, namely, propositional

functions; but for the present we shall conﬁne our attention to de-

scriptive functions, i.e. “the term having the relation R to x,” or, for

short, “the R of x,” where R is any one-many relation.

It will be observed that if “the R of x” is to describe a deﬁnite

term, x must be a term to which something has the relation R, and

there must not be more than one term having the relation R to x, since

“the,” correctly used, must imply uniqueness. Thus we may speak of

“the father of x” if x is any human being except Adam and Eve; but we

cannot speak of “the father of x” if x is a table or a chair or anything

else that does not have a father. We shall say that the R of x “exists”

when there is just one term, and no more, having the relation R to x.

Thus if R is a one-many relation, the R of x exists whenever x belongs

to the converse domain of R, and not otherwise. Regarding “the R

of x” as a function in the mathematical | ±; sense, we say that x is the

“argument” of the function, and if y is the term which has the relation

R to x, i.e. if y is the R of x, then y is the “value” of the function for

the argument x. If R is a one-many relation, the range of possible

arguments to the function is the converse domain of R, and the range

of values is the domain. Thus the range of possible arguments to the

function “the father of x” is all who have fathers, i.e. the converse

domain of the relation father, while the range of possible values for

the function is all fathers, i.e. the domain of the relation.

Many of the most important notions in the logic of relations are

descriptive functions, for example: converse, domain, converse domain,

ﬁeld. Other examples will occur as we proceed.

Among one-many relations, one-one relations are a specially im-

portant class. We have already had occasion to speak of one-one

relations in connection with the deﬁnition of number, but it is neces-

sary to be familiar with them, and not merely to know their formal

deﬁnition. Their formal deﬁnition may be derived from that of one-

many relations: they may be deﬁned as one-many relations which are

also the converses of one-many relations, i.e. as relations which are

Chap. V. Kinds of Relations j6

both one-many and many-one. One-many relations may be deﬁned

as relations such that, if x has the relation in question to y, there is no

other term x

**which also has the relation to y. Or, again, they may be
**

deﬁned as follows: Given two terms x and x

**, the terms to which x
**

has the given relation and those to which x

**has it have no member
**

in common. Or, again, they may be deﬁned as relations such that

the relative product of one of them and its converse implies identity,

where the “relative product” of two relations R and S is that relation

which holds between x and z when there is an intermediate term y,

such that x has the relation R to y and y has the relation S to z. Thus,

for example, if R is the relation of father to son, the relative product

of R and its converse will be the relation which holds between x and

a man z when there is a person y, such that x is the father of y and y

is the son of z. It is obvious that x and z must be | ±8 the same person. If,

on the other hand, we take the relation of parent and child, which is

not one-many, we can no longer argue that, if x is a parent of y and

y is a child of z, x and z must be the same person, because one may

be the father of y and the other the mother. This illustrates that it

is characteristic of one-many relations when the relative product of

a relation and its converse implies identity. In the case of one-one

relations this happens, and also the relative product of the converse

and the relation implies identity. Given a relation R, it is convenient,

if x has the relation R to y, to think of y as being reached from x by

an “R-step” or an “R-vector.” In the same case x will be reached from

y by a “backward R-step.” Thus we may state the characteristic of

one-many relations with which we have been dealing by saying that

an R-step followed by a backward R-step must bring us back to our

starting-point. With other relations, this is by no means the case; for

example, if R is the relation of child to parent, the relative product of

R and its converse is the relation “self or brother or sister,” and if R is

the relation of grandchild to grandparent, the relative product of R

and its converse is “self or brother or sister or ﬁrst cousin.” It will be

observed that the relative product of two relations is not in general

commutative, i.e. the relative product of R and S is not in general

the same relation as the relative product of S and R. E.g. the relative

product of parent and brother is uncle, but the relative product of

brother and parent is parent.

One-one relations give a correlation of two classes, term for term,

so that each term in either class has its correlate in the other. Such

correlations are simplest to grasp when the two classes have no mem-

bers in common, like the class of husbands and the class of wives;

for in that case we know at once whether a term is to be considered

as one from which the correlating relation R goes, or as one to which

it goes. It is convenient to use the word referent for the term from

which the relation goes, and the term relatum for the term to which it

Chap. V. Kinds of Relations j;

goes. Thus if x and y are husband and wife, then, with respect to the

relation | ±o “husband,” x is referent and y relatum, but with respect to

the relation “wife,” y is referent and x relatum. We say that a relation

and its converse have opposite “senses”; thus the “sense” of a relation

that goes from x to y is the opposite of that of the corresponding rela-

tion from y to x. The fact that a relation has a “sense” is fundamental,

and is part of the reason why order can be generated by suitable

relations. It will be observed that the class of all possible referents to

a given relation is its domain, and the class of all possible relata is its

converse domain.

But it very often happens that the domain and converse domain

of a one-one relation overlap. Take, for example, the ﬁrst ten integers

(excluding o), and add 1 to each; thus instead of the ﬁrst ten integers

we now have the integers

z, j, ±, ¸, 6, ;, 8, o, 1o, 11.

These are the same as those we had before, except that 1 has been

cut oﬀ at the beginning and 11 has been joined on at the end. There

are still ten integers: they are correlated with the previous ten by the

relation of n to n+1, which is a one-one relation. Or, again, instead of

adding 1 to each of our original ten integers, we could have doubled

each of them, thus obtaining the integers

z, ±, 6, 8, 1o, 1z, 1±, 16, 18, zo.

Here we still have ﬁve of our previous set of integers, namely, z, ±, 6,

8, 1o. The correlating relation in this case is the relation of a number

to its double, which is again a one-one relation. Or we might have

replaced each number by its square, thus obtaining the set

1, ±, o, 16, z¸, j6, ±o, 6±, 81, 1oo.

On this occasion only three of our original set are left, namely, 1, ±, o.

Such processes of correlation may be varied endlessly.

The most interesting case of the above kind is the case where our

one-one relation has a converse domain which is part, but | ¸o not the

whole, of the domain. If, instead of conﬁning the domain to the ﬁrst

ten integers, we had considered the whole of the inductive numbers,

the above instances would have illustrated this case. We may place

the numbers concerned in two rows, putting the correlate directly

under the number whose correlate it is. Thus when the correlator is

the relation of n to n +1, we have the two rows:

1, z, j, ±, ¸, . . . n . . .

z, j, ±, ¸, 6, . . . n +1 . . .

Chap. V. Kinds of Relations j8

When the correlator is the relation of a number to its double, we have

the two rows:

1, z, j, ±, ¸, . . . n . . .

z, ±, 6, 8, 1o, . . . zn . . .

When the correlator is the relation of a number to its square, the rows

are:

1, z, j, ±, ¸, . . . n . . .

1, ±, o, 16, z¸, . . . n

z

. . .

In all these cases, all inductive numbers occur in the top row, and

only some in the bottom row.

Cases of this sort, where the converse domain is a “proper part”

of the domain (i.e. a part not the whole), will occupy us again when

we come to deal with inﬁnity. For the present, we wish only to note

that they exist and demand consideration.

Another class of correlations which are often important is the

class called “permutations,” where the domain and converse domain

are identical. Consider, for example, the six possible arrangements of

three letters:

a, b, c

a, c, b

b, c, a

b, a, c

c, a, b

c, b, a |

¸1 Each of these can be obtained from any one of the others by means of

a correlation. Take, for example, the ﬁrst and last, (a, b, c) and (c, b, a).

Here a is correlated with c, b with itself, and c with a. It is obvious

that the combination of two permutations is again a permutation, i.e.

the permutations of a given class form what is called a “group.”

These various kinds of correlations have importance in various

connections, some for one purpose, some for another. The general

notion of one-one correlations has boundless importance in the phi-

losophy of mathematics, as we have partly seen already, but shall see

much more fully as we proceed. One of its uses will occupy us in our

next chapter.

CHAPTER VI

SIMILARITY OF RELATIONS

¸z We saw in Chapter II. that two classes have the same number of terms

when they are “similar,” i.e. when there is a one-one relation whose

domain is the one class and whose converse domain is the other. In

such a case we say that there is a “one-one correlation” between the

two classes.

In the present chapter we have to deﬁne a relation between rela-

tions, which will play the same part for them that similarity of classes

plays for classes. We will call this relation “similarity of relations,” or

“likeness” when it seems desirable to use a diﬀerent word from that

which we use for classes. How is likeness to be deﬁned?

We shall employ still the notion of correlation: we shall assume

that the domain of the one relation can be correlated with the domain

of the other, and the converse domain with the converse domain; but

that is not enough for the sort of resemblance which we desire to

have between our two relations. What we desire is that, whenever

either relation holds between two terms, the other relation shall hold

between the correlates of these two terms. The easiest example of the

sort of thing we desire is a map. When one place is north of another,

the place on the map corresponding to the one is above the place

on the map corresponding to the other; when one place is west of

another, the place on the map corresponding to the one is to the left

of the place on the map corresponding to the other; and so on. The

structure of the map corresponds with that of | ¸j the country of which

it is a map. The space-relations in the map have “likeness” to the

space-relations in the country mapped. It is this kind of connection

between relations that we wish to deﬁne.

We may, in the ﬁrst place, proﬁtably introduce a certain restriction.

We will conﬁne ourselves, in deﬁning likeness, to such relations as

have “ﬁelds,” i.e. to such as permit of the formation of a single class

out of the domain and the converse domain. This is not always

the case. Take, for example, the relation “domain,” i.e. the relation

which the domain of a relation has to the relation. This relation has all

classes for its domain, since every class is the domain of some relation;

jo

Chap. VI. Similarity of Relations ±o

and it has all relations for its converse domain, since every relation has

a domain. But classes and relations cannot be added together to form

a new single class, because they are of diﬀerent logical “types.” We

do not need to enter upon the diﬃcult doctrine of types, but it is well

to know when we are abstaining from entering upon it. We may say,

without entering upon the grounds for the assertion, that a relation

only has a “ﬁeld” when it is what we call “homogeneous,” i.e. when

its domain and converse domain are of the same logical type; and as a

rough-and-ready indication of what we mean by a “type,” we may say

that individuals, classes of individuals, relations between individuals,

relations between classes, relations of classes to individuals, and so

on, are diﬀerent types. Now the notion of likeness is not very useful

as applied to relations that are not homogeneous; we shall, therefore,

in deﬁning likeness, simplify our problem by speaking of the “ﬁeld”

of one of the relations concerned. This somewhat limits the generality

of our deﬁnition, but the limitation is not of any practical importance.

And having been stated, it need no longer be remembered.

We may deﬁne two relations P and Q as “similar,” or as having

“likeness,” when there is a one-one relation S whose domain is the

ﬁeld of P and whose converse domain is the ﬁeld of Q, and which is

such that, if one term has the relation P | ¸± to another, the correlate of

the one has the relation Q to the correlate of the other, and vice versa.

A ﬁgure will make this clearer. Let x and y be two terms having the

z w

x y

Q

P

S S

relation P. Then there are to be two

terms z, w, such that x has the relation

S to z, y has the relation S to w, and z

has the relation Q to w. If this happens

with every pair of terms such as x and y,

and if the converse happens with every

pair of terms such as z and w, it is clear

that for every instance in which the re-

lation P holds there is a corresponding

instance in which the relation Q holds,

and vice versa; and this is what we de-

sire to secure by our deﬁnition. We can eliminate some redundancies

in the above sketch of a deﬁnition, by observing that, when the above

conditions are realised, the relation P is the same as the relative prod-

uct of S and Q and the converse of S, i.e. the P-step from x to y may be

replaced by the succession of the S-step from x to z, the Q-step from

z to w, and the backward S-step from w to y. Thus we may set up the

following deﬁnitions:—

A relation S is said to be a “correlator” or an “ordinal correlator”

of two relations P and Q if S is one-one, has the ﬁeld of Q for its

converse domain, and is such that P is the relative product of S and

Q and the converse of S.

Chap. VI. Similarity of Relations ±1

Two relations P and Qare said to be “similar,” or to have “likeness,”

when there is at least one correlator of P and Q.

These deﬁnitions will be found to yield what we above decided to

be necessary.

It will be found that, when two relations are similar, they share

all properties which do not depend upon the actual terms in their

ﬁelds. For instance, if one implies diversity, so does the other; if one

is transitive, so is the other; if one is connected, so is the other. Hence

if one is serial, so is the other. Again, if one is one-many or one-one,

the other is one-many | ¸¸ or one-one; and so on, through all the general

properties of relations. Even statements involving the actual terms of

the ﬁeld of a relation, though they may not be true as they stand when

applied to a similar relation, will always be capable of translation

into statements that are analogous. We are led by such considerations

to a problem which has, in mathematical philosophy, an importance

by no means adequately recognised hitherto. Our problem may be

stated as follows:—

Given some statement in a language of which we know the gram-

mar and the syntax, but not the vocabulary, what are the possible

meanings of such a statement, and what are the meanings of the

unknown words that would make it true?

The reason that this question is important is that it represents,

much more nearly than might be supposed, the state of our knowl-

edge of nature. We know that certain scientiﬁc propositions—which,

in the most advanced sciences, are expressed in mathematical sym-

bols—are more or less true of the world, but we are very much at

sea as to the interpretation to be put upon the terms which occur

in these propositions. We know much more (to use, for a moment,

an old-fashioned pair of terms) about the form of nature than about

the matter. Accordingly, what we really know when we enunciate

a law of nature is only that there is probably some interpretation of

our terms which will make the law approximately true. Thus great

importance attaches to the question: What are the possible meanings

of a law expressed in terms of which we do not know the substantive

meaning, but only the grammar and syntax? And this question is the

one suggested above.

For the present we will ignore the general question, which will

occupy us again at a later stage; the subject of likeness itself must

ﬁrst be further investigated.

Owing to the fact that, when two relations are similar, their prop-

erties are the same except when they depend upon the ﬁelds being

composed of just the terms of which they are composed, it is desirable

to have a nomenclature which collects | ¸6 together all the relations that

are similar to a given relation. Just as we called the set of those classes

that are similar to a given class the “number” of that class, so we may

Chap. VI. Similarity of Relations ±z

call the set of all those relations that are similar to a given relation

the “number” of that relation. But in order to avoid confusion with

the numbers appropriate to classes, we will speak, in this case, of a

“relation-number.” Thus we have the following deﬁnitions:—

The “relation-number” of a given relation is the class of all those

relations that are similar to the given relation.

“Relation-numbers” are the set of all those classes of relations

that are relation-numbers of various relations; or, what comes to the

same thing, a relation-number is a class of relations consisting of all

those relations that are similar to one member of the class.

When it is necessary to speak of the numbers of classes in a way

which makes it impossible to confuse themwith relation-numbers, we

shall call them “cardinal numbers.” Thus cardinal numbers are the

numbers appropriate to classes. These include the ordinary integers

of daily life, and also certain inﬁnite numbers, of which we shall

speak later. When we speak of “numbers” without qualiﬁcation, we

are to be understood as meaning cardinal numbers. The deﬁnition of

a cardinal number, it will be remembered, is as follows:—

The “cardinal number” of a given class is the set of all those classes

that are similar to the given class.

The most obvious application of relation-numbers is to series.

Two series may be regarded as equally long when they have the

same relation-number. Two ﬁnite series will have the same relation-

number when their ﬁelds have the same cardinal number of terms,

and only then—i.e. a series of (say) 1¸ terms will have the same

relation-number as any other series of ﬁfteen terms, but will not have

the same relation-number as a series of 1± or 16 terms, nor, of course,

the same relation-number as a relation which is not serial. Thus, in

the quite special case of ﬁnite series, there is parallelism between

cardinal and relation-numbers. The relation-numbers applicable to

series may be | ¸; called “serial numbers” (what are commonly called

“ordinal numbers” are a sub-class of these); thus a ﬁnite serial number

is determinate when we know the cardinal number of terms in the

ﬁeld of a series having the serial number in question. If n is a ﬁnite

cardinal number, the relation-number of a series which has n terms

is called the “ordinal” number n. (There are also inﬁnite ordinal

numbers, but of them we shall speak in a later chapter.) When the

cardinal number of terms in the ﬁeld of a series is inﬁnite, the relation-

number of the series is not determined merely by the cardinal number,

indeed an inﬁnite number of relation-numbers exist for one inﬁnite

cardinal number, as we shall see when we come to consider inﬁnite

series. When a series is inﬁnite, what we may call its “length,” i.e. its

relation-number, may vary without change in the cardinal number;

but when a series is ﬁnite, this cannot happen.

We can deﬁne addition and multiplication for relation-numbers

as well as for cardinal numbers, and a whole arithmetic of relation-

Chap. VI. Similarity of Relations ±j

numbers can be developed. The manner in which this is to be done is

easily seen by considering the case of series. Suppose, for example,

that we wish to deﬁne the sum of two non-overlapping series in such

a way that the relation-number of the sum shall be capable of being

deﬁned as the sum of the relation-numbers of the two series. In the

ﬁrst place, it is clear that there is an order involved as between the

two series: one of them must be placed before the other. Thus if

P and Q are the generating relations of the two series, in the series

which is their sum with P put before Q, every member of the ﬁeld

of P will precede every member of the ﬁeld of Q. Thus the serial

relation which is to be deﬁned as the sum of P and Q is not “P or

Q” simply, but “P or Q or the relation of any member of the ﬁeld

of P to any member of the ﬁeld of Q.” Assuming that P and Q do

not overlap, this relation is serial, but “P or Q” is not serial, being

not connected, since it does not hold between a member of the ﬁeld

of P and a member of the ﬁeld of Q. Thus the sum of P and Q, as

above deﬁned, is what we need in order | ¸8 to deﬁne the sum of two

relation-numbers. Similar modiﬁcations are needed for products and

powers. The resulting arithmetic does not obey the commutative

law: the sum or product of two relation-numbers generally depends

upon the order in which they are taken. But it obeys the associative

law, one form of the distributive law, and two of the formal laws

for powers, not only as applied to serial numbers, but as applied

to relation-numbers generally. Relation-arithmetic, in fact, though

recent, is a thoroughly respectable branch of mathematics.

It must not be supposed, merely because series aﬀord the most

obvious application of the idea of likeness, that there are no other

applications that are important. We have already mentioned maps,

and we might extend our thoughts from this illustration to geometry

generally. If the system of relations by which a geometry is applied to

a certain set of terms can be brought fully into relations of likeness

with a system applying to another set of terms, then the geometry

of the two sets is indistinguishable from the mathematical point of

view, i.e. all the propositions are the same, except for the fact that

they are applied in one case to one set of terms and in the other to

another. We may illustrate this by the relations of the sort that may

be called “between,” which we considered in Chapter IV. We there

saw that, provided a three-term relation has certain formal logical

properties, it will give rise to series, and may be called a “between-

relation.” Given any two points, we can use the between-relation to

deﬁne the straight line determined by those two points; it consists

of a and b together with all points x, such that the between-relation

holds between the three points a, b, x in some order or other. It has

been shown by O. Veblen that we may regard our whole space as

the ﬁeld of a three-term between-relation, and deﬁne our geometry

Chap. VI. Similarity of Relations ±±

by the properties we assign to our between-relation.

1

Now likeness

is just as easily | ¸o deﬁnable between three-term relations as between

two-term relations. If B and B

**are two between-relations, so that
**

“xB(y, z)” means “x is between y and z with respect to B,” we shall

call S a correlator of B and B

if it has the ﬁeld of B

**for its converse
**

domain, and is such that the relation B holds between three terms

when B

**holds between their S-correlates, and only then. And we
**

shall say that B is like B

**when there is at least one correlator of B
**

with B

. The reader can easily convince himself that, if B is like B

in

this sense, there can be no diﬀerence between the geometry generated

by B and that generated by B

.

It follows from this that the mathematician need not concern him-

self with the particular being or intrinsic nature of his points, lines,

and planes, even when he is speculating as an applied mathematician.

We may say that there is empirical evidence of the approximate truth

of such parts of geometry as are not matters of deﬁnition. But there is

no empirical evidence as to what a “point” is to be. It has to be some-

thing that as nearly as possible satisﬁes our axioms, but it does not

have to be “very small” or “without parts.” Whether or not it is those

things is a matter of indiﬀerence, so long as it satisﬁes the axioms. If

we can, out of empirical material, construct a logical structure, no

matter how complicated, which will satisfy our geometrical axioms,

that structure may legitimately be called a “point.” We must not say

that there is nothing else that could legitimately be called a “point”;

we must only say: “This object we have constructed is suﬃcient for

the geometer; it may be one of many objects, any of which would be

suﬃcient, but that is no concern of ours, since this object is enough

to vindicate the empirical truth of geometry, in so far as geometry is

not a matter of deﬁnition.” This is only an illustration of the general

principle that what matters in mathematics, and to a very great extent

in physical science, is not the intrinsic nature of our terms, but the

logical nature of their interrelations.

We may say, of two similar relations, that they have the same

| 6o “structure.” For mathematical purposes (though not for those of

pure philosophy) the only thing of importance about a relation is the

cases in which it holds, not its intrinsic nature. Just as a class may

be deﬁned by various diﬀerent but co-extensive concepts—e.g. “man”

and “featherless biped”—so two relations which are conceptually

diﬀerent may hold in the same set of instances. An “instance” in

which a relation holds is to be conceived as a couple of terms, with an

order, so that one of the terms comes ﬁrst and the other second; the

couple is to be, of course, such that its ﬁrst term has the relation in

1

This does not apply to elliptic space, but only to spaces in which the straight

line is an open series. Modern Mathematics, edited by J. W. A. Young, pp. j–¸1

(monograph by O. Veblen on “The Foundations of Geometry”).

Chap. VI. Similarity of Relations ±¸

question to its second. Take (say) the relation “father”: we can deﬁne

what we may call the “extension” of this relation as the class of all

ordered couples (x, y) which are such that x is the father of y. From

the mathematical point of view, the only thing of importance about

the relation “father” is that it deﬁnes this set of ordered couples.

Speaking generally, we say:

The “extension” of a relation is the class of those ordered couples

(x, y) which are such that x has the relation in question to y.

We can now go a step further in the process of abstraction, and

consider what we mean by “structure.” Given any relation, we can, if

it is a suﬃciently simple one, construct a map of it. For the sake of def-

initeness, let us take a relation of which the extension is the following

couples: ab, ac, ad, bc, ce, dc, de, where a, b, c, d, e are ﬁve terms, no

d

c

a b

e

matter what. We may make a “map” of this

relation by taking ﬁve points on a plane and

connecting them by arrows, as in the accom-

panying ﬁgure. What is revealed by the map

is what we call the “structure” of the relation.

It is clear that the “structure” of the re-

lation does not depend upon the particular

terms that make up the ﬁeld of the relation.

The ﬁeld may be changed without chang-

ing the structure, and the structure may be

changed without changing the ﬁeld—for |

61 example, if we were to add the couple ae in 61

the above illustration we should alter the structure but not the ﬁeld.

Two relations have the same “structure,” we shall say, when the same

map will do for both—or, what comes to the same thing, when either

can be a map for the other (since every relation can be its own map).

And that, as a moment’s reﬂection shows, is the very same thing as

what we have called “likeness.” That is to say, two relations have the

same structure when they have likeness, i.e. when they have the same

relation-number. Thus what we deﬁned as the “relation-number” is

the very same thing as is obscurely intended by the word “structure”—

a word which, important as it is, is never (so far as we know) deﬁned

in precise terms by those who use it.

There has been a great deal of speculation in traditional philoso-

phy which might have been avoided if the importance of structure,

and the diﬃculty of getting behind it, had been realised. For example,

it is often said that space and time are subjective, but they have objec-

tive counterparts; or that phenomena are subjective, but are caused

by things in themselves, which must have diﬀerences inter se corre-

sponding with the diﬀerences in the phenomena to which they give

rise. Where such hypotheses are made, it is generally supposed that

we can know very little about the objective counterparts. In actual

Chap. VI. Similarity of Relations ±6

fact, however, if the hypotheses as stated were correct, the objective

counterparts would form a world having the same structure as the

phenomenal world, and allowing us to infer from phenomena the

truth of all propositions that can be stated in abstract terms and are

known to be true of phenomena. If the phenomenal world has three

dimensions, so must the world behind phenomena; if the phenomenal

world is Euclidean, so must the other be; and so on. In short, every

proposition having a communicable signiﬁcance must be true of both

worlds or of neither: the only diﬀerence must lie in just that essence

of individuality which always eludes words and baﬄes description,

but which, for that very reason, is irrelevant to science. Now the only

purpose that philosophers | 6z have in view in condemning phenomena 6z

is in order to persuade themselves and others that the real world is

very diﬀerent from the world of appearance. We can all sympathise

with their wish to prove such a very desirable proposition, but we

cannot congratulate them on their success. It is true that many of

them do not assert objective counterparts to phenomena, and these

escape from the above argument. Those who do assert counterparts

are, as a rule, very reticent on the subject, probably because they

feel instinctively that, if pursued, it will bring about too much of a

rapprochement between the real and the phenomenal world. If they

were to pursue the topic, they could hardly avoid the conclusions

which we have been suggesting. In such ways, as well as in many

others, the notion of structure or relation-number is important.

CHAPTER VII

RATIONAL, REAL, ANDCOMPLEX

NUMBERS

6j We have now seen how to deﬁne cardinal numbers, and also relation-

numbers, of which what are commonly called ordinal numbers are a

particular species. It will be found that each of these kinds of number

may be inﬁnite just as well as ﬁnite. But neither is capable, as it stands,

of the more familiar extensions of the idea of number, namely, the

extensions to negative, fractional, irrational, and complex numbers.

In the present chapter we shall brieﬂy supply logical deﬁnitions of

these various extensions.

One of the mistakes that have delayed the discovery of correct

deﬁnitions in this region is the common idea that each extension of

number included the previous sorts as special cases. It was thought

that, in dealing with positive and negative integers, the positive inte-

gers might be identiﬁed with the original signless integers. Again it

was thought that a fraction whose denominator is 1 may be identiﬁed

with the natural number which is its numerator. And the irrational

numbers, such as the square root of z, were supposed to ﬁnd their

place among rational fractions, as being greater than some of them

and less than the others, so that rational and irrational numbers could

be taken together as one class, called “real numbers.” And when the

idea of number was further extended so as to include “complex” num-

bers, i.e. numbers involving the square root of −1, it was thought that

real numbers could be regarded as those among complex numbers in

which the imaginary part (i.e. the part | 6± which was a multiple of the

square root of −1) was zero. All these suppositions were erroneous,

and must be discarded, as we shall ﬁnd, if correct deﬁnitions are to

be given.

Let us begin with positive and negative integers. It is obvious on a

moment’s consideration that +1 and −1 must both be relations, and

in fact must be each other’s converses. The obvious and suﬃcient

deﬁnition is that +1 is the relation of n +1 to n, and −1 is the relation

of n to n + 1. Generally, if m is any inductive number, +m will be

the relation of n +m to n (for any n), and −m will be the relation of

±;

Chap. VII. Rational, Real, and Complex Numbers ±8

n to n + m. According to this deﬁnition, +m is a relation which is

one-one so long as n is a cardinal number (ﬁnite or inﬁnite) and m

is an inductive cardinal number. But +m is under no circumstances

capable of being identiﬁed with m, which is not a relation, but a class

of classes. Indeed, +m is every bit as distinct from m as −m is.

Fractions are more interesting than positive or negative integers.

We need fractions for many purposes, but perhaps most obviously

for purposes of measurement. My friend and collaborator Dr A. N.

Whitehead has developed a theory of fractions specially adapted

for their application to measurement, which is set forth in Principia

Mathematica.

1

But if all that is needed is to deﬁne objects having

the required purely mathematical properties, this purpose can be

achieved by a simpler method, which we shall here adopt. We shall

deﬁne the fraction m/n as being that relation which holds between

two inductive numbers x, y when xn = ym. This deﬁnition enables us

to prove that m/n is a one-one relation, provided neither m nor n is

zero. And of course n/m is the converse relation to m/n.

From the above deﬁnition it is clear that the fraction m/1 is that

relation between two integers x and y which consists in the fact that

x = my. This relation, like the relation +m, is by no means capable

of being identiﬁed with the inductive cardinal number m, because a

relation and a class of classes are objects | 6¸ of utterly diﬀerent kinds.

z

It will be seen that o/n is always the same relation, whatever inductive

number n may be; it is, in short, the relation of o to any other inductive

cardinal. We may call this the zero of rational numbers; it is not, of

course, identical with the cardinal number o. Conversely, the relation

m/o is always the same, whatever inductive number m may be. There

is not any inductive cardinal to correspond to m/o. We may call it

“the inﬁnity of rationals.” It is an instance of the sort of inﬁnite that

is traditional in mathematics, and that is represented by “∞.” This

is a totally diﬀerent sort from the true Cantorian inﬁnite, which we

shall consider in our next chapter. The inﬁnity of rationals does

not demand, for its deﬁnition or use, any inﬁnite classes or inﬁnite

integers. It is not, in actual fact, a very important notion, and we

could dispense with it altogether if there were any object in doing so.

The Cantorian inﬁnite, on the other hand, is of the greatest and most

fundamental importance; the understanding of it opens the way to

whole new realms of mathematics and philosophy.

It will be observed that zero and inﬁnity, alone among ratios, are

not one-one. Zero is one-many, and inﬁnity is many-one.

1

Vol. iii. ∗jooﬀ., especially joj.

z

Of course in practice we shall continue to speak of a fraction as (say) greater or

less than 1, meaning greater or less than the ratio 1/1. So long as it is understood

that the ratio 1/1 and the cardinal number 1 are diﬀerent, it is not necessary to be

always pedantic in emphasising the diﬀerence.

Chap. VII. Rational, Real, and Complex Numbers ±o

There is not any diﬃculty in deﬁning greater and less among ratios

(or fractions). Given two ratios m/n and p/q, we shall say that m/n is

less than p/q if mq is less than pn. There is no diﬃculty in proving

that the relation “less than,” so deﬁned, is serial, so that the ratios

form a series in order of magnitude. In this series, zero is the smallest

term and inﬁnity is the largest. If we omit zero and inﬁnity from our

series, there is no longer any smallest or largest ratio; it is obvious

that if m/n is any ratio other than zero and inﬁnity, m/zn is smaller

and zm/n is larger, though neither is zero or inﬁnity, so that m/n is

neither the smallest | 66 nor the largest ratio, and therefore (when zero

and inﬁnity are omitted) there is no smallest or largest, since m/n was

chosen arbitrarily. In like manner we can prove that however nearly

equal two fractions may be, there are always other fractions between

them. For, let m/n and p/q be two fractions, of which p/q is the

greater. Then it is easy to see (or to prove) that (m+p)/(n +q) will be

greater than m/n and less than p/q. Thus the series of ratios is one in

which no two terms are consecutive, but there are always other terms

between any two. Since there are other terms between these others,

and so on ad inﬁnitum, it is obvious that there are an inﬁnite number

of ratios between any two, however nearly equal these two may be.

j

A

series having the property that there are always other terms between

any two, so that no two are consecutive, is called “compact.” Thus

the ratios in order of magnitude form a “compact” series. Such series

have many important properties, and it is important to observe that

ratios aﬀord an instance of a compact series generated purely logically,

without any appeal to space or time or any other empirical datum.

Positive and negative ratios can be deﬁned in a way analogous to

that in which we deﬁned positive and negative integers. Having ﬁrst

deﬁned the sum of two ratios m/n and p/q as (mq +pn)/nq, we deﬁne

+p/q as the relation of m/n +p/q to m/n, where m/n is any ratio; and

−p/q is of course the converse of +p/q. This is not the only possible

way of deﬁning positive and negative ratios, but it is a way which, for

our purpose, has the merit of being an obvious adaptation of the way

we adopted in the case of integers.

We come now to a more interesting extension of the idea of num-

ber, i.e. the extension to what are called “real” numbers, which are

the kind that embrace irrationals. In Chapter I. we had occasion to

mention “incommensurables” and their | 6; discovery by Pythagoras.

It was through them, i.e. through geometry, that irrational numbers

were ﬁrst thought of. A square of which the side is one inch long will

have a diagonal of which the length is the square root of z inches. But,

as the ancients discovered, there is no fraction of which the square

j

Strictly speaking, this statement, as well as those following to the end of the

paragraph, involves what is called the “axiom of inﬁnity,” which will be discussed

in a later chapter.

Chap. VII. Rational, Real, and Complex Numbers ¸o

is z. This proposition is proved in the tenth book of Euclid, which is

one of those books that schoolboys supposed to be fortunately lost

in the days when Euclid was still used as a text-book. The proof is

extraordinarily simple. If possible, let m/n be the square root of z,

so that m

z

/n

z

= z, i.e. m

z

= zn

z

. Thus m

z

is an even number, and

therefore m must be an even number, because the square of an odd

number is odd. Now if m is even, m

z

must divide by ±, for if m = zp,

then m

z

= ±p

z

. Thus we shall have ±p

z

= zn

z

, where p is half of m.

Hence zp

z

= n

z

, and therefore n/p will also be the square root of z.

But then we can repeat the argument: if n = zq, p/q will also be the

square root of z, and so on, through an unending series of numbers

that are each half of its predecessor. But this is impossible; if we

divide a number by z, and then halve the half, and so on, we must

reach an odd number after a ﬁnite number of steps. Or we may put

the argument even more simply by assuming that the m/n we start

with is in its lowest terms; in that case, m and n cannot both be even;

yet we have seen that, if m

z

/n

z

= z, they must be. Thus there cannot

be any fraction m/n whose square is z.

Thus no fraction will express exactly the length of the diagonal

of a square whose side is one inch long. This seems like a challenge

thrown out by nature to arithmetic. However the arithmetician may

boast (as Pythagoras did) about the power of numbers, nature seems

able to baﬄe him by exhibiting lengths which no numbers can esti-

mate in terms of the unit. But the problem did not remain in this

geometrical form. As soon as algebra was invented, the same problem

arose as regards the solution of equations, though here it took on a

wider form, since it also involved complex numbers.

It is clear that fractions can be found which approach nearer | 68 and

nearer to having their square equal to z. We can form an ascending

series of fractions all of which have their squares less than z, but

diﬀering from z in their later members by less than any assigned

amount. That is to say, suppose I assign some small amount in

advance, say one-billionth, it will be found that all the terms of our

series after a certain one, say the tenth, have squares that diﬀer from

z by less than this amount. And if I had assigned a still smaller

amount, it might have been necessary to go further along the series,

but we should have reached sooner or later a term in the series, say

the twentieth, after which all terms would have had squares diﬀering

from z by less than this still smaller amount. If we set to work to

extract the square root of z by the usual arithmetical rule, we shall

obtain an unending decimal which, taken to so-and-so many places,

exactly fulﬁls the above conditions. We can equally well form a

descending series of fractions whose squares are all greater than z,

but greater by continually smaller amounts as we come to later terms

of the series, and diﬀering, sooner or later, by less than any assigned

Chap. VII. Rational, Real, and Complex Numbers ¸1

amount. In this way we seemto be drawing a cordon round the square

root of z, and it may seem diﬃcult to believe that it can permanently

escape us. Nevertheless, it is not by this method that we shall actually

reach the square root of z.

If we divide all ratios into two classes, according as their squares

are less than z or not, we ﬁnd that, among those whose squares are

not less than z, all have their squares greater than z. There is no

maximum to the ratios whose square is less than z, and no minimum

to those whose square is greater than z. There is no lower limit short

of zero to the diﬀerence between the numbers whose square is a little

less than z and the numbers whose square is a little greater than z.

We can, in short, divide all ratios into two classes such that all the

terms in one class are less than all in the other, there is no maximum

to the one class, and there is no minimum to the other. Between these

two classes, where

√

z ought to be, there is nothing. Thus our | 6o cordon,

though we have drawn it as tight as possible, has been drawn in the

wrong place, and has not caught

√

z.

The above method of dividing all the terms of a series into two

classes, of which the one wholly precedes the other, was brought into

prominence by Dedekind,

±

and is therefore called a “Dedekind cut.”

With respect to what happens at the point of section, there are four

possibilities: (1) there may be a maximum to the lower section and a

minimum to the upper section, (z) there may be a maximum to the

one and no minimum to the other, (j) there may be no maximum

to the one, but a minimum to the other, (±) there may be neither a

maximum to the one nor a minimum to the other. Of these four cases,

the ﬁrst is illustrated by any series in which there are consecutive

terms: in the series of integers, for instance, a lower section must

end with some number n and the upper section must then begin with

n +1. The second case will be illustrated in the series of ratios if we

take as our lower section all ratios up to and including 1, and in our

upper section all ratios greater than 1. The third case is illustrated if

we take for our lower section all ratios less than 1, and for our upper

section all ratios from 1 upward (including 1 itself). The fourth case,

as we have seen, is illustrated if we put in our lower section all ratios

whose square is less than z, and in our upper section all ratios whose

square is greater than z.

We may neglect the ﬁrst of our four cases, since it only arises in

series where there are consecutive terms. In the second of our four

cases, we say that the maximum of the lower section is the lower limit

of the upper section, or of any set of terms chosen out of the upper

section in such a way that no term of the upper section is before all of

them. In the third of our four cases, we say that the minimum of the

±

Stetigkeit und irrationale Zahlen, znd edition, Brunswick, 18oz.

Chap. VII. Rational, Real, and Complex Numbers ¸z

upper section is the upper limit of the lower section, or of any set of

terms chosen out of the lower section in such a way that no term of

the lower section is after all of them. In the fourth case, we say that |

;o there is a “gap”: neither the upper section nor the lower has a limit or

a last term. In this case, we may also say that we have an “irrational

section,” since sections of the series of ratios have “gaps” when they

correspond to irrationals.

What delayed the true theory of irrationals was a mistaken belief

that there must be “limits” of series of ratios. The notion of “limit” is

of the utmost importance, and before proceeding further it will be

well to deﬁne it.

A term x is said to be an “upper limit” of a class α with respect

to a relation P if (1) α has no maximum in P, (z) every member of

α which belongs to the ﬁeld of P precedes x, (j) every member of

the ﬁeld of P which precedes x precedes some member of α. (By

“precedes” we mean “has the relation P to.”)

This presupposes the following deﬁnition of a “maximum”:—

A term x is said to be a “maximum” of a class α with respect to

a relation P if x is a member of α and of the ﬁeld of P and does not

have the relation P to any other member of α.

These deﬁnitions do not demand that the terms to which they

are applied should be quantitative. For example, given a series of

moments of time arranged by earlier and later, their “maximum” (if

any) will be the last of the moments; but if they are arranged by

later and earlier, their “maximum” (if any) will be the ﬁrst of the

moments.

The “minimum” of a class with respect to P is its maximum with

respect to the converse of P; and the “lower limit” with respect to P is

the upper limit with respect to the converse of P.

The notions of limit and maximum do not essentially demand

that the relation in respect to which they are deﬁned should be serial,

but they have few important applications except to cases when the

relation is serial or quasi-serial. A notion which is often important

is the notion “upper limit or maximum,” to which we may give the

name “upper boundary.” Thus the “upper boundary” of a set of terms

chosen out of a series is their last member if they have one, but, if not,

it is the ﬁrst term after all of them, if there is such a term. If there

is neither | ;1 a maximum nor a limit, there is no upper boundary. The

“lower boundary” is the lower limit or minimum.

Reverting to the four kinds of Dedekind section, we see that in

the case of the ﬁrst three kinds each section has a boundary (upper

or lower as the case may be), while in the fourth kind neither has

a boundary. It is also clear that, whenever the lower section has an

upper boundary, the upper section has a lower boundary. In the

second and third cases, the two boundaries are identical; in the ﬁrst,

they are consecutive terms of the series.

Chap. VII. Rational, Real, and Complex Numbers ¸j

A series is called “Dedekindian” when every section has a bound-

ary, upper or lower as the case may be.

We have seen that the series of ratios in order of magnitude is not

Dedekindian.

From the habit of being inﬂuenced by spatial imagination, people

have supposed that series must have limits in cases where it seems

odd if they do not. Thus, perceiving that there was no rational limit

to the ratios whose square is less than z, they allowed themselves

to “postulate” an irrational limit, which was to ﬁll the Dedekind gap.

Dedekind, in the above-mentioned work, set up the axiom that the

gap must always be ﬁlled, i.e. that every section must have a boundary.

It is for this reason that series where his axiom is veriﬁed are called

“Dedekindian.” But there are an inﬁnite number of series for which it

is not veriﬁed.

The method of “postulating” what we want has many advantages;

they are the same as the advantages of theft over honest toil. Let us

leave them to others and proceed with our honest toil.

It is clear that an irrational Dedekind cut in some way “represents”

an irrational. In order to make use of this, which to begin with is no

more than a vague feeling, we must ﬁnd some way of eliciting from

it a precise deﬁnition; and in order to do this, we must disabuse our

minds of the notion that an irrational must be the limit of a set of

ratios. Just as ratios whose denominator is 1 are not identical with

integers, so those rational | ;z numbers which can be greater or less

than irrationals, or can have irrationals as their limits, must not be

identiﬁed with ratios. We have to deﬁne a newkind of numbers called

“real numbers,” of which some will be rational and some irrational.

Those that are rational “correspond” to ratios, in the same kind of

way in which the ratio n/1 corresponds to the integer n; but they are

not the same as ratios. In order to decide what they are to be, let us

observe that an irrational is represented by an irrational cut, and a cut

is represented by its lower section. Let us conﬁne ourselves to cuts

in which the lower section has no maximum; in this case we will call

the lower section a “segment.” Then those segments that correspond

to ratios are those that consist of all ratios less than the ratio they

correspond to, which is their boundary; while those that represent

irrationals are those that have no boundary. Segments, both those

that have boundaries and those that do not, are such that, of any two

pertaining to one series, one must be part of the other; hence they can

all be arranged in a series by the relation of whole and part. A series

in which there are Dedekind gaps, i.e. in which there are segments

that have no boundary, will give rise to more segments than it has

terms, since each term will deﬁne a segment having that term for

boundary, and then the segments without boundaries will be extra.

We are now in a position to deﬁne a real number and an irrational

number.

Chap. VII. Rational, Real, and Complex Numbers ¸±

A “real number” is a segment of the series of ratios in order of

magnitude.

An “irrational number” is a segment of the series of ratios which

has no boundary.

A “rational real number” is a segment of the series of ratios which

has a boundary.

Thus a rational real number consists of all ratios less than a certain

ratio, and it is the rational real number corresponding to that ratio.

The real number 1, for instance, is the class of proper fractions. |

In ;j the cases in which we naturally supposed that an irrational

must be the limit of a set of ratios, the truth is that it is the limit of the

corresponding set of rational real numbers in the series of segments

ordered by whole and part. For example,

√

z is the upper limit of all

those segments of the series of ratios that correspond to ratios whose

square is less than z. More simply still,

√

z is the segment consisting

of all those ratios whose square is less than z.

It is easy to prove that the series of segments of any series is

Dedekindian. For, given any set of segments, their boundary will be

their logical sum, i.e. the class of all those terms that belong to at least

one segment of the set.

¸

The above deﬁnition of real numbers is an example of “construc-

tion” as against “postulation,” of which we had another example

in the deﬁnition of cardinal numbers. The great advantage of this

method is that it requires no new assumptions, but enables us to

proceed deductively from the original apparatus of logic.

There is no diﬃculty in deﬁning addition and multiplication for

real numbers as above deﬁned. Given two real numbers µ and ν, each

being a class of ratios, take any member of µ and any member of ν

and add them together according to the rule for the addition of ratios.

Form the class of all such sums obtainable by varying the selected

members of µ and ν. This gives a new class of ratios, and it is easy

to prove that this new class is a segment of the series of ratios. We

deﬁne it as the sum of µ and ν. We may state the deﬁnition more

shortly as follows:—

The arithmetical sum of two real numbers is the class of the arith-

metical sums of a member of the one and a member of the other

chosen in all possible ways. |

We ;± can deﬁne the arithmetical product of two real numbers in

exactly the same way, by multiplying a member of the one by a

member of the other in all possible ways. The class of ratios thus

generated is deﬁned as the product of the two real numbers. (In all

such deﬁnitions, the series of ratios is to be deﬁned as excluding o

and inﬁnity.)

¸

For a fuller treatment of the subject of segments and Dedekindian relations,

see Principia Mathematica, vol. ii. ∗z1o–z1±. For a fuller treatment of real numbers,

see ibid., vol. iii. ∗j1oﬀ., and Principles of Mathematics, chaps. xxxiii. and xxxiv.

Chap. VII. Rational, Real, and Complex Numbers ¸¸

There is no diﬃculty in extending our deﬁnitions to positive and

negative real numbers and their addition and multiplication.

It remains to give the deﬁnition of complex numbers.

Complex numbers, though capable of a geometrical interpretation,

are not demanded by geometry in the same imperative way in which

irrationals are demanded. A “complex” number means a number

involving the square root of a negative number, whether integral,

fractional, or real. Since the square of a negative number is positive, a

number whose square is to be negative has to be a new sort of number.

Using the letter i for the square root of −1, any number involving the

square root of a negative number can be expressed in the form x +yi,

where x and y are real. The part yi is called the “imaginary” part of

this number, x being the “real” part. (The reason for the phrase “real

numbers” is that they are contrasted with such as are “imaginary.”)

Complex numbers have been for a long time habitually used by

mathematicians, in spite of the absence of any precise deﬁnition. It

has been simply assumed that they would obey the usual arithmetical

rules, and on this assumption their employment has been found

proﬁtable. They are required less for geometry than for algebra and

analysis. We desire, for example, to be able to say that every quadratic

equation has two roots, and every cubic equation has three, and so

on. But if we are conﬁned to real numbers, such an equation as

x

z

+1 = o has no roots, and such an equation as x

j

−1 = o has only one.

Every generalisation of number has ﬁrst presented itself as needed

for some simple problem: negative numbers were needed in order

that subtraction might be always possible, since otherwise a−b would

be meaningless if a were less than b; fractions were needed | ;¸ in order

that division might be always possible; and complex numbers are

needed in order that extraction of roots and solution of equations

may be always possible. But extensions of number are not created

by the mere need for them: they are created by the deﬁnition, and it

is to the deﬁnition of complex numbers that we must now turn our

attention.

A complex number may be regarded and deﬁned as simply an

ordered couple of real numbers. Here, as elsewhere, many deﬁnitions

are possible. All that is necessary is that the deﬁnitions adopted shall

lead to certain properties. In the case of complex numbers, if they

are deﬁned as ordered couples of real numbers, we secure at once

some of the properties required, namely, that two real numbers are

required to determine a complex number, and that among these we

can distinguish a ﬁrst and a second, and that two complex numbers

are only identical when the ﬁrst real number involved in the one

is equal to the ﬁrst involved in the other, and the second to the

second. What is needed further can be secured by deﬁning the rules

of addition and multiplication. We are to have

Chap. VII. Rational, Real, and Complex Numbers ¸6

(x +yi) +(x

+y

i) = (x +x

) +(y +y

)i

(x +yi)(x

+y

i) = (xx

−yy

) +(xy

+x

y)i.

Thus we shall deﬁne that, given two ordered couples of real numbers,

(x, y) and (x

, y

), their sum is to be the couple (x +x

, y +y

), and their

product is to be the couple (xx

−yy

, xy

+x

y). By these deﬁnitions

we shall secure that our ordered couples shall have the properties

we desire. For example, take the product of the two couples (o, y)

and (o, y

). This will, by the above rule, be the couple (−yy

, o). Thus

the square of the couple (o, 1) will be the couple (−1, o). Now those

couples in which the second term is o are those which, according

to the usual nomenclature, have their imaginary part zero; in the

notation x + yi, they are x + oi, which it is natural to write simply

x. Just as it is natural (but erroneous) | ;6 to identify ratios whose

denominator is unity with integers, so it is natural (but erroneous)

to identify complex numbers whose imaginary part is zero with real

numbers. Although this is an error in theory, it is a convenience

in practice; “x +oi” may be replaced simply by “x” and “o +yi” by

“yi,” provided we remember that the “x” is not really a real number,

but a special case of a complex number. And when y is 1, “yi” may

of course be replaced by “i.” Thus the couple (o, 1) is represented

by i, and the couple (−1, o) is represented by −1. Now our rules of

multiplication make the square of (o, 1) equal to (−1, o), i.e. the square

of i is −1. This is what we desired to secure. Thus our deﬁnitions

serve all necessary purposes.

It is easy to give a geometrical interpretation of complex numbers

in the geometry of the plane. This subject was agreeably expounded

by W. K. Cliﬀord in his Common Sense of the Exact Sciences, a book

of great merit, but written before the importance of purely logical

deﬁnitions had been realised.

Complex numbers of a higher order, though much less useful

and important than those what we have been deﬁning, have certain

uses that are not without importance in geometry, as may be seen,

for example, in Dr Whitehead’s Universal Algebra. The deﬁnition

of complex numbers of order n is obtained by an obvious extension

of the deﬁnition we have given. We deﬁne a complex number of

order n as a one-many relation whose domain consists of certain

real numbers and whose converse domain consists of the integers

from 1 to n.

6

This is what would ordinarily be indicated by the

notation (x

1

, x

z

, x

j

, . . . x

n

), where the suﬃxes denote correlation with

the integers used as suﬃxes, and the correlation is one-many, not

necessarily one-one, because x

r

and x

s

may be equal when r and s are

not equal. The above deﬁnition, with a suitable rule of multiplication,

6

Cf. Principles of Mathematics, §j6o, p. j;o.

Chap. VII. Rational, Real, and Complex Numbers ¸;

will serve all purposes for which complex numbers of higher orders

are needed.

We have now completed our review of those extensions of number

which do not involve inﬁnity. The application of number to inﬁnite

collections must be our next topic.

CHAPTER VIII

INFINITE CARDINAL NUMBERS

;; The deﬁnition of cardinal numbers which we gave in Chapter II.

was applied in Chapter III. to ﬁnite numbers, i.e. to the ordinary

natural numbers. To these we gave the name “inductive numbers,”

because we found that they are to be deﬁned as numbers which

obey mathematical induction starting from o. But we have not yet

considered collections which do not have an inductive number of

terms, nor have we inquired whether such collections can be said

to have a number at all. This is an ancient problem, which has

been solved in our own day, chieﬂy by Georg Cantor. In the present

chapter we shall attempt to explain the theory of transﬁnite or inﬁnite

cardinal numbers as it results from a combination of his discoveries

with those of Frege on the logical theory of numbers.

It cannot be said to be certain that there are in fact any inﬁnite

collections in the world. The assumption that there are is what we call

the “axiom of inﬁnity.” Although various ways suggest themselves

by which we might hope to prove this axiom, there is reason to fear

that they are all fallacious, and that there is no conclusive logical

reason for believing it to be true. At the same time, there is certainly

no logical reason against inﬁnite collections, and we are therefore

justiﬁed, in logic, in investigating the hypothesis that there are such

collections. The practical form of this hypothesis, for our present

purposes, is the assumption that, if n is any inductive number, n

is not equal to n + 1. Various subtleties arise in identifying this

form of our assumption with | ;8 the form that asserts the existence

of inﬁnite collections; but we will leave these out of account until,

in a later chapter, we come to consider the axiom of inﬁnity on its

own account. For the present we shall merely assume that, if n is an

inductive number, n is not equal to n +1. This is involved in Peano’s

assumption that no two inductive numbers have the same successor;

for, if n = n + 1, then n − 1 and n have the same successor, namely

n. Thus we are assuming nothing that was not involved in Peano’s

primitive propositions.

Let us now consider the collection of the inductive numbers them-

selves. This is a perfectly well-deﬁned class. In the ﬁrst place, a

¸8

Chap. VIII. Inﬁnite Cardinal Numbers ¸o

cardinal number is a set of classes which are all similar to each other

and are not similar to anything except each other. We then deﬁne

as the “inductive numbers” those among cardinals which belong to

the posterity of o with respect to the relation of n to n +1, i.e. those

which possess every property possessed by o and by the successors

of possessors, meaning by the “successor” of n the number n + 1.

Thus the class of “inductive numbers” is perfectly deﬁnite. By our

general deﬁnition of cardinal numbers, the number of terms in the

class of inductive numbers is to be deﬁned as “all those classes that

are similar to the class of inductive numbers”—i.e. this set of classes

is the number of the inductive numbers according to our deﬁnitions.

Now it is easy to see that this number is not one of the inductive

numbers. If n is any inductive number, the number of numbers from

o to n (both included) is n+1; therefore the total number of inductive

numbers is greater than n, no matter which of the inductive numbers

n may be. If we arrange the inductive numbers in a series in order

of magnitude, this series has no last term; but if n is an inductive

number, every series whose ﬁeld has n terms has a last term, as it is

easy to prove. Such diﬀerences might be multiplied ad lib. Thus the

number of inductive numbers is a new number, diﬀerent from all of

them, not possessing all inductive properties. It may happen that o

has a certain | ;o property, and that if n has it so has n +1, and yet that

this newnumber does not have it. The diﬃculties that so long delayed

the theory of inﬁnite numbers were largely due to the fact that some,

at least, of the inductive properties were wrongly judged to be such

as must belong to all numbers; indeed it was thought that they could

not be denied without contradiction. The ﬁrst step in understanding

inﬁnite numbers consists in realising the mistakenness of this view.

The most noteworthy and astonishing diﬀerence between an in-

ductive number and this new number is that this new number is

unchanged by adding 1 or subtracting 1 or doubling or halving or

any of a number of other operations which we think of as necessarily

making a number larger or smaller. The fact of being not altered by

the addition of 1 is used by Cantor for the deﬁnition of what he calls

“transﬁnite” cardinal numbers; but for various reasons, some of which

will appear as we proceed, it is better to deﬁne an inﬁnite cardinal

number as one which does not possess all inductive properties, i.e.

simply as one which is not an inductive number. Nevertheless, the

property of being unchanged by the addition of 1 is a very important

one, and we must dwell on it for a time.

To say that a class has a number which is not altered by the

addition of 1 is the same thing as to say that, if we take a term x

which does not belong to the class, we can ﬁnd a one-one relation

whose domain is the class and whose converse domain is obtained by

adding x to the class. For in that case, the class is similar to the sum

Chap. VIII. Inﬁnite Cardinal Numbers 6o

of itself and the term x, i.e. to a class having one extra term; so that

it has the same number as a class with one extra term, so that if n is

this number, n = n +1. In this case, we shall also have n = n −1, i.e.

there will be one-one relations whose domains consist of the whole

class and whose converse domains consist of just one term short of

the whole class. It can be shown that the cases in which this happens

are the same as the apparently more general cases in which some part

(short of the whole) can be put into one-one relation with the whole.

When this can be done, | 8o the correlator by which it is done may be

said to “reﬂect” the whole class into a part of itself; for this reason,

such classes will be called “reﬂexive.” Thus:

A “reﬂexive” class is one which is similar to a proper part of itself.

(A “proper part” is a part short of the whole.)

A“reﬂexive” cardinal number is the cardinal number of a reﬂexive

class.

We have now to consider this property of reﬂexiveness.

One of the most striking instances of a “reﬂexion” is Royce’s

illustration of the map: he imagines it decided to make a map of

England upon a part of the surface of England. Amap, if it is accurate,

has a perfect one-one correspondence with its original; thus our map,

which is part, is in one-one relation with the whole, and must contain

the same number of points as the whole, which must therefore be a

reﬂexive number. Royce is interested in the fact that the map, if it is

correct, must contain a map of the map, which must in turn contain

a map of the map of the map, and so on ad inﬁnitum. This point is

interesting, but need not occupy us at this moment. In fact, we shall

do well to pass from picturesque illustrations to such as are more

completely deﬁnite, and for this purpose we cannot do better than

consider the number-series itself.

The relation of n to n +1, conﬁned to inductive numbers, is one-

one, has the whole of the inductive numbers for its domain, and all

except o for its converse domain. Thus the whole class of inductive

numbers is similar to what the same class becomes when we omit

o. Consequently it is a “reﬂexive” class according to the deﬁnition,

and the number of its terms is a “reﬂexive” number. Again, the

relation of n to zn, conﬁned to inductive numbers, is one-one, has

the whole of the inductive numbers for its domain, and the even

inductive numbers alone for its converse domain. Hence the total

number of inductive numbers is the same as the number of even

inductive numbers. This property was used by Leibniz (and many

others) as a proof that inﬁnite numbers are impossible; it was thought

self-contradictory that | 81 “the part should be equal to the whole.” But

this is one of those phrases that depend for their plausibility upon

an unperceived vagueness: the word “equal” has many meanings,

but if it is taken to mean what we have called “similar,” there is

Chap. VIII. Inﬁnite Cardinal Numbers 61

no contradiction, since an inﬁnite collection can perfectly well have

parts similar to itself. Those who regard this as impossible have,

unconsciously as a rule, attributed to numbers in general properties

which can only be proved by mathematical induction, and which

only their familiarity makes us regard, mistakenly, as true beyond

the region of the ﬁnite.

Whenever we can “reﬂect” a class into a part of itself, the same

relation will necessarily reﬂect that part into a smaller part, and so

on ad inﬁnitum. For example, we can reﬂect, as we have just seen, all

the inductive numbers into the even numbers; we can, by the same

relation (that of n to zn) reﬂect the even numbers into the multiples

of ±, these into the multiples of 8, and so on. This is an abstract

analogue to Royce’s problem of the map. The even numbers are a

“map” of all the inductive numbers; the multiples of ± are a map of

the map; the multiples of 8 are a map of the map of the map; and so

on. If we had applied the same process to the relation of n to n +1,

our “map” would have consisted of all the inductive numbers except

o; the map of the map would have consisted of all from z onward,

the map of the map of the map of all from j onward; and so on. The

chief use of such illustrations is in order to become familiar with the

idea of reﬂexive classes, so that apparently paradoxical arithmetical

propositions can be readily translated into the language of reﬂexions

and classes, in which the air of paradox is much less.

It will be useful to give a deﬁnition of the number which is that of

the inductive cardinals. For this purpose we will ﬁrst deﬁne the kind

of series exempliﬁed by the inductive cardinals in order of magnitude.

The kind of series which is called a “progression” has already been

considered in Chapter I. It is a series which can be generated by a

relation of consecutiveness: | 8z every member of the series is to have a

successor, but there is to be just one which has no predecessor, and

every member of the series is to be in the posterity of this term with

respect to the relation “immediate predecessor.” These characteristics

may be summed up in the following deﬁnition:—

1

A “progression” is a one-one relation such that there is just one

term belonging to the domain but not to the converse domain, and

the domain is identical with the posterity of this one term.

It is easy to see that a progression, so deﬁned, satisﬁes Peano’s ﬁve

axioms. The term belonging to the domain but not to the converse

domain will be what he calls “o”; the term to which a term has the

one-one relation will be the “successor” of the term; and the domain

of the one-one relation will be what he calls “number.” Taking his

ﬁve axioms in turn, we have the following translations:—

(1) “o is a number” becomes: “The member of the domain which

is not a member of the converse domain is a member of the domain.”

1

Cf. Principia Mathematica, vol. ii. ∗1zj.

Chap. VIII. Inﬁnite Cardinal Numbers 6z

This is equivalent to the existence of such a member, which is given

in our deﬁnition. We will call this member “the ﬁrst term.”

(z) “The successor of any number is a number” becomes: “The

term to which a given member of the domain has the relation in

question is again a member of the domain.” This is proved as follows:

By the deﬁnition, every member of the domain is a member of the

posterity of the ﬁrst term; hence the successor of a member of the

domain must be a member of the posterity of the ﬁrst term (because

the posterity of a term always contains its own successors, by the gen-

eral deﬁnition of posterity), and therefore a member of the domain,

because by the deﬁnition the posterity of the ﬁrst term is the same as

the domain.

(j) “No two numbers have the same successor.” This is only to

say that the relation is one-many, which it is by deﬁnition (being

one-one). |

(±) 8j “o is not the successor of any number” becomes: “The ﬁrst

term is not a member of the converse domain,” which is again an

immediate result of the deﬁnition.

(¸) This is mathematical induction, and becomes: “Every member

of the domain belongs to the posterity of the ﬁrst term,” which was

part of our deﬁnition.

Thus progressions as we have deﬁned them have the ﬁve formal

properties from which Peano deduces arithmetic. It is easy to show

that two progressions are “similar” in the sense deﬁned for similarity

of relations in Chapter VI. We can, of course, derive a relation which

is serial from the one-one relation by which we deﬁne a progression:

the method used is that explained in Chapter IV., and the relation is

that of a term to a member of its proper posterity with respect to the

original one-one relation.

Two transitive asymmetrical relations which generate progres-

sions are similar, for the same reasons for which the corresponding

one-one relations are similar. The class of all such transitive genera-

tors of progressions is a “serial number” in the sense of Chapter VI.; it

is in fact the smallest of inﬁnite serial numbers, the number to which

Cantor has given the name ω, by which he has made it famous.

But we are concerned, for the moment, with cardinal numbers.

Since two progressions are similar relations, it follows that their

domains (or their ﬁelds, which are the same as their domains) are

similar classes. The domains of progressions form a cardinal number,

since every class which is similar to the domain of a progression is

easily shown to be itself the domain of a progression. This cardinal

number is the smallest of the inﬁnite cardinal numbers; it is the one

to which Cantor has appropriated the Hebrew Aleph with the suﬃx

o, to distinguish it from larger inﬁnite cardinals, which have other

suﬃxes. Thus the name of the smallest of inﬁnite cardinals is ℵ

o

.

Chap. VIII. Inﬁnite Cardinal Numbers 6j

To say that a class has ℵ

o

terms is the same thing as to say that it is

a member of ℵ

o

, and this is the same thing as to say | 8± that the members

of the class can be arranged in a progression. It is obvious that any

progression remains a progression if we omit a ﬁnite number of terms

from it, or every other term, or all except every tenth term or every

hundredth term. These methods of thinning out a progression do not

make it cease to be a progression, and therefore do not diminish the

number of its terms, which remains ℵ

o

. In fact, any selection from a

progression is a progression if it has no last term, however sparsely it

may be distributed. Take (say) inductive numbers of the form n

n

, or

n

n

n

. Such numbers grow very rare in the higher parts of the number

series, and yet there are just as many of them as there are inductive

numbers altogether, namely, ℵ

o

.

Conversely, we can add terms to the inductive numbers without

increasing their number. Take, for example, ratios. One might be

inclined to think that there must be many more ratios than integers,

since ratios whose denominator is 1 correspond to the integers, and

seemto be only an inﬁnitesimal proportion of ratios. But in actual fact

the number of ratios (or fractions) is exactly the same as the number

of inductive numbers, namely, ℵ

o

. This is easily seen by arranging

ratios in a series on the following plan: If the sum of numerator and

denominator in one is less than in the other, put the one before the

other; if the sum is equal in the two, put ﬁrst the one with the smaller

numerator. This gives us the series

1,

1

z

, z,

1

j

, j,

1

±

,

z

j

,

j

z

, ±,

1

¸

, . . .

This series is a progression, and all ratios occur in it sooner or later.

Hence we can arrange all ratios in a progression, and their number is

therefore ℵ

o

.

It is not the case, however, that all inﬁnite collections have ℵ

o

terms. The number of real numbers, for example, is greater than

ℵ

o

; it is, in fact, z

ℵ

o

, and it is not hard to prove that z

n

is greater

than n even when n is inﬁnite. The easiest way of proving this is

to prove, ﬁrst, that if a class has n members, it contains z

n

sub-

classes—in other words, that there are z

n

ways | 8¸ of selecting some of

its members (including the extreme cases where we select all or none);

and secondly, that the number of sub-classes contained in a class is

always greater than the number of members of the class. Of these two

propositions, the ﬁrst is familiar in the case of ﬁnite numbers, and is

not hard to extend to inﬁnite numbers. The proof of the second is so

simple and so instructive that we shall give it:

In the ﬁrst place, it is clear that the number of sub-classes of a

given class (say α) is at least as great as the number of members, since

each member constitutes a sub-class, and we thus have a correlation

of all the members with some of the sub-classes. Hence it follows that,

Chap. VIII. Inﬁnite Cardinal Numbers 6±

if the number of sub-classes is not equal to the number of members,

it must be greater. Now it is easy to prove that the number is not

equal, by showing that, given any one-one relation whose domain is

the members and whose converse domain is contained among the

set of sub-classes, there must be at least one sub-class not belonging

to the converse domain. The proof is as follows:

z

When a one-one

correlation R is established between all the members of α and some

of the sub-classes, it may happen that a given member x is correlated

with a sub-class of which it is a member; or, again, it may happen that

x is correlated with a sub-class of which it is not a member. Let us

form the whole class, β say, of those members x which are correlated

with sub-classes of which they are not members. This is a sub-class

of α, and it is not correlated with any member of α. For, taking ﬁrst

the members of β, each of them is (by the deﬁnition of β) correlated

with some sub-class of which it is not a member, and is therefore not

correlated with β. Taking next the terms which are not members of β,

each of them (by the deﬁnition of β) is correlated with some sub-class

of which it is a member, and therefore again is not correlated with β.

Thus no member of α is correlated with β. Since R was any one-one

correlation of all members | 86 with some sub-classes, it follows that

there is no correlation of all members with all sub-classes. It does

not matter to the proof if β has no members: all that happens in

that case is that the sub-class which is shown to be omitted is the

null-class. Hence in any case the number of sub-classes is not equal

to the number of members, and therefore, by what was said earlier,

it is greater. Combining this with the proposition that, if n is the

number of members, z

n

is the number of sub-classes, we have the

theorem that z

n

is always greater than n, even when n is inﬁnite.

It follows from this proposition that there is no maximum to the

inﬁnite cardinal numbers. However great an inﬁnite number n may

be, z

n

will be still greater. The arithmetic of inﬁnite numbers is

somewhat surprising until one becomes accustomed to it. We have,

for example,

ℵ

o

+1 = ℵ

o

,

ℵ

o

+n = ℵ

o

, where n is any inductive number,

ℵ

o

z

= ℵ

o

.

z

This proof is taken from Cantor, with some simpliﬁcations: see Jahresbericht

der Deutschen Mathematiker-Vereinigung, i. (18oz), p. ;;.

Chap. VIII. Inﬁnite Cardinal Numbers 6¸

(This follows from the case of the ratios, for, since a ratio is deter-

mined by a pair of inductive numbers, it is easy to see that the number

of ratios is the square of the number of inductive numbers, i.e. it is

ℵ

o

z

; but we saw that it is also ℵ

o

.)

ℵ

o

n

= ℵ

o

, where n is any inductive number.

(This follows from ℵ

o

z

= ℵ

o

by induction; for if ℵ

o

n

= ℵ

o

,

then ℵ

o

n+1

= ℵ

o

z

= ℵ

o

.)

But z

ℵ

o

> ℵ

o

.

In fact, as we shall see later, z

ℵ

o

is a very important number, namely,

the number of terms in a series which has “continuity” in the sense

in which this word is used by Cantor. Assuming space and time to be

continuous in this sense (as we commonly do in analytical geometry

and kinematics), this will be the number of points in space or of

instants in time; it will also be the number of points in any ﬁnite

portion of space, whether | 8; line, area, or volume. After ℵ

o

, z

ℵ

o

is the

most important and interesting of inﬁnite cardinal numbers.

Although addition and multiplication are always possible with

inﬁnite cardinals, subtraction and division no longer give deﬁnite

results, and cannot therefore be employed as they are employed in

elementary arithmetic. Take subtraction to begin with: so long as

the number subtracted is ﬁnite, all goes well; if the other number is

reﬂexive, it remains unchanged. Thus ℵ

o

−n = ℵ

o

, if n is ﬁnite; so far,

subtraction gives a perfectly deﬁnite result. But it is otherwise when

we subtract ℵ

o

from itself; we may then get any result, from o up to

ℵ

o

. This is easily seen by examples. From the inductive numbers,

take away the following collections of ℵ

o

terms:—

(1) All the inductive numbers—remainder, zero.

(z) All the inductive numbers from n onwards—remainder, the

numbers from o to n−1, numbering n terms in all.

(j) All the odd numbers—remainder, all the even numbers, num-

bering ℵ

o

terms.

All these are diﬀerent ways of subtracting ℵ

o

fromℵ

o

, and all give

diﬀerent results.

As regards division, very similar results follow from the fact that

ℵ

o

is unchanged when multiplied by z or j or any ﬁnite number n or

by ℵ

o

. It follows that ℵ

o

divided by ℵ

o

may have any value from 1 up

to ℵ

o

.

From the ambiguity of subtraction and division it results that

negative numbers and ratios cannot be extended to inﬁnite numbers.

Addition, multiplication, and exponentiation proceed quite satisfacto-

rily, but the inverse operations—subtraction, division, and extraction

of roots—are ambiguous, and the notions that depend upon them fail

when inﬁnite numbers are concerned.

Chap. VIII. Inﬁnite Cardinal Numbers 66

The characteristic by which we deﬁned ﬁnitude was mathematical

induction, i.e. we deﬁned a number as ﬁnite when it obeys mathemat-

ical induction starting from o, and a class as ﬁnite when its number is

ﬁnite. This deﬁnition yields the sort of result that a deﬁnition ought

to yield, namely, that the ﬁnite | 88 numbers are those that occur in

the ordinary number-series o, 1, z, j, . . . But in the present chapter,

the inﬁnite numbers we have discussed have not merely been non-

inductive: they have also been reﬂexive. Cantor used reﬂexiveness

as the deﬁnition of the inﬁnite, and believes that it is equivalent to

non-inductiveness; that is to say, he believes that every class and

every cardinal is either inductive or reﬂexive. This may be true, and

may very possibly be capable of proof; but the proofs hitherto oﬀered

by Cantor and others (including the present author in former days)

are fallacious, for reasons which will be explained when we come

to consider the “multiplicative axiom.” At present, it is not known

whether there are classes and cardinals which are neither reﬂexive

nor inductive. If n were such a cardinal, we should not have n = n+1,

but n would not be one of the “natural numbers,” and would be

lacking in some of the inductive properties. All known inﬁnite classes

and cardinals are reﬂexive; but for the present it is well to preserve

an open mind as to whether there are instances, hitherto unknown,

of classes and cardinals which are neither reﬂexive nor inductive.

Meanwhile, we adopt the following deﬁnitions:—

A ﬁnite class or cardinal is one which is inductive.

An inﬁnite class or cardinal is one which is not inductive.

All reﬂexive classes and cardinals are inﬁnite; but it is not known at

present whether all inﬁnite classes and cardinals are reﬂexive. We

shall return to this subject in Chapter XII.

CHAPTER IX

INFINITE SERIES ANDORDINALS

8o An “inﬁnite series” may be deﬁned as a series of which the ﬁeld is an

inﬁnite class. We have already had occasion to consider one kind of

inﬁnite series, namely, progressions. In this chapter we shall consider

the subject more generally.

The most noteworthy characteristic of an inﬁnite series is that its

serial number can be altered by merely re-arranging its terms. In this

respect there is a certain oppositeness between cardinal and serial

numbers. It is possible to keep the cardinal number of a reﬂexive

class unchanged in spite of adding terms to it; on the other hand, it

is possible to change the serial number of a series without adding or

taking away any terms, by mere re-arrangement. At the same time,

in the case of any inﬁnite series it is also possible, as with cardinals,

to add terms without altering the serial number: everything depends

upon the way in which they are added.

In order to make matters clear, it will be best to begin with exam-

ples. Let us ﬁrst consider various diﬀerent kinds of series which can

be made out of the inductive numbers arranged on various plans. We

start with the series

1, z, j, ±, . . . n, . . .,

which, as we have already seen, represents the smallest of inﬁnite

serial numbers, the sort that Cantor calls ω. Let us proceed to thin out

this series by repeatedly performing the | oo operation of removing to the

end the ﬁrst even number that occurs. We thus obtain in succession

the various series:

1, j, ±, ¸, . . . n, . . . z,

1, j, ¸, 6, . . . n +1, . . . z, ±,

1, j, ¸, ;, . . . n +z, . . . z, ±, 6,

and so on. If we imagine this process carried on as long as possible,

we ﬁnally reach the series

1, j, ¸, ;, . . . zn +1, . . . z, ±, 6, 8, . . . zn, . . .,

6;

Chap. IX. Inﬁnite Series and Ordinals 68

in which we have ﬁrst all the odd numbers and then all the even

numbers.

The serial numbers of these various series are ω + 1, ω + z, ω +

j, . . . zω. Each of these numbers is “greater” than any of its predeces-

sors, in the following sense:—

One serial number is said to be “greater” than another if any series

having the ﬁrst number contains a part having the second number,

but no series having the second number contains a part having the

ﬁrst number.

If we compare the two series

1, z, j, ±, . . . n, . . .

1, j, ±, ¸, . . . n +1, . . . z,

we see that the ﬁrst is similar to the part of the second which omits

the last term, namely, the number z, but the second is not similar to

any part of the ﬁrst. (This is obvious, but is easily demonstrated.)

Thus the second series has a greater serial number than the ﬁrst,

according to the deﬁnition—i.e. ω+1 is greater than ω. But if we add

a term at the beginning of a progression instead of the end, we still

have a progression. Thus 1 +ω = ω. Thus 1 +ω is not equal to ω+1.

This is characteristic of relation-arithmetic generally: if µ and ν are

two relation-numbers, the general rule is that µ +ν is not equal to

ν +µ. The case of ﬁnite ordinals, in which there is equality, is quite

exceptional.

The series we ﬁnally reached just now consisted of ﬁrst all the

odd numbers and then all the even numbers, and its serial | o1 number

is zω. This number is greater than ω or ω+n, where n is ﬁnite. It is to

be observed that, in accordance with the general deﬁnition of order,

each of these arrangements of integers is to be regarded as resulting

from some deﬁnite relation. E.g. the one which merely removes z to

the end will be deﬁned by the following relation: “x and y are ﬁnite

integers, and either y is z and x is not z, or neither is z and x is less

than y.” The one which puts ﬁrst all the odd numbers and then all

the even ones will be deﬁned by: “x and y are ﬁnite integers, and

either x is odd and y is even or x is less than y and both are odd or

both are even.” We shall not trouble, as a rule, to give these formulæ

in future; but the fact that they could be given is essential.

The number which we have called zω, namely, the number of

a series consisting of two progressions, is sometimes called ω . z.

Multiplication, like addition, depends upon the order of the factors:

a progression of couples gives a series such as

x

1

, y

1

, x

z

, y

z

, x

j

, y

j

, . . . x

n

, y

n

, . . . ,

which is itself a progression; but a couple of progressions gives a

series which is twice as long as a progression. It is therefore necessary

Chap. IX. Inﬁnite Series and Ordinals 6o

to distinguish between zω and ω. z. Usage is variable; we shall use

zω for a couple of progressions and ω. z for a progression of couples,

and this decision of course governs our general interpretation of “α.β”

when α and β are relation-numbers: “α . β” will have to stand for a

suitably constructed sum of α relations each having β terms.

We can proceed indeﬁnitely with the process of thinning out the

inductive numbers. For example, we can place ﬁrst the odd numbers,

then their doubles, then the doubles of these, and so on. We thus

obtain the series

1, j, ¸, ;, . . .; z, 6, 1o, 1±, . . .; ±, 1z, zo, z8, . . .;

8, z±, ±o, ¸6, . . .,

of which the number is ω

z

, since it is a progression of progressions.

Any one of the progressions in this new series can of course be |

oz thinned out as we thinned out our original progression. We can

proceed to ω

j

, ω

±

, . . . ω

ω

, and so on; however far we have gone, we

can always go further.

The series of all the ordinals that can be obtained in this way, i.e.

all that can be obtained by thinning out a progression, is itself longer

than any series that can be obtained by re-arranging the terms of a

progression. (This is not diﬃcult to prove.) The cardinal number of

the class of such ordinals can be shown to be greater than ℵ

o

; it is the

number which Cantor calls ℵ

1

. The ordinal number of the series of all

ordinals that can be made out of an ℵ

o

, taken in order of magnitude,

is called ω

1

. Thus a series whose ordinal number is ω

1

has a ﬁeld

whose cardinal number is ℵ

1

.

We can proceed from ω

1

and ℵ

1

to ω

z

and ℵ

z

by a process exactly

analogous to that by which we advanced from ω and ℵ

o

to ω

1

and ℵ

1

.

And there is nothing to prevent us from advancing indeﬁnitely in this

way to new cardinals and new ordinals. It is not known whether z

ℵ

o

is equal to any of the cardinals in the series of Alephs. It is not even

known whether it is comparable with them in magnitude; for aught

we know, it may be neither equal to nor greater nor less than any

one of the Alephs. This question is connected with the multiplicative

axiom, of which we shall treat later.

All the series we have been considering so far in this chapter

have been what is called “well-ordered.” A well-ordered series is one

which has a beginning, and has consecutive terms, and has a termnext

after any selection of its terms, provided there are any terms after the

selection. This excludes, on the one hand, compact series, in which

there are terms between any two, and on the other hand series which

have no beginning, or in which there are subordinate parts having

no beginning. The series of negative integers in order of magnitude,

having no beginning, but ending with −1, is not well-ordered; but

Chap. IX. Inﬁnite Series and Ordinals ;o

taken in the reverse order, beginning with −1, it is well-ordered,

being in fact a progression. The deﬁnition is: |

A oj “well-ordered” series is one in which every sub-class (except, of

course, the null-class) has a ﬁrst term.

An “ordinal” number means the relation-number of a well-ordered

series. It is thus a species of serial number.

Among well-ordered series, a generalised form of mathematical

induction applies. A property may be said to be “transﬁnitely heredi-

tary” if, when it belongs to a certain selection of the terms in a series,

it belongs to their immediate successor provided they have one. In a

well-ordered series, a transﬁnitely hereditary property belonging to

the ﬁrst term of the series belongs to the whole series. This makes it

possible to prove many propositions concerning well-ordered series

which are not true of all series.

It is easy to arrange the inductive numbers in series which are

not well-ordered, and even to arrange them in compact series. For

example, we can adopt the following plan: consider the decimals

from ·1 (inclusive) to 1 (exclusive), arranged in order of magnitude.

These form a compact series; between any two there are always an

inﬁnite number of others. Now omit the dot at the beginning of each,

and we have a compact series consisting of all ﬁnite integers except

such as divide by 1o. If we wish to include those that divide by 1o,

there is no diﬃculty; instead of starting with ·1, we will include all

decimals less than 1, but when we remove the dot, we will transfer to

the right any o’s that occur at the beginning of our decimal. Omitting

these, and returning to the ones that have no o’s at the beginning,

we can state the rule for the arrangement of our integers as follows:

Of two integers that do not begin with the same digit, the one that

begins with the smaller digit comes ﬁrst. Of two that do begin with

the same digit, but diﬀer at the second digit, the one with the smaller

second digit comes ﬁrst, but ﬁrst of all the one with no second digit;

and so on. Generally, if two integers agree as regards the ﬁrst n dig-

its, but not as regards the (n + 1)

th

, that one comes ﬁrst which has

either no (n +1)

th

digit or a smaller one than the other. This rule of

arrangement, | o± as the reader can easily convince himself, gives rise

to a compact series containing all the integers not divisible by 1o;

and, as we saw, there is no diﬃculty about including those that are

divisible by 1o. It follows from this example that it is possible to

construct compact series having ℵ

o

terms. In fact, we have already

seen that there are ℵ

o

ratios, and ratios in order of magnitude form a

compact series; thus we have here another example. We shall resume

this topic in the next chapter.

Of the usual formal laws of addition, multiplication, and expo-

nentiation, all are obeyed by transﬁnite cardinals, but only some are

obeyed by transﬁnite ordinals, and those that are obeyed by them are

Chap. IX. Inﬁnite Series and Ordinals ;1

obeyed by all relation-numbers. By the “usual formal laws” we mean

the following:—

I. The commutative law:

α +β = β +α and α ×β = β ×α.

II. The associative law:

(α +β) +γ = α +(β +γ) and (α ×β) ×γ = α ×(β ×γ).

III. The distributive law:

α(β +γ) = αβ +αγ.

When the commutative law does not hold, the above form of the

distributive law must be distinguished from

(β +γ)α = βα +γα.

As we shall see immediately, one form may be true and the other

false.

IV. The laws of exponentiation:

α

β

. α

γ

= α

β+γ

, α

γ

. β

γ

= (αβ)

γ

, (α

β

)

γ

= α

βγ

.

All these laws hold for cardinals, whether ﬁnite or inﬁnite, and

for ﬁnite ordinals. But when we come to inﬁnite ordinals, or indeed

to relation-numbers in general, some hold and some do not. The

commutative law does not hold; the associative law does hold; the

distributive law (adopting the convention | o¸ we have adopted above as

regards the order of the factors in a product) holds in the form

(β +γ)α = βα +γα,

but not in the form

α(β +γ) = αβ +αγ;

the exponential laws

α

β

. α

γ

= α

β+γ

and (α

β

)

γ

= α

βγ

still hold, but not the law

α

γ

. β

γ

= (αβ)

γ

,

which is obviously connected with the commutative law for multipli-

cation.

The deﬁnitions of multiplication and exponentiation that are

assumed in the above propositions are somewhat complicated. The

reader who wishes to know what they are and how the above laws

are proved must consult the second volume of Principia Mathematica,

∗1;z–1;6.

Chap. IX. Inﬁnite Series and Ordinals ;z

Ordinal transﬁnite arithmetic was developed by Cantor at an ear-

lier stage than cardinal transﬁnite arithmetic, because it has various

technical mathematical uses which led him to it. But from the point

of view of the philosophy of mathematics it is less important and

less fundamental than the theory of transﬁnite cardinals. Cardinals

are essentially simpler than ordinals, and it is a curious historical

accident that they ﬁrst appeared as an abstraction from the latter,

and only gradually came to be studied on their own account. This

does not apply to Frege’s work, in which cardinals, ﬁnite and transﬁ-

nite, were treated in complete independence of ordinals; but it was

Cantor’s work that made the world aware of the subject, while Frege’s

remained almost unknown, probably in the main on account of the

diﬃculty of his symbolism. And mathematicians, like other people,

have more diﬃculty in understanding and using notions which are

comparatively “simple” in the logical sense than in manipulating

more complex notions which are | o6 more akin to their ordinary prac-

tice. For these reasons, it was only gradually that the true importance

of cardinals in mathematical philosophy was recognised. The im-

portance of ordinals, though by no means small, is distinctly less

than that of cardinals, and is very largely merged in that of the more

general conception of relation-numbers.

CHAPTER X

LIMITS ANDCONTINUITY

o; The conception of a “limit” is one of which the importance in math-

ematics has been found continually greater than had been thought.

The whole of the diﬀerential and integral calculus, indeed practically

everything in higher mathematics, depends upon limits. Formerly, it

was supposed that inﬁnitesimals were involved in the foundations of

these subjects, but Weierstrass showed that this is an error: wherever

inﬁnitesimals were thought to occur, what really occurs is a set of ﬁ-

nite quantities having zero for their lower limit. It used to be thought

that “limit” was an essentially quantitative notion, namely, the notion

of a quantity to which others approached nearer and nearer, so that

among those others there would be some diﬀering by less than any

assigned quantity. But in fact the notion of “limit” is a purely ordinal

notion, not involving quantity at all (except by accident when the

series concerned happens to be quantitative). A given point on a

line may be the limit of a set of points on the line, without its be-

ing necessary to bring in co-ordinates or measurement or anything

quantitative. The cardinal number ℵ

o

is the limit (in the order of

magnitude) of the cardinal numbers 1, z, j, . . . n, . . . , although the

numerical diﬀerence between ℵ

o

and a ﬁnite cardinal is constant

and inﬁnite: from a quantitative point of view, ﬁnite numbers get

no nearer to ℵ

o

as they grow larger. What makes ℵ

o

the limit of the

ﬁnite numbers is the fact that, in the series, it comes immediately

after them, which is an ordinal fact, not a quantitative fact. |

There o8 are various forms of the notion of “limit,” of increasing

complexity. The simplest and most fundamental form, from which

the rest are derived, has been already deﬁned, but we will here repeat

the deﬁnitions which lead to it, in a general form in which they do not

demand that the relation concerned shall be serial. The deﬁnitions

are as follows:—

The “minima” of a class α with respect to a relation P are those

members of α and the ﬁeld of P (if any) to which no member of α has

the relation P.

The “maxima” with respect to P are the minima with respect to

the converse of P.

;j

Chap. X. Limits and Continuity ;±

The “sequents” of a class α with respect to a relation P are the

minima of the “successors” of α, and the “successors” of α are those

members of the ﬁeld of P to which every member of the common part

of α and the ﬁeld of P has the relation P.

The “precedents” with respect to P are the sequents with respect

to the converse of P.

The “upper limits” of α with respect to P are the sequents pro-

vided α has no maximum; but if α has a maximum, it has no upper

limits.

The “lower limits” with respect to P are the upper limits with

respect to the converse of P.

Whenever P has connexity, a class can have at most one maximum,

one minimum, one sequent, etc. Thus, in the cases we are concerned

with in practice, we can speak of “the limit” (if any).

When P is a serial relation, we can greatly simplify the above

deﬁnition of a limit. We can, in that case, deﬁne ﬁrst the “boundary”

of a class α, i.e. its limit or maximum, and then proceed to distinguish

the case where the boundary is the limit from the case where it is a

maximum. For this purpose it is best to use the notion of “segment.”

We will speak of the “segment of P deﬁned by a class α” as all

those terms that have the relation P to some one or more of the

members of α. This will be a segment in the sense deﬁned | oo in

Chapter VII.; indeed, every segment in the sense there deﬁned is the

segment deﬁned by some class α. If P is serial, the segment deﬁned

by α consists of all the terms that precede some term or other of α.

If α has a maximum, the segment will be all the predecessors of the

maximum. But if α has no maximum, every member of α precedes

some other member of α, and the whole of α is therefore included in

the segment deﬁned by α. Take, for example, the class consisting of

the fractions

1

z

,

j

±

,

;

8

,

1¸

16

, . . .,

i.e. of all fractions of the form 1 −1/z

n

for diﬀerent ﬁnite values of

n. This series of fractions has no maximum, and it is clear that the

segment which it deﬁnes (in the whole series of fractions in order of

magnitude) is the class of all proper fractions. Or, again, consider the

prime numbers, considered as a selection from the cardinals (ﬁnite

and inﬁnite) in order of magnitude. In this case the segment deﬁned

consists of all ﬁnite integers.

Assuming that P is serial, the “boundary” of a class α will be the

term x (if it exists) whose predecessors are the segment deﬁned by α.

A “maximum” of α is a boundary which is a member of α.

An “upper limit” of α is a boundary which is not a member of α.

If a class has no boundary, it has neither maximum nor limit. This

is the case of an “irrational” Dedekind cut, or of what is called a

“gap.”

Chap. X. Limits and Continuity ;¸

Thus the “upper limit” of a set of terms α with respect to a series

P is that term x (if it exists) which comes after all the α’s, but is such

that every earlier term comes before some of the α’s.

We may deﬁne all the “upper limiting-points” of a set of terms

β as all those that are the upper limits of sets of terms chosen out

of β. We shall, of course, have to distinguish upper limiting-points

from lower limiting-points. If we consider, for example, the series of

ordinal numbers:

1, z, j, . . . ω, ω+1, . . . zω, zω+1, . . . jω, . . . ω

z

, . . . ω

j

, . . . , |

the 1oo upper limiting-points of the ﬁeld of this series are those that have

no immediate predecessors, i.e.

1, ω, zω, jω, . . . ω

z

, ω

z

+ω, . . . zω

z

, . . . ω

j

. . .

The upper limiting-points of the ﬁeld of this new series will be

1, ω

z

, zω

z

, . . . ω

j

, ω

j

+ω

z

. . .

On the other hand, the series of ordinals—and indeed every well-

ordered series—has no lower limiting-points, because there are no

terms except the last that have no immediate successors. But if we

consider such a series as the series of ratios, every member of this

series is both an upper and a lower limiting-point for suitably chosen

sets. If we consider the series of real numbers, and select out of it

the rational real numbers, this set (the rationals) will have all the real

numbers as upper and lower limiting-points. The limiting-points of

a set are called its “ﬁrst derivative,” and the limiting-points of the

ﬁrst derivative are called the second derivative, and so on.

With regard to limits, we may distinguish various grades of what

may be called “continuity” in a series. The word “continuity” had

been used for a long time, but had remained without any precise

deﬁnition until the time of Dedekind and Cantor. Each of these two

men gave a precise signiﬁcance to the term, but Cantor’s deﬁnition is

narrower than Dedekind’s: a series which has Cantorian continuity

must have Dedekindian continuity, but the converse does not hold.

The ﬁrst deﬁnition that would naturally occur to a man seeking

a precise meaning for the continuity of series would be to deﬁne it

as consisting in what we have called “compactness,” i.e. in the fact

that between any two terms of the series there are others. But this

would be an inadequate deﬁnition, because of the existence of “gaps”

in series such as the series of ratios. We saw in Chapter VII. that there

are innumerable ways in which the series of ratios can be divided

into two parts, of which one wholly precedes the other, and of which

the ﬁrst has no last term, | 1o1 while the second has no ﬁrst term. Such a

state of aﬀairs seems contrary to the vague feeling we have as to what

Chap. X. Limits and Continuity ;6

should characterise “continuity,” and, what is more, it shows that

the series of ratios is not the sort of series that is needed for many

mathematical purposes. Take geometry, for example: we wish to be

able to say that when two straight lines cross each other they have a

point in common, but if the series of points on a line were similar to

the series of ratios, the two lines might cross in a “gap” and have no

point in common. This is a crude example, but many others might

be given to show that compactness is inadequate as a mathematical

deﬁnition of continuity.

It was the needs of geometry, as much as anything, that led to

the deﬁnition of “Dedekindian” continuity. It will be remembered

that we deﬁned a series as Dedekindian when every sub-class of the

ﬁeld has a boundary. (It is suﬃcient to assume that there is always an

upper boundary, or that there is always a lower boundary. If one of

these is assumed, the other can be deduced.) That is to say, a series is

Dedekindian when there are no gaps. The absence of gaps may arise

either through terms having successors, or through the existence of

limits in the absence of maxima. Thus a ﬁnite series or a well-ordered

series is Dedekindian, and so is the series of real numbers. The former

sort of Dedekindian series is excluded by assuming that our series is

compact; in that case our series must have a property which may, for

many purposes, be ﬁttingly called continuity. Thus we are led to the

deﬁnition:

A series has “Dedekindian continuity” when it is Dedekindian

and compact.

But this deﬁnition is still too wide for many purposes. Suppose,

for example, that we desire to be able to assign such properties to

geometrical space as shall make it certain that every point can be

speciﬁed by means of co-ordinates which are real numbers: this is not

insured by Dedekindian continuity alone. We want to be sure that

every point which cannot be speciﬁed by rational co-ordinates can be

speciﬁed as the limit of a progression of points | 1oz whose co-ordinates

are rational, and this is a further property which our deﬁnition does

not enable us to deduce.

We are thus led to a closer investigation of series with respect to

limits. This investigation was made by Cantor and formed the basis

of his deﬁnition of continuity, although, in its simplest form, this

deﬁnition somewhat conceals the considerations which have given

rise to it. We shall, therefore, ﬁrst travel through some of Cantor’s

conceptions in this subject before giving his deﬁnition of continuity.

Cantor deﬁnes a series as “perfect” when all its points are limiting-

points and all its limiting-points belong to it. But this deﬁnition does

not express quite accurately what he means. There is no correction

required so far as concerns the property that all its points are to be

limiting-points; this is a property belonging to compact series, and to

Chap. X. Limits and Continuity ;;

no others if all points are to be upper limiting- or all lower limiting-

points. But if it is only assumed that they are limiting-points one

way, without specifying which, there will be other series that will

have the property in question—for example, the series of decimals

in which a decimal ending in a recurring o is distinguished from the

corresponding terminating decimal and placed immediately before it.

Such a series is very nearly compact, but has exceptional terms which

are consecutive, and of which the ﬁrst has no immediate predecessor,

while the second has no immediate successor. Apart from such series,

the series in which every point is a limiting-point are compact series;

and this holds without qualiﬁcation if it is speciﬁed that every point

is to be an upper limiting-point (or that every point is to be a lower

limiting-point).

Although Cantor does not explicitly consider the matter, we must

distinguish diﬀerent kinds of limiting-points according to the nature

of the smallest sub-series by which they can be deﬁned. Cantor

assumes that they are to be deﬁned by progressions, or by regressions

(which are the converses of progressions). When every member of

our series is the limit of a progression or regression, Cantor calls our

series “condensed in itself” (insichdicht). |

We 1oj come now to the second property by which perfection was to

be deﬁned, namely, the property which Cantor calls that of being

“closed” (abgeschlossen). This, as we saw, was ﬁrst deﬁned as consisting

in the fact that all the limiting-points of a series belong to it. But this

only has any eﬀective signiﬁcance if our series is given as contained

in some other larger series (as is the case, e.g., with a selection of

real numbers), and limiting-points are taken in relation to the larger

series. Otherwise, if a series is considered simply on its own account,

it cannot fail to contain its limiting-points. What Cantor means is not

exactly what he says; indeed, on other occasions he says something

rather diﬀerent, which is what he means. What he really means is that

every subordinate series which is of the sort that might be expected

to have a limit does have a limit within the given series; i.e. every

subordinate series which has no maximum has a limit, i.e. every

subordinate series has a boundary. But Cantor does not state this for

every subordinate series, but only for progressions and regressions.

(It is not clear how far he recognises that this is a limitation.) Thus,

ﬁnally, we ﬁnd that the deﬁnition we want is the following:—

A series is said to be “closed” (abgeschlossen) when every progres-

sion or regression contained in the series has a limit in the series.

We then have the further definition:—

A series is “perfect” when it is condensed in itself and closed, i.e.

when every term is the limit of a progression or regression, and every

progression or regression contained in the series has a limit in the

series.

Chap. X. Limits and Continuity ;8

In seeking a deﬁnition of continuity, what Cantor has in mind

is the search for a deﬁnition which shall apply to the series of real

numbers and to any series similar to that, but to no others. For

this purpose we have to add a further property. Among the real

numbers some are rational, some are irrational; although the number

of irrationals is greater than the number of rationals, yet there are

rationals between any two real numbers, however | 1o± little the two may

diﬀer. The number of rationals, as we saw, is ℵ

o

. This gives a further

property which suﬃces to characterise continuity completely, namely,

the property of containing a class of ℵ

o

members in such a way that

some of this class occur between any two terms of our series, however

near together. This property, added to perfection, suﬃces to deﬁne

a class of series which are all similar and are in fact a serial number.

This class Cantor deﬁnes as that of continuous series.

We may slightly simplify his deﬁnition. To begin with, we say:

A “median class” of a series is a sub-class of the ﬁeld such that

members of it are to be found between any two terms of the series.

Thus the rationals are a median class in the series of real numbers.

It is obvious that there cannot be median classes except in compact

series.

We then ﬁnd that Cantor’s deﬁnition is equivalent to the follow-

ing:—

A series is “continuous” when (1) it is Dedekindian, (z) it contains

a median class having ℵ

o

terms.

To avoid confusion, we shall speak of this kind as “Cantorian

continuity.” It will be seen that it implies Dedekindian continuity,

but the converse is not the case. All series having Cantorian continuity

are similar, but not all series having Dedekindian continuity.

The notions of limit and continuity which we have been deﬁning

must not be confounded with the notions of the limit of a function for

approaches to a given argument, or the continuity of a function in the

neighbourhood of a given argument. These are diﬀerent notions, very

important, but derivative from the above and more complicated. The

continuity of motion (if motion is continuous) is an instance of the

continuity of a function; on the other hand, the continuity of space

and time (if they are continuous) is an instance of the continuity of

series, or (to speak more cautiously) of a kind of continuity which

can, by suﬃcient mathematical | 1o¸ manipulation, be reduced to the

continuity of series. In viewof the fundamental importance of motion

in applied mathematics, as well as for other reasons, it will be well

to deal brieﬂy with the notions of limits and continuity as applied

to functions; but this subject will be best reserved for a separate

chapter.

The deﬁnitions of continuity which we have been considering,

namely, those of Dedekind and Cantor, do not correspond very closely

Chap. X. Limits and Continuity ;o

to the vague idea which is associated with the word in the mind of

the man in the street or the philosopher. They conceive continuity

rather as absence of separateness, the sort of general obliteration of

distinctions which characterises a thick fog. Afog gives an impression

of vastness without deﬁnite multiplicity or division. It is this sort of

thing that a metaphysician means by “continuity,” declaring it, very

truly, to be characteristic of his mental life and of that of children

and animals.

The general idea vaguely indicated by the word “continuity” when

so employed, or by the word “ﬂux,” is one which is certainly quite

diﬀerent from that which we have been deﬁning. Take, for example,

the series of real numbers. Each is what it is, quite deﬁnitely and

uncompromisingly; it does not pass over by imperceptible degrees

into another; it is a hard, separate unit, and its distance from every

other unit is ﬁnite, though it can be made less than any given ﬁnite

amount assigned in advance. The question of the relation between

the kind of continuity existing among the real numbers and the

kind exhibited, e.g. by what we see at a given time, is a diﬃcult

and intricate one. It is not to be maintained that the two kinds are

simply identical, but it may, I think, be very well maintained that

the mathematical conception which we have been considering in this

chapter gives the abstract logical scheme to which it must be possible

to bring empirical material by suitable manipulation, if that material

is to be called “continuous” in any precisely deﬁnable sense. It would

be quite impossible | 1o6 to justify this thesis within the limits of the

present volume. The reader who is interested may read an attempt

to justify it as regards time in particular by the present author in

the Monist for 1o1±−¸, as well as in parts of Our Knowledge of the

External World. With these indications, we must leave this problem,

interesting as it is, in order to return to topics more closely connected

with mathematics.

CHAPTER XI

LIMITS ANDCONTINUITY OF

FUNCTIONS

1o; In this chapter we shall be concerned with the deﬁnition of the limit

of a function (if any) as the argument approaches a given value, and

also with the deﬁnition of what is meant by a “continuous function.”

Both of these ideas are somewhat technical, and would hardly de-

mand treatment in a mere introduction to mathematical philosophy

but for the fact that, especially through the so-called inﬁnitesimal

calculus, wrong views upon our present topics have become so ﬁrmly

embedded in the minds of professional philosophers that a prolonged

and considerable eﬀort is required for their uprooting. It has been

thought ever since the time of Leibniz that the diﬀerential and in-

tegral calculus required inﬁnitesimal quantities. Mathematicians

(especially Weierstrass) proved that this is an error; but errors incor-

porated, e.g. in what Hegel has to say about mathematics, die hard,

and philosophers have tended to ignore the work of such men as

Weierstrass.

Limits and continuity of functions, in works on ordinary mathe-

matics, are deﬁned in terms involving number. This is not essential,

as Dr Whitehead has shown.

1

We will, however, begin with the deﬁ-

nitions in the text-books, and proceed afterwards to show how these

deﬁnitions can be generalised so as to apply to series in general, and

not only to such as are numerical or numerically measurable.

Let us consider any ordinary mathematical function f x, where

| 1o8 x and f x are both real numbers, and f x is one-valued—i.e. when

x is given, there is only one value that f x can have. We call x the

“argument,” and f x the “value for the argument x.” When a function

is what we call “continuous,” the rough idea for which we are seeking

a precise deﬁnition is that small diﬀerences in x shall correspond

to small diﬀerences in f x, and if we make the diﬀerences in x small

enough, we can make the diﬀerences in f x fall below any assigned

amount. We do not want, if a function is to be continuous, that

1

See Principia Mathematica, vol. ii. ∗zjo–zj±.

8o

Chap. XI. Limits and Continuity of Functions 81

there shall be sudden jumps, so that, for some value of x, any change,

however small, will make a change in f x which exceeds some assigned

ﬁnite amount. The ordinary simple functions of mathematics have

this property: it belongs, for example, to x

z

, x

j

, . . . logx, sinx, and so

on. But it is not at all diﬃcult to deﬁne discontinuous functions. Take,

as a non-mathematical example, “the place of birth of the youngest

person living at time t.” This is a function of t; its value is constant

from the time of one person’s birth to the time of the next birth, and

then the value changes suddenly from one birthplace to the other. An

analogous mathematical example would be “the integer next below

x,” where x is a real number. This function remains constant from

one integer to the next, and then gives a sudden jump. The actual fact

is that, though continuous functions are more familiar, they are the

exceptions: there are inﬁnitely more discontinuous functions than

continuous ones.

Many functions are discontinuous for one or several values of the

variable, but continuous for all other values. Take as an example

sin1/x. The function sin θ passes through all values from −1 to 1

every time that θ passes from −π/z to π/z, or from π/z to jπ/z, or

generally from (zn−1)π/z to (zn+1)π/z, where n is any integer. Now

if we consider 1/x when x is very small, we see that as x diminishes

1/x grows faster and faster, so that it passes more and more quickly

through the cycle of values from one multiple of π/z to another as x

becomes smaller and smaller. Consequently sin1/x passes more and

more quickly from−1 | 1oo to 1 and back again, as x grows smaller. In fact,

if we take any interval containing o, say the interval from − to +

where is some very small number, sin1/x will go through an inﬁnite

number of oscillations in this interval, and we cannot diminish the

oscillations by making the interval smaller. Thus round about the

argument o the function is discontinuous. It is easy to manufacture

functions which are discontinuous in several places, or in ℵ

o

places,

or everywhere. Examples will be found in any book on the theory of

functions of a real variable.

Proceeding now to seek a precise deﬁnition of what is meant

by saying that a function is continuous for a given argument, when

argument and value are both real numbers, let us ﬁrst deﬁne a “neigh-

bourhood” of a number x as all the numbers from x− to x+, where

is some number which, in important cases, will be very small. It is

clear that continuity at a given point has to do with what happens in

any neighbourhood of that point, however small.

What we desire is this: If a is the argument for which we wish

our function to be continuous, let us ﬁrst deﬁne a neighbourhood (α

say) containing the value f a which the function has for the argument

a; we desire that, if we take a suﬃciently small neighbourhood con-

taining a, all values for arguments throughout this neighbourhood

Chap. XI. Limits and Continuity of Functions 8z

shall be contained in the neighbourhood α, no matter how small we

may have made α. That is to say, if we decree that our function is

not to diﬀer from f a by more than some very tiny amount, we can

always ﬁnd a stretch of real numbers, having a in the middle of it,

such that throughout this stretch f x will not diﬀer from f a by more

than the prescribed tiny amount. And this is to remain true what-

ever tiny amount we may select. Hence we are led to the following

deﬁnition:—

The function f (x) is said to be “continuous” for the argument a

if, for every positive number σ, diﬀerent from o, but as small as we

please, there exists a positive number , diﬀerent from o, such that,

for all values of δ which are numerically | 11o less

z

than , the diﬀerence

f (a +δ) −f (a) is numerically less than σ.

In this deﬁnition, σ ﬁrst deﬁnes a neighbourhood of f (a), namely,

the neighbourhood from f (a) − σ to f (a) + σ. The deﬁnition then

proceeds to say that we can (by means of ) deﬁne a neighbourhood,

namely, that from a − to a +, such that, for all arguments within

this neighbourhood, the value of the function lies within the neigh-

bourhood from f (a) − σ to f (a) +σ. If this can be done, however σ

may be chosen, the function is “continuous” for the argument a.

So far we have not deﬁned the “limit” of a function for a given

argument. If we had done so, we could have deﬁned the continuity

of a function diﬀerently: a function is continuous at a point where

its value is the same as the limit of its values for approaches either

from above or from below. But it is only the exceptionally “tame”

function that has a deﬁnite limit as the argument approaches a given

point. The general rule is that a function oscillates, and that, given

any neighbourhood of a given argument, however small, a whole

stretch of values will occur for arguments within this neighbourhood.

As this is the general rule, let us consider it ﬁrst.

Let us consider what may happen as the argument approaches

some value a from below. That is to say, we wish to consider what

happens for arguments contained in the interval froma− to a, where

is some number which, in important cases, will be very small.

The values of the function for arguments from a − to a (a ex-

cluded) will be a set of real numbers which will deﬁne a certain

section of the set of real numbers, namely, the section consisting of

those numbers that are not greater than all the values for arguments

from a − to a. Given any number in this section, there are values at

least as great as this number for arguments between a − and a, i.e.

for arguments that fall very little short | 111 of a (if is very small). Let

us take all possible ’s and all possible corresponding sections. The

common part of all these sections we will call the “ultimate section”

z

A number is said to be “numerically less” than when it lies between − and

+.

Chap. XI. Limits and Continuity of Functions 8j

as the argument approaches a. To say that a number z belongs to the

ultimate section is to say that, however small we may make , there

are arguments between a− and a for which the value of the function

is not less than z.

We may apply exactly the same process to upper sections, i.e. to

sections that go from some point up to the top, instead of from the

bottom up to some point. Here we take those numbers that are not

less than all the values for arguments from a − to a; this deﬁnes an

upper section which will vary as varies. Taking the common part of

all such sections for all possible ’s, we obtain the “ultimate upper

section.” To say that a number z belongs to the ultimate upper section

is to say that, however small we make , there are arguments between

a − and a for which the value of the function is not greater than z.

If a term z belongs both to the ultimate section and to the ultimate

upper section, we shall say that it belongs to the “ultimate oscillation.”

We may illustrate the matter by considering once more the function

sin1/x as x approaches the value o. We shall assume, in order to ﬁt in

with the above deﬁnitions, that this value is approached from below.

Let us begin with the “ultimate section.” Between − and o, what-

ever may be, the function will assume the value 1 for certain argu-

ments, but will never assume any greater value. Hence the ultimate

section consists of all real numbers, positive and negative, up to and

including 1; i.e. it consists of all negative numbers together with o,

together with the positive numbers up to and including 1.

Similarly the “ultimate upper section” consists of all positive

numbers together with o, together with the negative numbers down

to and including −1.

Thus the “ultimate oscillation” consists of all real numbers from

−1 to 1, both included. |

We 11z may say generally that the “ultimate oscillation” of a function

as the argument approaches a from below consists of all those num-

bers x which are such that, however near we come to a, we shall still

ﬁnd values as great as x and values as small as x.

The ultimate oscillation may contain no terms, or one term, or

many terms. In the ﬁrst two cases the function has a deﬁnite limit for

approaches from below. If the ultimate oscillation has one term, this

is fairly obvious. It is equally true if it has none; for it is not diﬃcult

to prove that, if the ultimate oscillation is null, the boundary of the

ultimate section is the same as that of the ultimate upper section, and

may be deﬁned as the limit of the function for approaches from below.

But if the ultimate oscillation has many terms, there is no deﬁnite

limit to the function for approaches from below. In this case we

can take the lower and upper boundaries of the ultimate oscillation

(i.e. the lower boundary of the ultimate upper section and the upper

boundary of the ultimate section) as the lower and upper limits of

Chap. XI. Limits and Continuity of Functions 8±

its “ultimate” values for approaches from below. Similarly we obtain

lower and upper limits of the “ultimate” values for approaches from

above. Thus we have, in the general case, four limits to a function for

approaches to a given argument. The limit for a given argument a only

exists when all these four are equal, and is then their common value.

If it is also the value for the argument a, the function is continuous

for this argument. This may be taken as deﬁning continuity: it is

equivalent to our former deﬁnition.

We can deﬁne the limit of a function for a given argument (if it

exists) without passing through the ultimate oscillation and the four

limits of the general case. The deﬁnition proceeds, in that case, just

as the earlier deﬁnition of continuity proceeded. Let us deﬁne the

limit for approaches from below. If there is to be a deﬁnite limit

for approaches to a from below, it is necessary and suﬃcient that,

given any small number σ, two values for arguments suﬃciently

near to a (but both less than a) will diﬀer | 11j by less than σ; i.e. if is

suﬃciently small, and our arguments both lie between a − and a (a

excluded), then the diﬀerence between the values for these arguments

will be less than σ. This is to hold for any σ, however small; in that

case the function has a limit for approaches from below. Similarly

we deﬁne the case when there is a limit for approaches from above.

These two limits, even when both exist, need not be identical; and if

they are identical, they still need not be identical with the value for

the argument a. It is only in this last case that we call the function

continuous for the argument a.

A function is called “continuous” (without qualiﬁcation) when it

is continuous for every argument.

Another slightly diﬀerent method of reaching the deﬁnition of

continuity is the following:—

Let us say that a function “ultimately converges into a class α”

if there is some real number such that, for this argument and all

arguments greater than this, the value of the function is a member of

the class α. Similarly we shall say that a function “converges into α

as the argument approaches x from below” if there is some argument

y less than x such that throughout the interval from y (included) to x

(excluded) the function has values which are members of α. We may

now say that a function is continuous for the argument a, for which

it has the value f a, if it satisﬁes four conditions, namely:—

(1) Given any real number less than f a, the function converges

into the successors of this number as the argument approaches a from

below;

(z) Given any real number greater than f a, the function converges

into the predecessors of this number as the argument approaches a

from below;

(j) and (±) Similar conditions for approaches to a from above.

Chap. XI. Limits and Continuity of Functions 8¸

The advantage of this form of deﬁnition is that it analyses the con-

ditions of continuity into four, derived from considering arguments

and values respectively greater or less than the argument and value

for which continuity is to be deﬁned. |

We 11± may now generalise our deﬁnitions so as to apply to series

which are not numerical or known to be numerically measurable.

The case of motion is a convenient one to bear in mind. There is a

story by H. G. Wells which will illustrate, from the case of motion,

the diﬀerence between the limit of a function for a given argument

and its value for the same argument. The hero of the story, who

possessed, without his knowledge, the power of realising his wishes,

was being attacked by a policeman, but on ejaculating “Go to——”

he found that the policeman disappeared. If f (t) was the policeman’s

position at time t, and t

o

the moment of the ejaculation, the limit of

the policeman’s positions as t approached to t

o

from below would be

in contact with the hero, whereas the value for the argument t

o

was

—. But such occurrences are supposed to be rare in the real world, and

it is assumed, though without adequate evidence, that all motions

are continuous, i.e. that, given any body, if f (t) is its position at time

t, f (t) is a continuous function of t. It is the meaning of “continuity”

involved in such statements which we now wish to deﬁne as simply

as possible.

The deﬁnitions given for the case of functions where argument

and value are real numbers can readily be adapted for more general

use.

Let P and Q be two relations, which it is well to imagine serial,

though it is not necessary to our deﬁnitions that they should be so.

Let R be a one-many relation whose domain is contained in the ﬁeld

of P, while its converse domain is contained in the ﬁeld of Q. Then

R is (in a generalised sense) a function, whose arguments belong to

the ﬁeld of Q, while its values belong to the ﬁeld of P. Suppose, for

example, that we are dealing with a particle moving on a line: let

Q be the time-series, P the series of points on our line from left to

right, R the relation of the position of our particle on the line at time

a to the time a, so that “the R of a” is its position at time a. This

illustration may be borne in mind throughout our deﬁnitions.

We shall say that the function R is continuous for the argument

| 11¸ a if, given any interval α on the P-series containing the value of

the function for the argument a, there is an interval on the Q-series

containing a not as an end-point and such that, throughout this

interval, the function has values which are members of α. (We mean

by an “interval” all the terms between any two; i.e. if x and y are two

members of the ﬁeld of P, and x has the relation P to y, we shall mean

by the “P-interval x to y” all terms z such that x has the relation P to z

and z has the relation P to y—together, when so stated, with x or y

themselves.)

Chap. XI. Limits and Continuity of Functions 86

We can easily deﬁne the “ultimate section” and the “ultimate

oscillation.” To deﬁne the “ultimate section” for approaches to the

argument a frombelow, take any argument y which precedes a (i.e. has

the relation Q to a), take the values of the function for all arguments

up to and including y, and form the section of P deﬁned by these

values, i.e. those members of the P-series which are earlier than or

identical with some of these values. Form all such sections for all y’s

that precede a, and take their common part; this will be the ultimate

section. The ultimate upper section and the ultimate oscillation are

then deﬁned exactly as in the previous case.

The adaptation of the deﬁnition of convergence and the resulting

alternative deﬁnition of continuity oﬀers no diﬃculty of any kind.

We say that a function R is “ultimately Q-convergent into α” if

there is a member y of the converse domain of R and the ﬁeld of Q

such that the value of the function for the argument y and for any

argument to which y has the relation Q is a member of α. We say that

R “Q-converges into α as the argument approaches a given argument

a” if there is a term y having the relation Q to a and belonging to the

converse domain of R and such that the value of the function for any

argument in the Q-interval from y (inclusive) to a (exclusive) belongs

to α.

Of the four conditions that a function must fulﬁl in order to be

continuous for the argument a, the ﬁrst is, putting b for the value for

the argument a: |

Given 116 any term having the relation P to b, R Q-converges into the

successors of b (with respect to P) as the argument approaches a from

below.

The second condition is obtained by replacing P by its converse;

the third and fourth are obtained from the ﬁrst and second by replac-

ing Q by its converse.

There is thus nothing, in the notions of the limit of a function or

the continuity of a function, that essentially involves number. Both

can be deﬁned generally, and many propositions about them can

be proved for any two series (one being the argument-series and

the other the value-series). It will be seen that the deﬁnitions do

not involve inﬁnitesimals. They involve inﬁnite classes of intervals,

growing smaller without any limit short of zero, but they do not

involve any intervals that are not ﬁnite. This is analogous to the

fact that if a line an inch long be halved, then halved again, and so

on indeﬁnitely, we never reach inﬁnitesimals in this way: after n

bisections, the length of our bit is 1/z

n

of an inch; and this is ﬁnite

whatever ﬁnite number n may be. The process of successive bisection

does not lead to divisions whose ordinal number is inﬁnite, since it

is essentially a one-by-one process. Thus inﬁnitesimals are not to be

reached in this way. Confusions on such topics have had much to

Chap. XI. Limits and Continuity of Functions 8;

do with the diﬃculties which have been found in the discussion of

inﬁnity and continuity.

CHAPTER XII

SELECTIONS ANDTHE MULTIPLICATIVE

AXIOM

11; In this chapter we have to consider an axiomwhich can be enunciated,

but not proved, in terms of logic, and which is convenient, though not

indispensable, in certain portions of mathematics. It is convenient, in

the sense that many interesting propositions, which it seems natural

to suppose true, cannot be proved without its help; but it is not

indispensable, because even without those propositions the subjects

in which they occur still exist, though in a somewhat mutilated form.

Before enunciating the multiplicative axiom, we must ﬁrst explain

the theory of selections, and the deﬁnition of multiplication when

the number of factors may be inﬁnite.

In deﬁning the arithmetical operations, the only correct procedure

is to construct an actual class (or relation, in the case of relation-

numbers) having the required number of terms. This sometimes

demands a certain amount of ingenuity, but it is essential in order

to prove the existence of the number deﬁned. Take, as the simplest

example, the case of addition. Suppose we are given a cardinal

number µ, and a class α which has µ terms. How shall we deﬁne

µ + µ? For this purpose we must have two classes having µ terms,

and they must not overlap. We can construct such classes from α in

various ways, of which the following is perhaps the simplest: Form

ﬁrst all the ordered couples whose ﬁrst term is a class consisting of a

single member of α, and whose second term is the null-class; then,

secondly, form all the ordered couples whose ﬁrst term is | 118 the null-

class and whose second term is a class consisting of a single member

of α. These two classes of couples have no member in common, and

the logical sum of the two classes will have µ + µ terms. Exactly

analogously we can deﬁne µ +ν, given that µ is the number of some

class α and ν is the number of some class β.

Such deﬁnitions, as a rule, are merely a question of a suitable

technical device. But in the case of multiplication, where the num-

ber of factors may be inﬁnite, important problems arise out of the

deﬁnition.

88

Chap. XII. Selections and the Multiplicative Axiom 8o

Multiplication when the number of factors is ﬁnite oﬀers no diﬃ-

culty. Given two classes α and β, of which the ﬁrst has µ terms and

the second ν terms, we can deﬁne µ × ν as the number of ordered

couples that can be formed by choosing the ﬁrst term out of α and the

second out of β. It will be seen that this deﬁnition does not require

that α and β should not overlap; it even remains adequate when α

and β are identical. For example, let α be the class whose members

are x

1

, x

z

, x

j

. Then the class which is used to deﬁne the product µ×µ

is the class of couples:

(x

1

, x

1

), (x

1

, x

z

), (x

1

, x

j

); (x

z

, x

1

), (x

z

, x

z

), (x

z

, x

j

); (x

j

, x

1

),

(x

j

, x

z

), (x

j

, x

j

).

This deﬁnition remains applicable when µ or ν or both are inﬁnite,

and it can be extended step by step to three or four or any ﬁnite

number of factors. No diﬃculty arises as regards this deﬁnition,

except that it cannot be extended to an inﬁnite number of factors.

The problem of multiplication when the number of factors may

be inﬁnite arises in this way: Suppose we have a class κ consisting

of classes; suppose the number of terms in each of these classes is

given. How shall we deﬁne the product of all these numbers? If we

can frame our deﬁnition generally, it will be applicable whether κ is

ﬁnite or inﬁnite. It is to be observed that the problem is to be able

to deal with the case when κ is inﬁnite, not with the case when its

members are. If | 11o κ is not inﬁnite, the method deﬁned above is just as

applicable when its members are inﬁnite as when they are ﬁnite. It is

the case when κ is inﬁnite, even though its members may be ﬁnite,

that we have to ﬁnd a way of dealing with.

The following method of deﬁning multiplication generally is due

to Dr Whitehead. It is explained and treated at length in Principia

Mathematica, vol. i. ∗8oﬀ., and vol. ii. ∗11±.

Let us suppose to begin with that κ is a class of classes no two

of which overlap—say the constituencies in a country where there

is no plural voting, each constituency being considered as a class

of voters. Let us now set to work to choose one term out of each

class to be its representative, as constituencies do when they elect

members of Parliament, assuming that by law each constituency

has to elect a man who is a voter in that constituency. We thus

arrive at a class of representatives, who make up our Parliament,

one being selected out of each constituency. How many diﬀerent

possible ways of choosing a Parliament are there? Each constituency

can select any one of its voters, and therefore if there are µ voters in

a constituency, it can make µ choices. The choices of the diﬀerent

constituencies are independent; thus it is obvious that, when the total

number of constituencies is ﬁnite, the number of possible Parliaments

is obtained by multiplying together the numbers of voters in the

Chap. XII. Selections and the Multiplicative Axiom oo

various constituencies. When we do not know whether the number of

constituencies is ﬁnite or inﬁnite, we may take the number of possible

Parliaments as deﬁning the product of the numbers of the separate

constituencies. This is the method by which inﬁnite products are

deﬁned. We must now drop our illustration, and proceed to exact

statements.

Let κ be a class of classes, and let us assume to begin with that

no two members of κ overlap, i.e. that if α and β are two diﬀerent

members of κ, then no member of the one is a member of the other.

We shall call a class a “selection” from κ when it consists of just one

term from each member of κ; i.e. µ is a “selection” from κ if every

member of µ belongs to some member | 1zo of κ, and if α be any member

of κ, µ and α have exactly one term in common. The class of all

“selections” from κ we shall call the “multiplicative class” of κ. The

number of terms in the multiplicative class of κ, i.e. the number of

possible selections from κ, is deﬁned as the product of the numbers

of the members of κ. This deﬁnition is equally applicable whether κ

is ﬁnite or inﬁnite.

Before we can be wholly satisﬁed with these deﬁnitions, we must

remove the restriction that no two members of κ are to overlap. For

this purpose, instead of deﬁning ﬁrst a class called a “selection,” we

will deﬁne ﬁrst a relation which we will call a “selector.” A relation R

will be called a “selector” from κ if, from every member of κ, it picks

out one term as the representative of that member, i.e. if, given any

member α of κ, there is just one term x which is a member of α and

has the relation R to α; and this is to be all that R does. The formal

deﬁnition is:

A“selector” froma class of classes κ is a one-many relation, having

κ for its converse domain, and such that, if x has the relation to α,

then x is a member of α.

If R is a selector from κ, and α is a member of κ, and x is the term

which has the relation R to α, we call x the “representative” of α in

respect of the relation R.

A “selection” from κ will now be deﬁned as the domain of a

selector; and the multiplicative class, as before, will be the class of

selections.

But when the members of κ overlap, there may be more selectors

than selections, since a term x which belongs to two classes α and β

may be selected once to represent α and once to represent β, giving

rise to diﬀerent selectors in the two cases, but to the same selection.

For purposes of deﬁning multiplication, it is the selectors we require

rather than the selections. Thus we deﬁne:

“The product of the numbers of the members of a class of classes

κ” is the number of selectors from κ.

We can deﬁne exponentiation by an adaptation of the above |

1z1 plan. We might, of course, deﬁne µ

ν

as the number of selectors from

Chap. XII. Selections and the Multiplicative Axiom o1

ν classes, each of which has µ terms. But there are objections to

this deﬁnition, derived from the fact that the multiplicative axiom

(of which we shall speak shortly) is unnecessarily involved if it is

adopted. We adopt instead the following construction:—

Let α be a class having µ terms, and β a class having ν terms.

Let y be a member of β, and form the class of all ordered couples

that have y for their second term and a member of α for their ﬁrst

term. There will be µ such couples for a given y, since any member

of α may be chosen for the ﬁrst term, and α has µ members. If we

now form all the classes of this sort that result from varying y, we

obtain altogether ν classes, since y may be any member of β, and β

has ν members. These ν classes are each of them a class of couples,

namely, all the couples that can be formed of a variable member of

α and a ﬁxed member of β. We deﬁne µ

ν

as the number of selectors

from the class consisting of these ν classes. Or we may equally well

deﬁne µ

ν

as the number of selections, for, since our classes of couples

are mutually exclusive, the number of selectors is the same as the

number of selections. A selection from our class of classes will be a

set of ordered couples, of which there will be exactly one having any

given member of β for its second term, and the ﬁrst term may be any

member of α. Thus µ

ν

is deﬁned by the selectors from a certain set

of ν classes each having µ terms, but the set is one having a certain

structure and a more manageable composition than is the case in

general. The relevance of this to the multiplicative axiom will appear

shortly.

What applies to exponentiation applies also to the product of two

cardinals. We might deﬁne “µ ×ν” as the sum of the numbers of ν

classes each having µ terms, but we prefer to deﬁne it as the number

of ordered couples to be formed consisting of a member of α followed

by a member of β, where α has µ terms and β has ν terms. This

deﬁnition, also, is designed to evade the necessity of assuming the

multiplicative axiom. |

With 1zz our deﬁnitions, we can prove the usual formal laws of multi-

plication and exponentiation. But there is one thing we cannot prove:

we cannot prove that a product is only zero when one of its factors

is zero. We can prove this when the number of factors is ﬁnite, but

not when it is inﬁnite. In other words, we cannot prove that, given

a class of classes none of which is null, there must be selectors from

them; or that, given a class of mutually exclusive classes, there must

be at least one class consisting of one term out of each of the given

classes. These things cannot be proved; and although, at ﬁrst sight,

they seem obviously true, yet reﬂection brings gradually increasing

doubt, until at last we become content to register the assumption and

its consequences, as we register the axiom of parallels, without as-

suming that we can know whether it is true or false. The assumption,

Chap. XII. Selections and the Multiplicative Axiom oz

loosely worded, is that selectors and selections exist when we should

expect them. There are many equivalent ways of stating it precisely.

We may begin with the following:—

“Given any class of mutually exclusive classes, of which none is

null, there is at least one class which has exactly one term in common

with each of the given classes.”

This proposition we will call the “multiplicative axiom.”

1

We

will ﬁrst give various equivalent forms of the proposition, and then

consider certain ways in which its truth or falsehood is of interest to

mathematics.

The multiplicative axiom is equivalent to the proposition that a

product is only zero when at least one of its factors is zero; i.e. that, if

any number of cardinal numbers be multiplied together, the result

cannot be o unless one of the numbers concerned is o.

The multiplicative axiom is equivalent to the proposition that, if R

be any relation, and κ any class contained in the converse domain of

R, then there is at least one one-many relation implying R and having

κ for its converse domain.

The multiplicative axiom is equivalent to the assumption that if α

be any class, and κ all the sub-classes of α with the exception | 1zj of the

null-class, then there is at least one selector fromκ. This is the formin

which the axiom was ﬁrst brought to the notice of the learned world

by Zermelo, in his “Beweis, dass jede Menge wohlgeordnet werden

kann.”

z

Zermelo regards the axiom as an unquestionable truth. It

must be confessed that, until he made it explicit, mathematicians

had used it without a qualm; but it would seem that they had done

so unconsciously. And the credit due to Zermelo for having made it

explicit is entirely independent of the question whether it is true or

false.

The multiplicative axiom has been shown by Zermelo, in the

above-mentioned proof, to be equivalent to the proposition that every

class can be well-ordered, i.e. can be arranged in a series in which

every sub-class has a ﬁrst term (except, of course, the null-class). The

full proof of this proposition is diﬃcult, but it is not diﬃcult to see

the general principle upon which it proceeds. It uses the form which

we call “Zermelo’s axiom,” i.e. it assumes that, given any class α, there

is at least one one-many relation R whose converse domain consists

of all existent sub-classes of α and which is such that, if x has the

relation R to ξ, then x is a member of ξ. Such a relation picks out a

“representative” from each sub-class; of course, it will often happen

that two sub-classes have the same representative. What Zermelo

does, in eﬀect, is to count oﬀ the members of α, one by one, by means

1

See Principia Mathematica, vol. i. ∗88. Also vol. iii. ∗z¸;–z¸8.

z

Mathematische Annalen, vol. lix. pp. ¸1±–6. In this form we shall speak of it as

Zermelo’s axiom.

Chap. XII. Selections and the Multiplicative Axiom oj

of R and transﬁnite induction. We put ﬁrst the representative of

α; call it x

1

. Then take the representative of the class consisting of

all of α except x

1

; call it x

z

. It must be diﬀerent from x

1

, because

every representative is a member of its class, and x

1

is shut out

from this class. Proceed similarly to take away x

z

, and let x

j

be the

representative of what is left. In this way we ﬁrst obtain a progression

x

1

, x

z

, . . . x

n

, . . ., assuming that α is not ﬁnite. We then take away

the whole progression; let x

ω

be the representative of what is left

of α. In this way we can go on until nothing is left. The successive

representatives will form a | 1z± well-ordered series containing all the

members of α. (The above is, of course, only a hint of the general

lines of the proof.) This proposition is called “Zermelo’s theorem.”

The multiplicative axiom is also equivalent to the assumption that

of any two cardinals which are not equal, one must be the greater.

If the axiom is false, there will be cardinals µ and ν such that µ is

neither less than, equal to, nor greater than ν. We have seen that ℵ

1

and z

ℵ

o

possibly form an instance of such a pair.

Many other forms of the axiom might be given, but the above are

the most important of the forms known at present. As to the truth

or falsehood of the axiom in any of its forms, nothing is known at

present.

The propositions that depend upon the axiom, without being

known to be equivalent to it, are numerous and important. Take ﬁrst

the connection of addition and multiplication. We naturally think

that the sum of ν mutually exclusive classes, each having µ terms,

must have µ×ν terms. When ν is ﬁnite, this can be proved. But when

ν is inﬁnite, it cannot be proved without the multiplicative axiom,

except where, owing to some special circumstance, the existence of

certain selectors can be proved. The way the multiplicative axiom

enters in is as follows: Suppose we have two sets of ν mutually

exclusive classes, each having µ terms, and we wish to prove that the

sum of one set has as many terms as the sum of the other. In order to

prove this, we must establish a one-one relation. Now, since there are

in each case ν classes, there is some one-one relation between the two

sets of classes; but what we want is a one-one relation between their

terms. Let us consider some one-one relation S between the classes.

Then if κ and λ are the two sets of classes, and α is some member of

κ, there will be a member β of λ which will be the correlate of α with

respect to S. Nowα and β each have µ terms, and are therefore similar.

There are, accordingly, one-one correlations of α and β. The trouble

is that there are so many. In order to obtain a one-one correlation of

the sum of κ with the sum of λ, we have to pick out one correlator of

α with β, and similarly for every other pair. This requires a selection

from a set of classes | 1z¸ of correlators, one class of the set being all the

one-one correlators of α with β. If κ and λ are inﬁnite, we cannot in

Chap. XII. Selections and the Multiplicative Axiom o±

general know that such a selection exists, unless we can know that

the multiplicative axiom is true. Hence we cannot establish the usual

kind of connection between addition and multiplication.

This fact has various curious consequences. To begin with, we

know that ℵ

o

z

= ℵ

o

×ℵ

o

= ℵ

o

. It is commonly inferred from this that

the sum of ℵ

o

classes each having ℵ

o

members must itself have ℵ

o

members, but this inference is fallacious, since we do not know that

the number of terms in such a sum is ℵ

o

×ℵ

o

, nor consequently that it

is ℵ

o

. This has a bearing upon the theory of transﬁnite ordinals. It is

easy to prove that an ordinal which has ℵ

o

predecessors must be one

of what Cantor calls the “second class,” i.e. such that a series having

this ordinal number will have ℵ

o

terms in its ﬁeld. It is also easy to

see that, if we take any progression of ordinals of the second class,

the predecessors of their limit form at most the sum of ℵ

o

classes

each having ℵ

o

terms. It is inferred thence—fallaciously, unless the

multiplicative axiom is true—that the predecessors of the limit are

ℵ

o

in number, and therefore that the limit is a number of the “second

class.” That is to say, it is supposed to be proved that any progression

of ordinals of the second class has a limit which is again an ordinal

of the second class. This proposition, with the corollary that ω

1

(the

smallest ordinal of the third class) is not the limit of any progression,

is involved in most of the recognised theory of ordinals of the second

class. In viewof the way in which the multiplicative axiomis involved,

the proposition and its corollary cannot be regarded as proved. They

may be true, or they may not. All that can be said at present is that

we do not know. Thus the greater part of the theory of ordinals of the

second class must be regarded as unproved.

Another illustration may help to make the point clearer. We

know that z×ℵ

o

= ℵ

o

. Hence we might suppose that the sum of ℵ

o

pairs must have ℵ

o

terms. But this, though we can prove that it is

sometimes the case, cannot be proved to happen always | 1z6 unless we

assume the multiplicative axiom. This is illustrated by the millionaire

who bought a pair of socks whenever he bought a pair of boots, and

never at any other time, and who had such a passion for buying

both that at last he had ℵ

o

pairs of boots and ℵ

o

pairs of socks. The

problem is: How many boots had he, and how many socks? One

would naturally suppose that he had twice as many boots and twice

as many socks as he had pairs of each, and that therefore he had ℵ

o

of each, since that number is not increased by doubling. But this is

an instance of the diﬃculty, already noted, of connecting the sum of

ν classes each having µ terms with µ×ν. Sometimes this can be done,

sometimes it cannot. In our case it can be done with the boots, but

not with the socks, except by some very artiﬁcial device. The reason

for the diﬀerence is this: Among boots we can distinguish right and

left, and therefore we can make a selection of one out of each pair,

Chap. XII. Selections and the Multiplicative Axiom o¸

namely, we can choose all the right boots or all the left boots; but with

socks no such principle of selection suggests itself, and we cannot be

sure, unless we assume the multiplicative axiom, that there is any

class consisting of one sock out of each pair. Hence the problem.

We may put the matter in another way. To prove that a class has

ℵ

o

terms, it is necessary and suﬃcient to ﬁnd some way of arranging

its terms in a progression. There is no diﬃculty in doing this with the

boots. The pairs are given as forming an ℵ

o

, and therefore as the ﬁeld

of a progression. Within each pair, take the left boot ﬁrst and the

right second, keeping the order of the pair unchanged; in this way we

obtain a progression of all the boots. But with the socks we shall have

to choose arbitrarily, with each pair, which to put ﬁrst; and an inﬁnite

number of arbitrary choices is an impossibility. Unless we can ﬁnd a

rule for selecting, i.e. a relation which is a selector, we do not know

that a selection is even theoretically possible. Of course, in the case

of objects in space, like socks, we always can ﬁnd some principle of

selection. For example, take the centres of mass of the socks: there

will be points p in space such that, with any | 1z; pair, the centres of mass

of the two socks are not both at exactly the same distance from p;

thus we can choose, from each pair, that sock which has its centre of

mass nearer to p. But there is no theoretical reason why a method of

selection such as this should always be possible, and the case of the

socks, with a little goodwill on the part of the reader, may serve to

show how a selection might be impossible.

It is to be observed that, if it were impossible to select one out

of each pair of socks, it would follow that the socks could not be

arranged in a progression, and therefore that there were not ℵ

o

of

them. This case illustrates that, if µ is an inﬁnite number, one set of µ

pairs may not contain the same number of terms as another set of µ

pairs; for, given ℵ

o

pairs of boots, there are certainly ℵ

o

boots, but we

cannot be sure of this in the case of the socks unless we assume the

multiplicative axiom or fall back upon some fortuitous geometrical

method of selection such as the above.

Another important problem involving the multiplicative axiom is

the relation of reﬂexiveness to non-inductiveness. It will be remem-

bered that in Chapter VIII. we pointed out that a reﬂexive number

must be non-inductive, but that the converse (so far as is known at

present) can only be proved if we assume the multiplicative axiom.

The way in which this comes about is as follows:—

It is easy to prove that a reﬂexive class is one which contains

sub-classes having ℵ

o

terms. (The class may, of course, itself have

ℵ

o

terms.) Thus we have to prove, if we can, that, given any non-

inductive class, it is possible to choose a progression out of its terms.

Now there is no diﬃculty in showing that a non-inductive class must

contain more terms than any inductive class, or, what comes to the

Chap. XII. Selections and the Multiplicative Axiom o6

same thing, that if α is a non-inductive class and ν is any inductive

number, there are sub-classes of α that have ν terms. Thus we can

form sets of ﬁnite sub-classes of α: First one class having no terms,

then classes having 1 term (as many as there are members of α), then

classes having | 1z8 z terms, and so on. We thus get a progression of sets

of sub-classes, each set consisting of all those that have a certain given

ﬁnite number of terms. So far we have not used the multiplicative

axiom, but we have only proved that the number of collections of

sub-classes of α is a reﬂexive number, i.e. that, if µ is the number of

members of α, so that z

µ

is the number of sub-classes of α and z

z

µ

is the number of collections of sub-classes, then, provided µ is not

inductive, z

z

µ

must be reﬂexive. But this is a long way from what we

set out to prove.

In order to advance beyond this point, we must employ the multi-

plicative axiom. From each set of sub-classes let us choose out one,

omitting the sub-class consisting of the null-class alone. That is to say,

we select one sub-class containing one term, α

1

, say; one containing

two terms, α

z

, say; one containing three, α

j

, say; and so on. (We can

do this if the multiplicative axiom is assumed; otherwise, we do not

knowwhether we can always do it or not.) We have nowa progression

α

1

, α

z

, α

j

, . . . of sub-classes of α, instead of a progression of collec-

tions of sub-classes; thus we are one step nearer to our goal. We now

know that, assuming the multiplicative axiom, if µ is a non-inductive

number, z

µ

must be a reﬂexive number.

The next step is to notice that, although we cannot be sure that

new members of α come in at any one speciﬁed stage in the progres-

sion α

1

, α

z

, α

j

, . . . we can be sure that new members keep on coming

in from time to time. Let us illustrate. The class α

1

, which consists

of one term, is a new beginning; let the one term be x

1

. The class

α

z

, consisting of two terms, may or may not contain x

1

; if it does, it

introduces one new term; and if it does not, it must introduce two

new terms, say x

z

, x

j

. In this case it is possible that α

j

consists of

x

1

, x

z

, x

j

, and so introduces no new terms, but in that case α

±

must

introduce a new term. The ﬁrst ν classes α

1

, α

z

, α

j

, . . . α

ν

contain,

at the very most, 1 +z +j +. . . +ν terms, i.e. ν(ν +1)/z terms; thus it

would be possible, if there were no repetitions in the ﬁrst ν classes, to

go on with only repetitions from the (ν +1)

th

| 1zo class to the ν(ν +1)/z

th

class. But by that time the old terms would no longer be suﬃciently

numerous to form a next class with the right number of members, i.e.

ν(ν +1)/z +1, therefore new terms must come in at this point if not

sooner. It follows that, if we omit from our progression α

1

, α

z

, α

j

, . . .

all those classes that are composed entirely of members that have

occurred in previous classes, we shall still have a progression. Let

our new progression be called β

1

, β

z

, β

j

. . . (We shall have α

1

= β

1

and α

z

= β

z

, because α

1

and α

z

must introduce new terms. We may

Chap. XII. Selections and the Multiplicative Axiom o;

or may not have α

j

= β

j

, but, speaking generally, β

µ

will be α

ν

, where

ν is some number greater than µ; i.e. the β’s are some of the α’s.) Now

these β’s are such that any one of them, say β

µ

, contains members

which have not occurred in any of the previous β’s. Let γ

µ

be the part

of β

µ

which consists of new members. Thus we get a new progression

γ

1

, γ

z

, γ

j

, . . . (Again γ

1

will be identical with β

1

and with α

1

; if α

z

does not contain the one member of α

1

, we shall have γ

z

= β

z

= α

z

,

but if α

z

does contain this one member, γ

z

will consist of the other

member of α

z

.) This new progression of γ’s consists of mutually

exclusive classes. Hence a selection from them will be a progression;

i.e. if x

1

is the member of γ

1

, x

z

is a member of γ

z

, x

j

is a member of

γ

j

, and so on; then x

1

, x

z

, x

j

, . . . is a progression, and is a sub-class of

α. Assuming the multiplicative axiom, such a selection can be made.

Thus by twice using this axiom we can prove that, if the axiom is

true, every non-inductive cardinal must be reﬂexive. This could also

be deduced from Zermelo’s theorem, that, if the axiom is true, every

class can be well-ordered; for a well-ordered series must have either

a ﬁnite or a reﬂexive number of terms in its ﬁeld.

There is one advantage in the above direct argument, as against

deduction from Zermelo’s theorem, that the above argument does

not demand the universal truth of the multiplicative axiom, but only

its truth as applied to a set of ℵ

o

classes. It may happen that the

axiom holds for ℵ

o

classes, though not for larger numbers of classes.

For this reason it is better, when | 1jo it is possible, to content ourselves

with the more restricted assumption. The assumption made in the

above direct argument is that a product of ℵ

o

factors is never zero

unless one of the factors is zero. We may state this assumption in

the form: “ℵ

o

is a multipliable number,” where a number ν is deﬁned

as “multipliable” when a product of ν factors is never zero unless

one of the factors is zero. We can prove that a ﬁnite number is always

multipliable, but we cannot prove that any inﬁnite number is so. The

multiplicative axiom is equivalent to the assumption that all cardinal

numbers are multipliable. But in order to identify the reﬂexive with

the non-inductive, or to deal with the problem of the boots and socks,

or to show that any progression of numbers of the second class is of

the second class, we only need the very much smaller assumption

that ℵ

o

is multipliable.

It is not improbable that there is much to be discovered in regard

to the topics discussed in the present chapter. Cases may be found

where propositions which seem to involve the multiplicative axiom

can be proved without it. It is conceivable that the multiplicative

axiom in its general form may be shown to be false. From this point

of view, Zermelo’s theorem oﬀers the best hope: the continuum or

some still more dense series might be proved to be incapable of hav-

ing its terms well-ordered, which would prove the multiplicative

Chap. XII. Selections and the Multiplicative Axiom o8

axiom false, in virtue of Zermelo’s theorem. But so far, no method of

obtaining such results has been discovered, and the subject remains

wrapped in obscurity.

CHAPTER XIII

THE AXIOMOF INFINITY ANDLOGICAL

TYPES

1j1 The axiom of inﬁnity is an assumption which may be enunciated as

follows:—

“If n be any inductive cardinal number, there is at least one class

of individuals having n terms.”

If this is true, it follows, of course, that there are many classes of

individuals having n terms, and that the total number of individuals

in the world is not an inductive number. For, by the axiom, there

is at least one class having n + 1 terms, from which it follows that

there are many classes of n terms and that n is not the number of

individuals in the world. Since n is any inductive number, it follows

that the number of individuals in the world must (if our axiom be

true) exceed any inductive number. In view of what we found in the

preceding chapter, about the possibility of cardinals which are neither

inductive nor reﬂexive, we cannot infer from our axiom that there are

at least ℵ

o

individuals, unless we assume the multiplicative axiom.

But we do know that there are at least ℵ

o

classes of classes, since the

inductive cardinals are classes of classes, and form a progression if

our axiom is true.

The way in which the need for this axiom arises may be explained

as follows. One of Peano’s assumptions is that no two inductive

cardinals have the same successor, i.e. that we shall not have m+1 =

n + 1 unless m = n, if m and n are inductive cardinals. In Chapter

VIII. we had occasion to use what is virtually the same as the above

assumption of Peano’s, namely, that, if n is an inductive cardinal, | 1jz n

is not equal to n +1. It might be thought that this could be proved.

We can prove that, if α is an inductive class, and n is the number of

members of α, then n is not equal to n +1. This proposition is easily

proved by induction, and might be thought to imply the other. But in

fact it does not, since there might be no such class as α. What it does

imply is this: If n is an inductive cardinal such that there is at least

one class having n members, then n is not equal to n +1. The axiom

of inﬁnity assures us (whether truly or falsely) that there are classes

oo

Chap. XIII. The Axiom of Inﬁnity and Logical Types 1oo

having n members, and thus enables us to assert that n is not equal

to n +1. But without this axiom we should be left with the possibility

that n and n +1 might both be the null-class.

Let us illustrate this possibility by an example: Suppose there

were exactly nine individuals in the world. (As to what is meant by

the word “individual,” I must ask the reader to be patient.) Then the

inductive cardinals from o up to o would be such as we expect, but

1o (deﬁned as o +1) would be the null-class. It will be remembered

that n+1 may be deﬁned as follows: n+1 is the collection of all those

classes which have a term x such that, when x is taken away, there

remains a class of n terms. Now applying this deﬁnition, we see that,

in the case supposed, o +1 is a class consisting of no classes, i.e. it is

the null-class. The same will be true of o +z, or generally of o +n,

unless n is zero. Thus 1o and all subsequent inductive cardinals will

all be identical, since they will all be the null-class. In such a case

the inductive cardinals will not form a progression, nor will it be

true that no two have the same successor, for o and 1o will both be

succeeded by the null-class (1o being itself the null-class). It is in

order to prevent such arithmetical catastrophes that we require the

axiom of inﬁnity.

As a matter of fact, so long as we are content with the arithmetic of

ﬁnite integers, and do not introduce either inﬁnite integers or inﬁnite

classes or series of ﬁnite integers or ratios, it is possible to obtain all

desired results without the axiom of inﬁnity. That is to say, we can

deal with the addition, | 1jj multiplication, and exponentiation of ﬁnite

integers and of ratios, but we cannot deal with inﬁnite integers or

with irrationals. Thus the theory of the transﬁnite and the theory of

real numbers fails us. How these various results come about must

now be explained.

Assuming that the number of individuals in the world is n, the

number of classes of individuals will be z

n

. This is in virtue of the

general proposition mentioned in Chapter VIII. that the number of

classes contained in a class which has n members is z

n

. Now z

n

is

always greater than n. Hence the number of classes in the world is

greater than the number of individuals. If, now, we suppose the num-

ber of individuals to be o, as we did just now, the number of classes

will be z

o

, i.e. ¸1z. Thus if we take our numbers as being applied to

the counting of classes instead of to the counting of individuals, our

arithmetic will be normal until we reach ¸1z: the ﬁrst number to be

null will be ¸1j. And if we advance to classes of classes we shall do

still better: the number of them will be z

¸1z

, a number which is so

large as to stagger imagination, since it has about 1¸j digits. And if

we advance to classes of classes of classes, we shall obtain a number

represented by z raised to a power which has about 1¸j digits; the

number of digits in this number will be about three times 1o

1¸z

. In a

Chap. XIII. The Axiom of Inﬁnity and Logical Types 1o1

time of paper shortage it is undesirable to write out this number, and

if we want larger ones we can obtain them by travelling further along

the logical hierarchy. In this way any assigned inductive cardinal can

be made to ﬁnd its place among numbers which are not null, merely

by travelling along the hierarchy for a suﬃcient distance.

1

As regards ratios, we have a very similar state of aﬀairs. If a ratio

µ/ν is to have the expected properties, there must be enough objects

of whatever sort is being counted to insure that the null-class does

not suddenly obtrude itself. But this can be insured, for any given

ratio µ/ν, without the axiom of | 1j± inﬁnity, by merely travelling up the

hierarchy a suﬃcient distance. If we cannot succeed by counting

individuals, we can try counting classes of individuals; if we still

do not succeed, we can try classes of classes, and so on. Ultimately,

however few individuals there may be in the world, we shall reach a

stage where there are many more than µ objects, whatever inductive

number µ may be. Even if there were no individuals at all, this would

still be true, for there would then be one class, namely, the null-class,

z classes of classes (namely, the null-class of classes and the class

whose only member is the null-class of individuals), ± classes of

classes of classes, 16 at the next stage, 6¸,¸j6 at the next stage, and

so on. Thus no such assumption as the axiom of inﬁnity is required

in order to reach any given ratio or any given inductive cardinal.

It is when we wish to deal with the whole class or series of induc-

tive cardinals or of ratios that the axiom is required. We need the

whole class of inductive cardinals in order to establish the existence

of ℵ

o

, and the whole series in order to establish the existence of pro-

gressions: for these results, it is necessary that we should be able to

make a single class or series in which no inductive cardinal is null.

We need the whole series of ratios in order of magnitude in order

to deﬁne real numbers as segments: this deﬁnition will not give the

desired result unless the series of ratios is compact, which it cannot

be if the total number of ratios, at the stage concerned, is ﬁnite.

It would be natural to suppose—as I supposed myself in former

days—that, by means of constructions such as we have been con-

sidering, the axiom of inﬁnity could be proved. It may be said: Let

us assume that the number of individuals is n, where n may be o

without spoiling our argument; then if we form the complete set of

individuals, classes, classes of classes, etc., all taken together, the

number of terms in our whole set will be

n +z

n

+z

z

n

. . . ad inf.,

which is ℵ

o

. Thus taking all kinds of objects together, and not |

1j¸ conﬁning ourselves to objects of any one type, we shall certainly

1

On this subject see Principia Mathematica, vol. ii. ∗1zoﬀ. On the corresponding

problems as regards ratio, see ibid., vol. iii. ∗jojﬀ.

Chap. XIII. The Axiom of Inﬁnity and Logical Types 1oz

obtain an inﬁnite class, and shall therefore not need the axiom of

inﬁnity. So it might be said.

Now, before going into this argument, the ﬁrst thing to observe

is that there is an air of hocus-pocus about it: something reminds

one of the conjurer who brings things out of the hat. The man who

has lent his hat is quite sure there wasn’t a live rabbit in it before,

but he is at a loss to say how the rabbit got there. So the reader, if he

has a robust sense of reality, will feel convinced that it is impossible

to manufacture an inﬁnite collection out of a ﬁnite collection of in-

dividuals, though he may be unable to say where the ﬂaw is in the

above construction. It would be a mistake to lay too much stress on

such feelings of hocus-pocus; like other emotions, they may easily

lead us astray. But they aﬀord a prima facie ground for scrutinising

very closely any argument which arouses them. And when the above

argument is scrutinised it will, in my opinion, be found to be fal-

lacious, though the fallacy is a subtle one and by no means easy to

avoid consistently.

The fallacy involved is the fallacy which may be called “confusion

of types.” To explain the subject of “types” fully would require a

whole volume; moreover, it is the purpose of this book to avoid

those parts of the subjects which are still obscure and controversial,

isolating, for the convenience of beginners, those parts which can

be accepted as embodying mathematically ascertained truths. Now

the theory of types emphatically does not belong to the ﬁnished

and certain part of our subject: much of this theory is still inchoate,

confused, and obscure. But the need of some doctrine of types is

less doubtful than the precise form the doctrine should take; and in

connection with the axiom of inﬁnity it is particularly easy to see the

necessity of some such doctrine.

This necessity results, for example, from the “contradiction of

the greatest cardinal.” We saw in Chapter VIII. that the number of

classes contained in a given class is always greater than the | 1j6 number

of members of the class, and we inferred that there is no greatest

cardinal number. But if we could, as we suggested a moment ago, add

together into one class the individuals, classes of individuals, classes

of classes of individuals, etc., we should obtain a class of which its

own sub-classes would be members. The class consisting of all objects

that can be counted, of whatever sort, must, if there be such a class,

have a cardinal number which is the greatest possible. Since all its

sub-classes will be members of it, there cannot be more of them than

there are members. Hence we arrive at a contradiction.

When I ﬁrst came upon this contradiction, in the year 1oo1, I

attempted to discover some ﬂaw in Cantor’s proof that there is no

greatest cardinal, which we gave in Chapter VIII. Applying this proof

to the supposed class of all imaginable objects, I was led to a new and

simpler contradiction, namely, the following:—

Chap. XIII. The Axiom of Inﬁnity and Logical Types 1oj

The comprehensive class we are considering, which is to embrace

everything, must embrace itself as one of its members. In other

words, if there is such a thing as “everything,” then “everything” is

something, and is a member of the class “everything.” But normally

a class is not a member of itself. Mankind, for example, is not a man.

Form now the assemblage of all classes which are not members of

themselves. This is a class: is it a member of itself or not? If it is, it is

one of those classes that are not members of themselves, i.e. it is not a

member of itself. If it is not, it is not one of those classes that are not

members of themselves, i.e. it is a member of itself. Thus of the two

hypotheses—that it is, and that it is not, a member of itself—each

implies its contradictory. This is a contradiction.

There is no diﬃculty in manufacturing similar contradictions ad

lib. The solution of such contradictions by the theory of types is set

forth fully in Principia Mathematica,

z

and also, more brieﬂy, in articles

by the present author in the American Journal | 1j; of Mathematics

j

and in

the Revue de M´ etaphysique et de Morale.

±

For the present an outline of

the solution must suﬃce.

The fallacy consists in the formation of what we may call “impure”

classes, i.e. classes which are not pure as to “type.” As we shall see

in a later chapter, classes are logical ﬁctions, and a statement which

appears to be about a class will only be signiﬁcant if it is capable

of translation into a form in which no mention is made of the class.

This places a limitation upon the ways in which what are nominally,

though not really, names for classes can occur signiﬁcantly: a sentence

or set of symbols in which such pseudo-names occur in wrong ways

is not false, but strictly devoid of meaning. The supposition that a

class is, or that it is not, a member of itself is meaningless in just this

way. And more generally, to suppose that one class of individuals is a

member, or is not a member, of another class of individuals will be

to suppose nonsense; and to construct symbolically any class whose

members are not all of the same grade in the logical hierarchy is to use

symbols in a way which makes them no longer symbolise anything.

Thus if there are n individuals in the world, and z

n

classes of indi-

viduals, we cannot form a new class, consisting of both individuals

and classes and having n +z

n

members. In this way the attempt to

escape from the need for the axiom of inﬁnity breaks down. I do not

pretend to have explained the doctrine of types, or done more than

indicate, in rough outline, why there is need of such a doctrine. I have

aimed only at saying just so much as was required in order to show

that we cannot prove the existence of inﬁnite numbers and classes by

z

Vol. i., Introduction, chap. ii., ∗1z and ∗zo; vol. ii., Prefatory Statement.

j

“Mathematical Logic as based on the Theory of Types,” vol. xxx., 1oo8, pp.

zzz–z6z.

±

“Les paradoxes de la logique,” 1oo6, pp. 6z;–6¸o.

Chap. XIII. The Axiom of Inﬁnity and Logical Types 1o±

such conjurer’s methods as we have been examining. There remain,

however, certain other possible methods which must be considered.

Various arguments professing to prove the existence of inﬁnite

classes are given in the Principles of Mathematics, §jjo (p. j¸;). | 1j8 In so

far as these arguments assume that, if n is an inductive cardinal, n

is not equal to n +1, they have been already dealt with. There is an

argument, suggested by a passage in Plato’s Parmenides, to the eﬀect

that, if there is such a number as 1, then 1 has being; but 1 is not

identical with being, and therefore 1 and being are two, and therefore

there is such a number as z, and z together with 1 and being gives

a class of three terms, and so on. This argument is fallacious, partly

because “being” is not a term having any deﬁnite meaning, and still

more because, if a deﬁnite meaning were invented for it, it would be

found that numbers do not have being—they are, in fact, what are

called “logical ﬁctions,” as we shall see when we come to consider

the deﬁnition of classes.

The argument that the number of numbers from o to n (both inclu-

sive) is n +1 depends upon the assumption that up to and including

n no number is equal to its successor, which, as we have seen, will

not be always true if the axiom of inﬁnity is false. It must be under-

stood that the equation n = n +1, which might be true for a ﬁnite n

if n exceeded the total number of individuals in the world, is quite

diﬀerent from the same equation as applied to a reﬂexive number.

As applied to a reﬂexive number, it means that, given a class of n

terms, this class is “similar” to that obtained by adding another term.

But as applied to a number which is too great for the actual world, it

merely means that there is no class of n individuals, and no class of

n +1 individuals; it does not mean that, if we mount the hierarchy

of types suﬃciently far to secure the existence of a class of n terms,

we shall then ﬁnd this class “similar” to one of n +1 terms, for if n is

inductive this will not be the case, quite independently of the truth

or falsehood of the axiom of inﬁnity.

There is an argument employed by both Bolzano

¸

and Dedekind

6

to prove the existence of reﬂexive classes. The argument, in brief, is

this: An object is not identical with the idea of the | 1jo object, but there

is (at least in the realm of being) an idea of any object. The relation of

an object to the idea of it is one-one, and ideas are only some among

objects. Hence the relation “idea of” constitutes a reﬂexion of the

whole class of objects into a part of itself, namely, into that part which

consists of ideas. Accordingly, the class of objects and the class of

ideas are both inﬁnite. This argument is interesting, not only on its

own account, but because the mistakes in it (or what I judge to be

¸

Bolzano, Paradoxien des Unendlichen, 1j.

6

Dedekind, Was sind und was sollen die Zahlen? No. 66.

Chap. XIII. The Axiom of Inﬁnity and Logical Types 1o¸

mistakes) are of a kind which it is instructive to note. The main error

consists in assuming that there is an idea of every object. It is, of

course, exceedingly diﬃcult to decide what is meant by an “idea”;

but let us assume that we know. We are then to suppose that, starting

(say) with Socrates, there is the idea of Socrates, and then the idea

of the idea of Socrates, and so on ad inf. Now it is plain that this is

not the case in the sense that all these ideas have actual empirical

existence in people’s minds. Beyond the third or fourth stage they be-

come mythical. If the argument is to be upheld, the “ideas” intended

must be Platonic ideas laid up in heaven, for certainly they are not on

earth. But then it at once becomes doubtful whether there are such

ideas. If we are to know that there are, it must be on the basis of some

logical theory, proving that it is necessary to a thing that there should

be an idea of it. We certainly cannot obtain this result empirically, or

apply it, as Dedekind does, to “meine Gedankenwelt”—the world of

my thoughts.

If we were concerned to examine fully the relation of idea and

object, we should have to enter upon a number of psychological and

logical inquiries, which are not relevant to our main purpose. But

a few further points should be noted. If “idea” is to be understood

logically, it may be identical with the object, or it may stand for a

description (in the sense to be explained in a subsequent chapter). In

the former case the argument fails, because it was essential to the

proof of reﬂexiveness that object and idea should be distinct. In the

second case the argument also fails, because the relation of object

and description is not | 1±o one-one: there are innumerable correct de-

scriptions of any given object. Socrates (e.g.) may be described as

“the master of Plato,” or as “the philosopher who drank the hem-

lock,” or as “the husband of Xantippe.” If—to take up the remaining

hypothesis—“idea” is to be interpreted psychologically, it must be

maintained that there is not any one deﬁnite psychological entity

which could be called the idea of the object: there are innumerable

beliefs and attitudes, each of which could be called an idea of the

object in the sense in which we might say “my idea of Socrates is

quite diﬀerent from yours,” but there is not any central entity (ex-

cept Socrates himself) to bind together various “ideas of Socrates,”

and thus there is not any such one-one relation of idea and object

as the argument supposes. Nor, of course, as we have already noted,

is it true psychologically that there are ideas (in however extended

a sense) of more than a tiny proportion of the things in the world.

For all these reasons, the above argument in favour of the logical

existence of reﬂexive classes must be rejected.

It might be thought that, whatever may be said of logical argu-

ments, the empirical arguments derivable from space and time, the

diversity of colours, etc., are quite suﬃcient to prove the actual exis-

Chap. XIII. The Axiom of Inﬁnity and Logical Types 1o6

tence of an inﬁnite number of particulars. I do not believe this. We

have no reason except prejudice for believing in the inﬁnite extent

of space and time, at any rate in the sense in which space and time

are physical facts, not mathematical ﬁctions. We naturally regard

space and time as continuous, or, at least, as compact; but this again

is mainly prejudice. The theory of “quanta” in physics, whether true

or false, illustrates the fact that physics can never aﬀord proof of

continuity, though it might quite possibly aﬀord disproof. The senses

are not suﬃciently exact to distinguish between continuous motion

and rapid discrete succession, as anyone may discover in a cinema. A

world in which all motion consisted of a series of small ﬁnite jerks

would be empirically indistinguishable from one in which motion

was continuous. It would take up too much space to | 1±1 defend these

theses adequately; for the present I am merely suggesting them for

the reader’s consideration. If they are valid, it follows that there is

no empirical reason for believing the number of particulars in the

world to be inﬁnite, and that there never can be; also that there is at

present no empirical reason to believe the number to be ﬁnite, though

it is theoretically conceivable that some day there might be evidence

pointing, though not conclusively, in that direction.

From the fact that the inﬁnite is not self-contradictory, but is also

not demonstrable logically, we must conclude that nothing can be

known a priori as to whether the number of things in the world is

ﬁnite or inﬁnite. The conclusion is, therefore, to adopt a Leibnizian

phraseology, that some of the possible worlds are ﬁnite, some inﬁnite,

and we have no means of knowing to which of these two kinds our

actual world belongs. The axiom of inﬁnity will be true in some

possible worlds and false in others; whether it is true or false in this

world, we cannot tell.

Throughout this chapter the synonyms “individual” and “partic-

ular” have been used without explanation. It would be impossible

to explain them adequately without a longer disquisition on the the-

ory of types than would be appropriate to the present work, but a

few words before we leave this topic may do something to diminish

the obscurity which would otherwise envelop the meaning of these

words.

In an ordinary statement we can distinguish a verb, expressing

an attribute or relation, from the substantives which express the

subject of the attribute or the terms of the relation. “Cæsar lived”

ascribes an attribute to Cæsar; “Brutus killed Cæsar” expresses a

relation between Brutus and Cæsar. Using the word “subject” in a

generalised sense, we may call both Brutus and Cæsar subjects of

this proposition: the fact that Brutus is grammatically subject and

Cæsar object is logically irrelevant, since the same occurrence may be

expressed in the words “Cæsar was killed by Brutus,” where Cæsar

Chap. XIII. The Axiom of Inﬁnity and Logical Types 1o;

is the grammatical subject. | 1±z Thus in the simpler sort of proposition

we shall have an attribute or relation holding of or between one, two

or more “subjects” in the extended sense. (A relation may have more

than two terms: e.g. “A gives B to C” is a relation of three terms.)

Now it often happens that, on a closer scrutiny, the apparent subjects

are found to be not really subjects, but to be capable of analysis; the

only result of this, however, is that new subjects take their places.

It also happens that the verb may grammatically be made subject:

e.g. we may say, “Killing is a relation which holds between Brutus

and Cæsar.” But in such cases the grammar is misleading, and in

a straightforward statement, following the rules that should guide

philosophical grammar, Brutus and Cæsar will appear as the subjects

and killing as the verb.

We are thus led to the conception of terms which, when they occur

in propositions, can only occur as subjects, and never in any other

way. This is part of the old scholastic deﬁnition of substance; but

persistence through time, which belonged to that notion, forms no

part of the notion with which we are concerned. We shall deﬁne

“proper names” as those terms which can only occur as subjects in

propositions (using “subject” in the extended sense just explained).

We shall further deﬁne “individuals” or “particulars” as the objects

that can be named by proper names. (It would be better to deﬁne

them directly, rather than by means of the kind of symbols by which

they are symbolised; but in order to do that we should have to plunge

deeper into metaphysics than is desirable here.) It is, of course,

possible that there is an endless regress: that whatever appears as a

particular is really, on closer scrutiny, a class or some kind of complex.

If this be the case, the axiom of inﬁnity must of course be true. But if

it be not the case, it must be theoretically possible for analysis to reach

ultimate subjects, and it is these that give the meaning of “particulars”

or “individuals.” It is to the number of these that the axiom of inﬁnity

is assumed to apply. If it is true of them, it is true | 1±j of classes of them,

and classes of classes of them, and so on; similarly if it is false of them,

it is false throughout this hierarchy. Hence it is natural to enunciate

the axiom concerning them rather than concerning any other stage in

the hierarchy. But whether the axiom is true or false, there seems no

known method of discovering.

CHAPTER XIV

INCOMPATIBILITY ANDTHE THEORY OF

DEDUCTION

1±± We have now explored, somewhat hastily it is true, that part of the

philosophy of mathematics which does not demand a critical exam-

ination of the idea of class. In the preceding chapter, however, we

found ourselves confronted by problems which make such an exam-

ination imperative. Before we can undertake it, we must consider

certain other parts of the philosophy of mathematics, which we have

hitherto ignored. In a synthetic treatment, the parts which we shall

now be concerned with come ﬁrst: they are more fundamental than

anything that we have discussed hitherto. Three topics will concern

us before we reach the theory of classes, namely: (1) the theory of

deduction, (z) propositional functions, (j) descriptions. Of these, the

third is not logically presupposed in the theory of classes, but it is a

simpler example of the kind of theory that is needed in dealing with

classes. It is the ﬁrst topic, the theory of deduction, that will concern

us in the present chapter.

Mathematics is a deductive science: starting from certain pre-

misses, it arrives, by a strict process of deduction, at the various

theorems which constitute it. It is true that, in the past, mathematical

deductions were often greatly lacking in rigour; it is true also that

perfect rigour is a scarcely attainable ideal. Nevertheless, in so far as

rigour is lacking in a mathematical proof, the proof is defective; it is

no defence to urge that common sense shows the result to be correct,

for if we were to rely upon that, it would be better to dispense with

argument altogether, | 1±¸ rather than bring fallacy to the rescue of com-

mon sense. No appeal to common sense, or “intuition,” or anything

except strict deductive logic, ought to be needed in mathematics after

the premisses have been laid down.

Kant, having observed that the geometers of his day could not

prove their theorems by unaided argument, but required an appeal

to the ﬁgure, invented a theory of mathematical reasoning according

to which the inference is never strictly logical, but always requires

the support of what is called “intuition.” The whole trend of modern

1o8

Chap. XIV. Incompatibility and the Theory of Deduction 1oo

mathematics, with its increased pursuit of rigour, has been against

this Kantian theory. The things in the mathematics of Kant’s day

which cannot be proved, cannot be known—for example, the axiom

of parallels. What can be known, in mathematics and by mathemat-

ical methods, is what can be deduced from pure logic. What else

is to belong to human knowledge must be ascertained otherwise—

empirically, through the senses or through experience in some form,

but not a priori. The positive grounds for this thesis are to be found in

Principia Mathematica, passim; a controversial defence of it is given in

the Principles of Mathematics. We cannot here do more than refer the

reader to those works, since the subject is too vast for hasty treatment.

Meanwhile, we shall assume that all mathematics is deductive, and

proceed to inquire as to what is involved in deduction.

In deduction, we have one or more propositions called premisses,

from which we infer a proposition called the conclusion. For our

purposes, it will be convenient, when there are originally several

premisses, to amalgamate them into a single proposition, so as to be

able to speak of the premiss as well as of the conclusion. Thus we may

regard deduction as a process by which we pass from knowledge of

a certain proposition, the premiss, to knowledge of a certain other

proposition, the conclusion. But we shall not regard such a process as

logical deduction unless it is correct, i.e. unless there is such a relation

between premiss and conclusion that we have a right to believe the

conclusion | 1±6 if we know the premiss to be true. It is this relation that

is chieﬂy of interest in the logical theory of deduction.

In order to be able validly to infer the truth of a proposition, we

must know that some other proposition is true, and that there is

between the two a relation of the sort called “implication,” i.e. that

(as we say) the premiss “implies” the conclusion. (We shall deﬁne this

relation shortly.) Or we may know that a certain other proposition is

false, and that there is a relation between the two of the sort called

“disjunction,” expressed by “p or q,”

1

so that the knowledge that the

one is false allows us to infer that the other is true. Again, what

we wish to infer may be the falsehood of some proposition, not its

truth. This may be inferred from the truth of another proposition,

provided we know that the two are “incompatible,” i.e. that if one

is true, the other is false. It may also be inferred from the falsehood

of another proposition, in just the same circumstances in which the

truth of the other might have been inferred from the truth of the one;

i.e. from the falsehood of p we may infer the falsehood of q, when q

implies p. All these four are cases of inference. When our minds are

ﬁxed upon inference, it seems natural to take “implication” as the

primitive fundamental relation, since this is the relation which must

1

We shall use the letters p, q, r, s, t to denote variable propositions.

Chap. XIV. Incompatibility and the Theory of Deduction 11o

hold between p and q if we are to be able to infer the truth of q from

the truth of p. But for technical reasons this is not the best primitive

idea to choose. Before proceeding to primitive ideas and deﬁnitions,

let us consider further the various functions of propositions suggested

by the above-mentioned relations of propositions.

The simplest of such functions is the negative, “not-p.” This is

that function of p which is true when p is false, and false when p is

true. It is convenient to speak of the truth of a proposition, or its

falsehood, as its “truth-value”

z

; i.e. truth is the “truth-value” of a true

proposition, and falsehood of a false one. Thus not-p has the opposite

truth-value to p. |

We 1±; may take next disjunction, “p or q.” This is a function whose

truth-value is truth when p is true and also when q is true, but is

falsehood when both p and q are false.

Next we may take conjunction, “p and q.” This has truth for its

truth-value when p and q are both true; otherwise it has falsehood for

its truth-value.

Take next incompatibility, i.e. “p and q are not both true.” This is

the negation of conjunction; it is also the disjunction of the negations

of p and q, i.e. it is “not-p or not-q.” Its truth-value is truth when p is

false and likewise when q is false; its truth-value is falsehood when p

and q are both true.

Last take implication, i.e. “p implies q,” or “if p, then q.” This is

to be understood in the widest sense that will allow us to infer the

truth of q if we know the truth of p. Thus we interpret it as meaning:

“Unless p is false, q is true,” or “either p is false or q is true.” (The fact

that “implies” is capable of other meanings does not concern us; this

is the meaning which is convenient for us.) That is to say, “p implies

q” is to mean “not-p or q”: its truth-value is to be truth if p is false,

likewise if q is true, and is to be falsehood if p is true and q is false.

We have thus ﬁve functions: negation, disjunction, conjunction,

incompatibility, and implication. We might have added others, for

example, joint falsehood, “not-p and not-q,” but the above ﬁve will

suﬃce. Negation diﬀers from the other four in being a function of

one proposition, whereas the others are functions of two. But all ﬁve

agree in this, that their truth-value depends only upon that of the

propositions which are their arguments. Given the truth or falsehood

of p, or of p and q (as the case may be), we are given the truth or

falsehood of the negation, disjunction, conjunction, incompatibility,

or implication. A function of propositions which has this property is

called a “truth-function.”

The whole meaning of a truth-function is exhausted by the state-

ment of the circumstances under which it is true or false. “Not-p,”

z

This term is due to Frege.

Chap. XIV. Incompatibility and the Theory of Deduction 111

for example, is simply that function of p which is true when p is false,

and false when p is true: there is no further | 1±8 meaning to be assigned

to it. The same applies to “p or q” and the rest. It follows that two

truth-functions which have the same truth-value for all values of

the argument are indistinguishable. For example, “p and q” is the

negation of “not-p or not-q” and vice versa; thus either of these may

be deﬁned as the negation of the other. There is no further meaning

in a truth-function over and above the conditions under which it is

true or false.

It is clear that the above ﬁve truth-functions are not all indepen-

dent. We can deﬁne some of themin terms of others. There is no great

diﬃculty in reducing the number to two; the two chosen in Principia

Mathematica are negation and disjunction. Implication is then de-

ﬁned as “not-p or q”; incompatibility as “not-p or not-q”; conjunction

as the negation of incompatibility. But it has been shown by Shef-

fer

j

that we can be content with one primitive idea for all ﬁve, and

by Nicod

±

that this enables us to reduce the primitive propositions

required in the theory of deduction to two non-formal principles and

one formal one. For this purpose, we may take as our one indeﬁnable

either incompatibility or joint falsehood. We will choose the former.

Our primitive idea, now, is a certain truth-function called “incom-

patibility,” which we will denote by p / q. Negation can be at once

deﬁned as the incompatibility of a proposition with itself, i.e. “not-p”

is deﬁned as “p / p.” Disjunction is the incompatibility of not-p and

not-q, i.e. it is (p / p) | (q / q). Implication is the incompatibility of p and

not-q, i.e. p | (q / q). Conjunction is the negation of incompatibility, i.e.

it is (p / q) | (p / q). Thus all our four other functions are deﬁned in

terms of incompatibility.

It is obvious that there is no limit to the manufacture of truth-

functions, either by introducing more arguments or by repeating

arguments. What we are concerned with is the connection of this

subject with inference. |

If 1±o we know that p is true and that p implies q, we can proceed

to assert q. There is always unavoidably something psychological

about inference: inference is a method by which we arrive at new

knowledge, and what is not psychological about it is the relation

which allows us to infer correctly; but the actual passage from the

assertion of p to the assertion of q is a psychological process, and we

must not seek to represent it in purely logical terms.

In mathematical practice, when we infer, we have always some

expression containing variable propositions, say p and q, which is

known, in virtue of its form, to be true for all values of p and q;

we have also some other expression, part of the former, which is

j

Trans. Am. Math. Soc., vol. xiv. pp. ±81–±88.

±

Proc. Camb. Phil. Soc., vol. xix., i., January 1o1;.

Chap. XIV. Incompatibility and the Theory of Deduction 11z

also known to be true for all values of p and q; and in virtue of the

principles of inference, we are able to drop this part of our original

expression, and assert what is left. This somewhat abstract account

may be made clearer by a few examples.

Let us assume that we know the ﬁve formal principles of deduc-

tion enumerated in Principia Mathematica. (M. Nicod has reduced

these to one, but as it is a complicated proposition, we will begin with

the ﬁve.) These ﬁve propositions are as follows:—

(1) “p or p” implies p—i.e. if either p is true or p is true, then p is

true.

(z) q implies “p or q”—i.e. the disjunction “p or q” is true when

one of its alternatives is true.

(j) “p or q” implies “q or p.” This would not be required if we

had a theoretically more perfect notation, since in the conception

of disjunction there is no order involved, so that “p or q” and “q or

p” should be identical. But since our symbols, in any convenient

form, inevitably introduce an order, we need suitable assumptions

for showing that the order is irrelevant.

(±) If either p is true or “q or r” is true, then either q is true or

“p or r” is true. (The twist in this proposition serves to increase its

deductive power.) |

(¸) 1¸o If q implies r, then “p or q” implies “p or r.”

These are the formal principles of deduction employed in Principia

Mathematica. A formal principle of deduction has a double use, and

it is in order to make this clear that we have cited the above ﬁve

propositions. It has a use as the premiss of an inference, and a use as

establishing the fact that the premiss implies the conclusion. In the

schema of an inference we have a proposition p, and a proposition “p

implies q,” from which we infer q. Now when we are concerned with

the principles of deduction, our apparatus of primitive propositions

has to yield both the p and the “p implies q” of our inferences. That is

to say, our rules of deduction are to be used, not only as rules, which

is their use for establishing “p implies q,” but also as substantive

premisses, i.e. as the p of our schema. Suppose, for example, we

wish to prove that if p implies q, then if q implies r it follows that p

implies r. We have here a relation of three propositions which state

implications. Put

p

1

= p implies q, p

z

= q implies r, p

j

= p implies r.

Then we have to prove that p

1

implies that p

z

implies p

j

. Now take

the ﬁfth of our above principles, substitute not-p for p, and remember

that “not-p or q” is by deﬁnition the same as “p implies q.” Thus our

ﬁfth principle yields:

“If q implies r, then ‘p implies q’ implies ‘p implies r,’” i.e. “p

z

implies that p

1

implies p

j

.” Call this proposition A.

Chap. XIV. Incompatibility and the Theory of Deduction 11j

But the fourth of our principles, when we substitute not-p, not-q, for

p and q, and remember the deﬁnition of implication, becomes:

“If p implies that q implies r, then q implies that p implies r.”

Writing p

z

in place of p, p

1

in place of q, and p

j

in place of r, this

becomes:

“If p

z

implies that p

1

implies p

j

, then p

1

implies that p

z

implies

p

j

.” Call this B. |

1¸1 Now we proved by means of our ﬁfth principle that

“p

z

implies that p

1

implies p

j

,” which was what we called A.

Thus we have here an instance of the schema of inference, since A

represents the p of our scheme, and B represents the “p implies q.”

Hence we arrive at q, namely,

“p

1

implies that p

z

implies p

j

,”

which was the proposition to be proved. In this proof, the adaptation

of our ﬁfth principle, which yields A, occurs as a substantive premiss;

while the adaptation of our fourth principle, which yields B, is used to

give the form of the inference. The formal and material employments

of premisses in the theory of deduction are closely intertwined, and

it is not very important to keep them separated, provided we realise

that they are in theory distinct.

The earliest method of arriving at new results from a premiss is

one which is illustrated in the above deduction, but which itself can

hardly be called deduction. The primitive propositions, whatever

they may be, are to be regarded as asserted for all possible values

of the variable propositions p, q, r which occur in them. We may

therefore substitute for (say) p any expression whose value is always

a proposition, e.g. not-p, “s implies t,” and so on. By means of such

substitutions we really obtain sets of special cases of our original

proposition, but from a practical point of view we obtain what are vir-

tually new propositions. The legitimacy of substitutions of this kind

has to be insured by means of a non-formal principle of inference.

¸

We may now state the one formal principle of inference to which

M. Nicod has reduced the ﬁve given above. For this purpose we will

ﬁrst show how certain truth-functions can be deﬁned in terms of

incompatibility. We saw already that

p | (q / q) means “p implies q.” |

¸

No such principle is enunciated in Principia Mathematica or in M. Nicod’s

article mentioned above. But this would seem to be an omission.

Chap. XIV. Incompatibility and the Theory of Deduction 11±

1¸z We now observe that

p | (q / r) means “p implies both q and r.”

For this expression means “p is incompatible with the incompatibility

of q and r,” i.e. “p implies that q and r are not incompatible,” i.e. “p

implies that q and r are both true”—for, as we saw, the conjunction of

q and r is the negation of their incompatibility.

Observe next that t | (t / t) means “t implies itself.” This is a

particular case of p | (q / q).

Let us write p for the negation of p; thus p / s will mean the nega-

tion of p / s, i.e. it will mean the conjunction of p and s. It follows

that

(s / q) | p / s

expresses the incompatibility of s / q with the conjunction of p and s;

in other words, it states that if p and s are both true, s / q is false, i.e. s

and q are both true; in still simpler words, it states that p and s jointly

imply s and q jointly.

Now, put P = p | (q / r),

π = t | (t / t),

Q = (s / q) | p / s.

Then M. Nicod’s sole formal principle of deduction is

P | π/ Q,

in other words, P implies both π and Q.

He employs in addition one non-formal principle belonging to the

theory of types (which need not concern us), and one corresponding

to the principle that, given p, and given that p implies q, we can assert

q. This principle is:

“If p | (r / q) is true, and p is true, then q is true.” From this

apparatus the whole theory of deduction follows, except in so far

as we are concerned with deduction from or to the existence or the

universal truth of “propositional functions,” which we shall consider

in the next chapter.

There is, if I am not mistaken, a certain confusion in the | 1¸j minds

of some authors as to the relation, between propositions, in virtue

of which an inference is valid. In order that it may be valid to infer

q from p, it is only necessary that p should be true and that the

proposition “not-p or q” should be true. Whenever this is the case, it

is clear that q must be true. But inference will only in fact take place

when the proposition “not-p or q” is known otherwise than through

knowledge of not-p or knowledge of q. Whenever p is false, “not-p or

q” is true, but is useless for inference, which requires that p should

Chap. XIV. Incompatibility and the Theory of Deduction 11¸

be true. Whenever q is already known to be true, “not-p or q” is of

course also known to be true, but is again useless for inference, since

q is already known, and therefore does not need to be inferred. In

fact, inference only arises when “not-p or q” can be known without

our knowing already which of the two alternatives it is that makes

the disjunction true. Now, the circumstances under which this occurs

are those in which certain relations of form exist between p and q. For

example, we knowthat if r implies the negation of s, then s implies the

negation of r. Between “r implies not-s” and “s implies not-r” there

is a formal relation which enables us to know that the ﬁrst implies

the second, without having ﬁrst to know that the ﬁrst is false or to

know that the second is true. It is under such circumstances that the

relation of implication is practically useful for drawing inferences.

But this formal relation is only required in order that we may be

able to know that either the premiss is false or the conclusion is true.

It is the truth of “not-p or q” that is required for the validity of the

inference; what is required further is only required for the practi-

cal feasibility of the inference. Professor C. I. Lewis

6

has especially

studied the narrower, formal relation which we may call “formal

deducibility.” He urges that the wider relation, that expressed by

“not-p or q,” should not be called “implication.” That is, however, a

matter of words. | 1¸± Provided our use of words is consistent, it matters

little how we deﬁne them. The essential point of diﬀerence between

the theory which I advocate and the theory advocated by Professor

Lewis is this: He maintains that, when one proposition q is “formally

deducible” from another p, the relation which we perceive between

them is one which he calls “strict implication,” which is not the rela-

tion expressed by “not-p or q” but a narrower relation, holding only

when there are certain formal connections between p and q. I main-

tain that, whether or not there be such a relation as he speaks of, it is

in any case one that mathematics does not need, and therefore one

that, on general grounds of economy, ought not to be admitted into

our apparatus of fundamental notions; that, whenever the relation

of “formal deducibility” holds between two propositions, it is the

case that we can see that either the ﬁrst is false or the second true,

and that nothing beyond this fact is necessary to be admitted into

our premisses; and that, ﬁnally, the reasons of detail which Professor

Lewis adduces against the view which I advocate can all be met in de-

tail, and depend for their plausibility upon a covert and unconscious

assumption of the point of view which I reject. I conclude, therefore,

that there is no need to admit as a fundamental notion any form of

implication not expressible as a truth-function.

6

See Mind, vol. xxi., 1o1z, pp. ¸zz–¸j1; and vol. xxiii., 1o1±, pp. z±o–z±;.

CHAPTER XV

PROPOSITIONAL FUNCTIONS

1¸¸ When, in the preceding chapter, we were discussing propositions,

we did not attempt to give a deﬁnition of the word “proposition.”

But although the word cannot be formally deﬁned, it is necessary to

say something as to its meaning, in order to avoid the very common

confusion with “propositional functions,” which are to be the topic

of the present chapter.

We mean by a “proposition” primarily a form of words which

expresses what is either true or false. I say “primarily,” because

I do not wish to exclude other than verbal symbols, or even mere

thoughts if they have a symbolic character. But I think the word

“proposition” should be limited to what may, in some sense, be called

“symbols,” and further to such symbols as give expression to truth and

falsehood. Thus “two and two are four” and “two and two are ﬁve”

will be propositions, and so will “Socrates is a man” and “Socrates

is not a man.” The statement: “Whatever numbers a and b may

be, (a + b)

z

= a

z

+ zab + b

z

” is a proposition; but the bare formula

“(a +b)

z

= a

z

+zab +b

z

” alone is not, since it asserts nothing deﬁnite

unless we are further told, or led to suppose, that a and b are to

have all possible values, or are to have such-and-such values. The

former of these is tacitly assumed, as a rule, in the enunciation of

mathematical formulæ, which thus become propositions; but if no

such assumption were made, they would be “propositional functions.”

A “propositional function,” in fact, is an expression containing one

or more undetermined constituents, | 1¸6 such that, when values are

assigned to these constituents, the expression becomes a proposition.

In other words, it is a function whose values are propositions. But this

latter deﬁnition must be used with caution. A descriptive function,

e.g. “the hardest proposition in A’s mathematical treatise,” will not

be a propositional function, although its values are propositions. But

in such a case the propositions are only described: in a propositional

function, the values must actually enunciate propositions.

Examples of propositional functions are easy to give: “x is human”

is a propositional function; so long as x remains undetermined, it is

116

Chap. XV. Propositional Functions 11;

neither true nor false, but when a value is assigned to x it becomes

a true or false proposition. Any mathematical equation is a proposi-

tional function. So long as the variables have no deﬁnite value, the

equation is merely an expression awaiting determination in order to

become a true or false proposition. If it is an equation containing one

variable, it becomes true when the variable is made equal to a root of

the equation, otherwise it becomes false; but if it is an “identity” it

will be true when the variable is any number. The equation to a curve

in a plane or to a surface in space is a propositional function, true for

values of the co-ordinates belonging to points on the curve or surface,

false for other values. Expressions of traditional logic such as “all A

is B” are propositional functions: A and B have to be determined as

deﬁnite classes before such expressions become true or false.

The notion of “cases” or “instances” depends upon propositional

functions. Consider, for example, the kind of process suggested by

what is called “generalisation,” and let us take some very primitive

example, say, “lightning is followed by thunder.” We have a number

of “instances” of this, i.e. a number of propositions such as: “this

is a ﬂash of lightning and is followed by thunder.” What are these

occurrences “instances” of? They are instances of the propositional

function: “If x is a ﬂash of lightning, x is followed by thunder.” The

process of generalisation (with whose validity we are | 1¸; fortunately

not concerned) consists in passing from a number of such instances

to the universal truth of the propositional function: “If x is a ﬂash

of lightning, x is followed by thunder.” It will be found that, in an

analogous way, propositional functions are always involved whenever

we talk of instances or cases or examples.

We do not need to ask, or attempt to answer, the question: “What

is a propositional function?” A propositional function standing all

alone may be taken to be a mere schema, a mere shell, an empty

receptacle for meaning, not something already signiﬁcant. We are

concerned with propositional functions, broadly speaking, in two

ways: ﬁrst, as involved in the notions “true in all cases” and “true

in some cases”; secondly, as involved in the theory of classes and

relations. The second of these topics we will postpone to a later

chapter; the ﬁrst must occupy us now.

When we say that something is “always true” or “true in all cases,”

it is clear that the “something” involved cannot be a proposition. A

proposition is just true or false, and there is an end of the matter.

There are no instances or cases of “Socrates is a man” or “Napoleon

died at St Helena.” These are propositions, and it would be mean-

ingless to speak of their being true “in all cases.” This phrase is only

applicable to propositional functions. Take, for example, the sort of

thing that is often said when causation is being discussed. (We are

not concerned with the truth or falsehood of what is said, but only

Chap. XV. Propositional Functions 118

with its logical analysis.) We are told that A is, in every instance,

followed by B. Now if there are “instances” of A, A must be some

general concept of which it is signiﬁcant to say “x

1

is A,” “x

z

is A,”

“x

j

is A,” and so on, where x

1

, x

z

, x

j

are particulars which are not

identical one with another. This applies, e.g., to our previous case

of lightning. We say that lightning (A) is followed by thunder (B).

But the separate ﬂashes are particulars, not identical, but sharing the

common property of being lightning. The only way of expressing a

| 1¸8 common property generally is to say that a common property of a

number of objects is a propositional function which becomes true

when any one of these objects is taken as the value of the variable. In

this case all the objects are “instances” of the truth of the proposi-

tional function—for a propositional function, though it cannot itself

be true or false, is true in certain instances and false in certain others,

unless it is “always true” or “always false.” When, to return to our

example, we say that A is in every instance followed by B, we mean

that, whatever x may be, if x is an A, it is followed by a B; that is, we

are asserting that a certain propositional function is “always true.”

Sentences involving such words as “all,” “every,” “a,” “the,” “some”

require propositional functions for their interpretation. The way in

which propositional functions occur can be explained by means of

two of the above words, namely, “all” and “some.”

There are, in the last analysis, only two things that can be done

with a propositional function: one is to assert that it is true in all cases,

the other to assert that it is true in at least one case, or in some cases (as

we shall say, assuming that there is to be no necessary implication of

a plurality of cases). All the other uses of propositional functions can

be reduced to these two. When we say that a propositional function

is true “in all cases,” or “always” (as we shall also say, without any

temporal suggestion), we mean that all its values are true. If “φx” is

the function, and a is the right sort of object to be an argument to “φx,”

then φa is to be true, however a may have been chosen. For example,

“if a is human, a is mortal” is true whether a is human or not; in

fact, every proposition of this form is true. Thus the propositional

function “if x is human, x is mortal” is “always true,” or “true in

all cases.” Or, again, the statement “there are no unicorns” is the

same as the statement “the propositional function ‘x is not a unicorn’

is true in all cases.” The assertions in the preceding chapter about

propositions, e.g. “‘p or q’ implies ‘q or p,’” are really assertions | 1¸o that

certain propositional functions are true in all cases. We do not assert

the above principle, for example, as being true only of this or that

particular p or q, but as being true of any p or q concerning which

it can be made signiﬁcantly. The condition that a function is to be

signiﬁcant for a given argument is the same as the condition that it

shall have a value for that argument, either true or false. The study

Chap. XV. Propositional Functions 11o

of the conditions of signiﬁcance belongs to the doctrine of types,

which we shall not pursue beyond the sketch given in the preceding

chapter.

Not only the principles of deduction, but all the primitive proposi-

tions of logic, consist of assertions that certain propositional functions

are always true. If this were not the case, they would have to mention

particular things or concepts—Socrates, or redness, or east and west,

or what not—and clearly it is not the province of logic to make asser-

tions which are true concerning one such thing or concept but not

concerning another. It is part of the deﬁnition of logic (but not the

whole of its deﬁnition) that all its propositions are completely general,

i.e. they all consist of the assertion that some propositional function

containing no constant terms is always true. We shall return in our

ﬁnal chapter to the discussion of propositional functions containing

no constant terms. For the present we will proceed to the other thing

that is to be done with a propositional function, namely, the assertion

that it is “sometimes true,” i.e. true in at least one instance.

When we say “there are men,” that means that the propositional

function “x is a man” is sometimes true. When we say “some men

are Greeks,” that means that the propositional function “x is a man

and a Greek” is sometimes true. When we say “cannibals still exist in

Africa,” that means that the propositional function “x is a cannibal

now in Africa” is sometimes true, i.e. is true for some values of x. To

say “there are at least n individuals in the world” is to say that the

propositional function “α is a class of individuals and a member of

the cardinal number n” is sometimes true, or, as we may say, is true

for certain | 16o values of α. This form of expression is more convenient

when it is necessary to indicate which is the variable constituent

which we are taking as the argument to our propositional function.

For example, the above propositional function, which we may shorten

to “α is a class of n individuals,” contains two variables, α and n. The

axiom of inﬁnity, in the language of propositional functions, is: “The

propositional function ‘if n is an inductive number, it is true for some

values of α that α is a class of n individuals’ is true for all possible

values of n.” Here there is a subordinate function, “α is a class of n

individuals,” which is said to be, in respect of α, sometimes true; and

the assertion that this happens if n is an inductive number is said to

be, in respect of n, always true.

The statement that a function φx is always true is the negation of

the statement that not-φx is sometimes true, and the statement that

φx is sometimes true is the negation of the statement that not-φx is

always true. Thus the statement “all men are mortals” is the negation

of the statement that the function “x is an immortal man” is some-

times true. And the statement “there are unicorns” is the negation of

Chap. XV. Propositional Functions 1zo

the statement that the function “x is not a unicorn” is always true.

1

We say that φx is “never true” or “always false” if not-φx is always

true. We can, if we choose, take one of the pair “always,” “sometimes”

as a primitive idea, and deﬁne the other by means of the one and

negation. Thus if we choose “sometimes” as our primitive idea, we

can deﬁne: “‘φx is always true’ is to mean ‘it is false that not-φx is

sometimes true.’” But for reasons connected with the theory of types

it seems more correct to take both “always” and “sometimes” as prim-

itive ideas, and deﬁne by their means the negation of propositions

in which they occur. That is to say, assuming that we have already |

161 deﬁned (or adopted as a primitive idea) the negation of propositions

of the type to which φx belongs, we deﬁne: “The negation of ‘φx

always’ is ‘not-φx sometimes’; and the negation of ‘φx sometimes’ is

‘not-φx always.’” In like manner we can re-deﬁne disjunction and the

other truth-functions, as applied to propositions containing apparent

variables, in terms of the deﬁnitions and primitive ideas for propo-

sitions containing no apparent variables. Propositions containing

no apparent variables are called “elementary propositions.” From

these we can mount up step by step, using such methods as have

just been indicated, to the theory of truth-functions as applied to

propositions containing one, two, three . . . variables, or any number

up to n, where n is any assigned ﬁnite number.

z

The forms which are taken as simplest in traditional formal logic

are really far from being so, and all involve the assertion of all values

or some values of a compound propositional function. Take, to begin

with, “all S is P.” We will take it that S is deﬁned by a propositional

function φx, and P by a propositional function ψx. E.g., if S is men,

φx will be “x is human”; if P is mortals, ψx will be “there is a time

at which x dies.” Then “all S is P” means: “‘φx implies ψx’ is always

true.” It is to be observed that “all S is P” does not apply only to those

terms that actually are S’s; it says something equally about terms

which are not S’s. Suppose we come across an x of which we do not

know whether it is an S or not; still, our statement “all S is P” tells

us something about x, namely, that if x is an S, then x is a P. And this

is every bit as true when x is not an S as when x is an S. If it were

not equally true in both cases, the reductio ad absurdum would not

be a valid method; for the essence of this method consists in using

implications in cases where (as it afterwards turns out) the hypothesis

is false. We may put the matter another way. In order to understand

“all S is P,” it is not necessary to be able to enumerate what terms are

S’s; provided we know what is meant by being an S and what by being

1

For linguistic reasons, to avoid suggesting either the plural or the singular, it is

often convenient to say “φx is not always false” rather than “φx sometimes” or “φx

is sometimes true.”

z

The method of deduction is given in Principia Mathematica, vol. i. ∗o.

Chap. XV. Propositional Functions 1z1

a P, we can understand completely what is actually aﬃrmed | 16z by “all

S is P,” however little we may know of actual instances of either. This

shows that it is not merely the actual terms that are S’s that are rele-

vant in the statement “all S is P,” but all the terms concerning which

the supposition that they are S’s is signiﬁcant, i.e. all the terms that

are S’s, together with all the terms that are not S’s—i.e. the whole of

the appropriate logical “type.” What applies to statements about all

applies also to statements about some. “There are men,” e.g., means

that “x is human” is true for some values of x. Here all values of x

(i.e. all values for which “x is human” is signiﬁcant, whether true or

false) are relevant, and not only those that in fact are human. (This

becomes obvious if we consider how we could prove such a statement

to be false.) Every assertion about “all” or “some” thus involves not

only the arguments that make a certain function true, but all that

make it signiﬁcant, i.e. all for which it has a value at all, whether true

or false.

We may now proceed with our interpretation of the traditional

forms of the old-fashioned formal logic. We assume that S is those

terms x for which φx is true, and P is those for which ψx is true. (As

we shall see in a later chapter, all classes are derived in this way from

propositional functions.) Then:

“All S is P” means “‘φx implies ψx’ is always true.”

“Some S is P” means “‘φx and ψx’ is sometimes true.”

“No S is P” means “‘φx implies not-ψx’ is always true.”

“Some S is not P” means “‘φx and not-ψx’ is sometimes true.”

It will be observed that the propositional functions which are here

asserted for all or some values are not φx and ψx themselves, but

truth-functions of φx and ψx for the same argument x. The easiest

way to conceive of the sort of thing that is intended is to start not

from φx and ψx in general, but from φa and ψa, where a is some

constant. Suppose we are considering “all men are mortal”: we will

begin with

“If Socrates is human, Socrates is mortal,” |

16j and then we will regard “Socrates” as replaced by a variable x wher-

ever “Socrates” occurs. The object to be secured is that, although x

remains a variable, without any deﬁnite value, yet it is to have the

same value in “φx” as in “ψx” when we are asserting that “φx implies

ψx” is always true. This requires that we shall start with a function

whose values are such as “φa implies ψa,” rather than with two sepa-

rate functions φx and ψx; for if we start with two separate functions

we can never secure that the x, while remaining undetermined, shall

have the same value in both.

Chap. XV. Propositional Functions 1zz

For brevity we say “φx always implies ψx” when we mean that

“φx implies ψx” is always true. Propositions of the form “φx always

implies ψx” are called “formal implications”; this name is given

equally if there are several variables.

The above deﬁnitions show how far removed from the simplest

forms are such propositions as “all S is P,” with which traditional logic

begins. It is typical of the lack of analysis involved that traditional

logic treats “all S is P” as a proposition of the same form as “x is P”—

e.g., it treats “all men are mortal” as of the same form as “Socrates

is mortal.” As we have just seen, the ﬁrst is of the form “φx always

implies ψx,” while the second is of the form “ψx.” The emphatic

separation of these two forms, which was eﬀected by Peano and Frege,

was a very vital advance in symbolic logic.

It will be seen that “all S is P” and “no S is P” do not really diﬀer

in form, except by the substitution of not-ψx for ψx, and that the

same applies to “some S is P” and “some S is not P.” It should also

be observed that the traditional rules of conversion are faulty, if we

adopt the view, which is the only technically tolerable one, that such

propositions as “all S is P” do not involve the “existence” of S’s, i.e.

do not require that there should be terms which are S’s. The above

deﬁnitions lead to the result that, if φx is always false, i.e. if there are

no S’s, then “all S is P” and “no S is P” will both be true, | 16± whatever

P may be. For, according to the deﬁnition in the last chapter, “φx

implies ψx” means “not-φx or ψx,” which is always true if not-φx is

always true. At the ﬁrst moment, this result might lead the reader

to desire diﬀerent deﬁnitions, but a little practical experience soon

shows that any diﬀerent deﬁnitions would be inconvenient and would

conceal the important ideas. The proposition “φx always implies ψx,

and φx is sometimes true” is essentially composite, and it would

be very awkward to give this as the deﬁnition of “all S is P,” for

then we should have no language left for “φx always implies ψx,”

which is needed a hundred times for once that the other is needed.

But, with our deﬁnitions, “all S is P” does not imply “some S is P,”

since the ﬁrst allows the non-existence of S and the second does not;

thus conversion per accidens becomes invalid, and some moods of the

syllogism are fallacious, e.g. Darapti: “All M is S, all M is P, therefore

some S is P,” which fails if there is no M.

The notion of “existence” has several forms, one of which will

occupy us in the next chapter; but the fundamental formis that which

is derived immediately from the notion of “sometimes true.” We say

that an argument a “satisﬁes” a function φx if φa is true; this is the

same sense in which the roots of an equation are said to satisfy the

equation. Now if φx is sometimes true, we may say there are x’s for

which it is true, or we may say “arguments satisfying φx exist.” This

is the fundamental meaning of the word “existence.” Other meanings

Chap. XV. Propositional Functions 1zj

are either derived from this, or embody mere confusion of thought.

We may correctly say “men exist,” meaning that “x is a man” is some-

times true. But if we make a pseudo-syllogism: “Men exist, Socrates

is a man, therefore Socrates exists,” we are talking nonsense, since

“Socrates” is not, like “men,” merely an undetermined argument to a

given propositional function. The fallacy is closely analogous to that

of the argument: “Men are numerous, Socrates is a man, therefore

Socrates is numerous.” In this case it is obvious that the conclusion is

nonsensical, but | 16¸ in the case of existence it is not obvious, for reasons

which will appear more fully in the next chapter. For the present let

us merely note the fact that, though it is correct to say “men exist,”

it is incorrect, or rather meaningless, to ascribe existence to a given

particular x who happens to be a man. Generally, “terms satisfying

φx exist” means “φx is sometimes true”; but “a exists” (where a is a

term satisfying φx) is a mere noise or shape, devoid of signiﬁcance.

It will be found that by bearing in mind this simple fallacy we can

solve many ancient philosophical puzzles concerning the meaning of

existence.

Another set of notions as to which philosophy has allowed itself

to fall into hopeless confusions through not suﬃciently separating

propositions and propositional functions are the notions of “modal-

ity”: necessary, possible, and impossible. (Sometimes contingent or

assertoric is used instead of possible.) The traditional view was that,

among true propositions, some were necessary, while others were

merely contingent or assertoric; while among false propositions some

were impossible, namely, those whose contradictories were neces-

sary, while others merely happened not to be true. In fact, however,

there was never any clear account of what was added to truth by the

conception of necessity. In the case of propositional functions, the

threefold division is obvious. If “φx” is an undetermined value of

a certain propositional function, it will be necessary if the function

is always true, possible if it is sometimes true, and impossible if it is

never true. This sort of situation arises in regard to probability, for

example. Suppose a ball x is drawn from a bag which contains a

number of balls: if all the balls are white, “x is white” is necessary; if

some are white, it is possible; if none, it is impossible. Here all that

is known about x is that it satisﬁes a certain propositional function,

namely, “x was a ball in the bag.” This is a situation which is general

in probability problems and not uncommon in practical life—e.g.

when a person calls of whom we know nothing except that he brings

a letter of introduction from our friend so-and-so. In all such | 166 cases,

as in regard to modality in general, the propositional function is rele-

vant. For clear thinking, in many very diverse directions, the habit of

keeping propositional functions sharply separated from propositions

Chap. XV. Propositional Functions 1z±

is of the utmost importance, and the failure to do so in the past has

been a disgrace to philosophy.

CHAPTER XVI

DESCRIPTIONS

16; We dealt in the preceding chapter with the words all and some; in

this chapter we shall consider the word the in the singular, and in

the next chapter we shall consider the word the in the plural. It may

be thought excessive to devote two chapters to one word, but to the

philosophical mathematician it is a word of very great importance:

like Browning’s Grammarian with the enclitic δ, I would give the

doctrine of this word if I were “dead from the waist down” and not

merely in a prison.

We have already had occasion to mention “descriptive functions,”

i.e. such expressions as “the father of x” or “the sine of x.” These are

to be deﬁned by ﬁrst deﬁning “descriptions.”

A “description” may be of two sorts, deﬁnite and indeﬁnite (or

ambiguous). An indeﬁnite description is a phrase of the form “a

so-and-so,” and a deﬁnite description is a phrase of the form “the

so-and-so” (in the singular). Let us begin with the former.

“Who did you meet?” “I met a man.” “That is a very indeﬁnite

description.” We are therefore not departing from usage in our termi-

nology. Our question is: What do I really assert when I assert “I met

a man”? Let us assume, for the moment, that my assertion is true,

and that in fact I met Jones. It is clear that what I assert is not “I met

Jones.” I may say “I met a man, but it was not Jones”; in that case,

though I lie, I do not contradict myself, as I should do if when I say

I met a | 168 man I really mean that I met Jones. It is clear also that the

person to whom I am speaking can understand what I say, even if he

is a foreigner and has never heard of Jones.

But we may go further: not only Jones, but no actual man, enters

into my statement. This becomes obvious when the statement is false,

since then there is no more reason why Jones should be supposed to

enter into the proposition than why anyone else should. Indeed the

statement would remain signiﬁcant, though it could not possibly be

true, even if there were no man at all. “I met a unicorn” or “I met a

sea-serpent” is a perfectly signiﬁcant assertion, if we know what it

would be to be a unicorn or a sea-serpent, i.e. what is the deﬁnition of

1z¸

Chap. XVI. Descriptions 1z6

these fabulous monsters. Thus it is only what we may call the concept

that enters into the proposition. In the case of “unicorn,” for example,

there is only the concept: there is not also, somewhere among the

shades, something unreal which may be called “a unicorn.” Therefore,

since it is signiﬁcant (though false) to say “I met a unicorn,” it is clear

that this proposition, rightly analysed, does not contain a constituent

“a unicorn,” though it does contain the concept “unicorn.”

The question of “unreality,” which confronts us at this point, is a

very important one. Misled by grammar, the great majority of those

logicians who have dealt with this question have dealt with it on mis-

taken lines. They have regarded grammatical form as a surer guide

in analysis than, in fact, it is. And they have not known what diﬀer-

ences in grammatical form are important. “I met Jones” and “I met

a man” would count traditionally as propositions of the same form,

but in actual fact they are of quite diﬀerent forms: the ﬁrst names

an actual person, Jones; while the second involves a propositional

function, and becomes, when made explicit: “The function ‘I met x

and x is human’ is sometimes true.” (It will be remembered that we

adopted the convention of using “sometimes” as not implying more

than once.) This proposition is obviously not of the form “I met x,”

which accounts | 16o for the existence of the proposition “I met a unicorn”

in spite of the fact that there is no such thing as “a unicorn.”

For want of the apparatus of propositional functions, many logi-

cians have been driven to the conclusion that there are unreal objects.

It is argued, e.g. by Meinong,

1

that we can speak about “the golden

mountain,” “the round square,” and so on; we can make true propo-

sitions of which these are the subjects; hence they must have some

kind of logical being, since otherwise the propositions in which they

occur would be meaningless. In such theories, it seems to me, there is

a failure of that feeling for reality which ought to be preserved even

in the most abstract studies. Logic, I should maintain, must no more

admit a unicorn than zoology can; for logic is concerned with the

real world just as truly as zoology, though with its more abstract and

general features. To say that unicorns have an existence in heraldry,

or in literature, or in imagination, is a most pitiful and paltry evasion.

What exists in heraldry is not an animal, made of ﬂesh and blood,

moving and breathing of its own initiative. What exists is a picture,

or a description in words. Similarly, to maintain that Hamlet, for ex-

ample, exists in his own world, namely, in the world of Shakespeare’s

imagination, just as truly as (say) Napoleon existed in the ordinary

world, is to say something deliberately confusing, or else confused to

a degree which is scarcely credible. There is only one world, the “real”

world: Shakespeare’s imagination is part of it, and the thoughts that

1

Untersuchungen zur Gegenstandstheorie und Psychologie, 1oo±.

Chap. XVI. Descriptions 1z;

he had in writing Hamlet are real. So are the thoughts that we have

in reading the play. But it is of the very essence of ﬁction that only

the thoughts, feelings, etc., in Shakespeare and his readers are real,

and that there is not, in addition to them, an objective Hamlet. When

you have taken account of all the feelings roused by Napoleon in

writers and readers of history, you have not touched the actual man;

but in the case of Hamlet you have come to the end of him. If no one

thought about Hamlet, there would be nothing | 1;o left of him; if no one

had thought about Napoleon, he would have soon seen to it that some

one did. The sense of reality is vital in logic, and whoever juggles

with it by pretending that Hamlet has another kind of reality is doing

a disservice to thought. A robust sense of reality is very necessary

in framing a correct analysis of propositions about unicorns, golden

mountains, round squares, and other such pseudo-objects.

In obedience to the feeling of reality, we shall insist that, in the

analysis of propositions, nothing “unreal” is to be admitted. But,

after all, if there is nothing unreal, how, it may be asked, could we ad-

mit anything unreal? The reply is that, in dealing with propositions,

we are dealing in the ﬁrst instance with symbols, and if we attribute

signiﬁcance to groups of symbols which have no signiﬁcance, we shall

fall into the error of admitting unrealities, in the only sense in which

this is possible, namely, as objects described. In the proposition “I

met a unicorn,” the whole four words together make a signiﬁcant

proposition, and the word “unicorn” by itself is signiﬁcant, in just

the same sense as the word “man.” But the two words “a unicorn” do

not form a subordinate group having a meaning of its own. Thus if

we falsely attribute meaning to these two words, we ﬁnd ourselves

saddled with “a unicorn,” and with the problem how there can be

such a thing in a world where there are no unicorns. “A unicorn” is an

indeﬁnite description which describes nothing. It is not an indeﬁnite

description which describes something unreal. Such a proposition

as “x is unreal” only has meaning when “x” is a description, deﬁ-

nite or indeﬁnite; in that case the proposition will be true if “x” is

a description which describes nothing. But whether the description

“x” describes something or describes nothing, it is in any case not a

constituent of the proposition in which it occurs; like “a unicorn” just

now, it is not a subordinate group having a meaning of its own. All

this results from the fact that, when “x” is a description, “x is unreal”

or “x does not exist” is not nonsense, but is always signiﬁcant and

sometimes true. |

1;1 We may now proceed to deﬁne generally the meaning of propo-

sitions which contain ambiguous descriptions. Suppose we wish to

make some statement about “a so-and-so,” where “so-and-so’s” are

those objects that have a certain property φ, i.e. those objects x for

which the propositional function φx is true. (E.g. if we take “a man”

Chap. XVI. Descriptions 1z8

as our instance of “a so-and-so,” φx will be “x is human.”) Let us

now wish to assert the property ψ of “a so-and-so,” i.e. we wish to

assert that “a so-and-so” has that property which x has when ψx is

true. (E.g. in the case of “I met a man,” ψx will be “I met x.”) Now the

proposition that “a so-and-so” has the property ψ is not a proposition

of the form “ψx.” If it were, “a so-and-so” would have to be identical

with x for a suitable x; and although (in a sense) this may be true in

some cases, it is certainly not true in such a case as “a unicorn.” It is

just this fact, that the statement that a so-and-so has the property ψ

is not of the form ψx, which makes it possible for “a so-and-so” to

be, in a certain clearly deﬁnable sense, “unreal.” The deﬁnition is as

follows:—

The statement that “an object having the property φ has the prop-

erty ψ”

means:

“The joint assertion of φx and ψx is not always false.”

So far as logic goes, this is the same proposition as might be

expressed by “some φ’s are ψ’s”; but rhetorically there is a diﬀerence,

because in the one case there is a suggestion of singularity, and in the

other case of plurality. This, however, is not the important point. The

important point is that, when rightly analysed, propositions verbally

about “a so-and-so” are found to contain no constituent represented

by this phrase. And that is why such propositions can be signiﬁcant

even when there is no such thing as a so-and-so.

The deﬁnition of existence, as applied to ambiguous descriptions,

results fromwhat was said at the end of the preceding chapter. We say

that “men exist” or “a man exists” if the | 1;z propositional function “x is

human” is sometimes true; and generally “a so-and-so” exists if “x is

so-and-so” is sometimes true. We may put this in other language. The

proposition “Socrates is a man” is no doubt equivalent to “Socrates is

human,” but it is not the very same proposition. The is of “Socrates

is human” expresses the relation of subject and predicate; the is of

“Socrates is a man” expresses identity. It is a disgrace to the human

race that it has chosen to employ the same word “is” for these two

entirely diﬀerent ideas—a disgrace which a symbolic logical language

of course remedies. The identity in “Socrates is a man” is identity

between an object named (accepting “Socrates” as a name, subject to

qualiﬁcations explained later) and an object ambiguously described.

An object ambiguously described will “exist” when at least one such

proposition is true, i.e. when there is at least one true proposition of

the form “x is a so-and-so,” where “x” is a name. It is characteristic

of ambiguous (as opposed to deﬁnite) descriptions that there may

be any number of true propositions of the above form—Socrates is a

Chap. XVI. Descriptions 1zo

man, Plato is a man, etc. Thus “a man exists” follows from Socrates,

or Plato, or anyone else. With deﬁnite descriptions, on the other

hand, the corresponding form of proposition, namely, “x is the so-

and-so” (where “x” is a name), can only be true for one value of x at

most. This brings us to the subject of deﬁnite descriptions, which

are to be deﬁned in a way analogous to that employed for ambiguous

descriptions, but rather more complicated.

We come now to the main subject of the present chapter, namely,

the deﬁnition of the word the (in the singular). One very important

point about the deﬁnition of “a so-and-so” applies equally to “the

so-and-so”; the deﬁnition to be sought is a deﬁnition of propositions

in which this phrase occurs, not a deﬁnition of the phrase itself in

isolation. In the case of “a so-and-so,” this is fairly obvious: no one

could suppose that “a man” was a deﬁnite object, which could be

deﬁned by itself. | 1;j Socrates is a man, Plato is a man, Aristotle is a

man, but we cannot infer that “a man” means the same as “Socrates”

means and also the same as “Plato” means and also the same as

“Aristotle” means, since these three names have diﬀerent meanings.

Nevertheless, when we have enumerated all the men in the world,

there is nothing left of which we can say, “This is a man, and not

only so, but it is the ‘a man,’ the quintessential entity that is just an

indeﬁnite man without being anybody in particular.” It is of course

quite clear that whatever there is in the world is deﬁnite: if it is a

man it is one deﬁnite man and not any other. Thus there cannot be

such an entity as “a man” to be found in the world, as opposed to

speciﬁc men. And accordingly it is natural that we do not deﬁne “a

man” itself, but only the propositions in which it occurs.

In the case of “the so-and-so” this is equally true, though at ﬁrst

sight less obvious. We may demonstrate that this must be the case,

by a consideration of the diﬀerence between a name and a deﬁnite

description. Take the proposition, “Scott is the author of Waverley.”

We have here a name, “Scott,” and a description, “the author of

Waverley,” which are asserted to apply to the same person. The

distinction between a name and all other symbols may be explained

as follows:—

A name is a simple symbol whose meaning is something that can

only occur as subject, i.e. something of the kind that, in Chapter

XIII., we deﬁned as an “individual” or a “particular.” And a “simple”

symbol is one which has no parts that are symbols. Thus “Scott”

is a simple symbol, because, though it has parts (namely, separate

letters), these parts are not symbols. On the other hand, “the author

of Waverley” is not a simple symbol, because the separate words

that compose the phrase are parts which are symbols. If, as may be

the case, whatever seems to be an “individual” is really capable of

further analysis, we shall have to content ourselves with what may be

Chap. XVI. Descriptions 1jo

called “relative individuals,” which will be terms that, throughout the

context in question, are never analysed and never occur | 1;± otherwise

than as subjects. And in that case we shall have correspondingly

to content ourselves with “relative names.” From the standpoint

of our present problem, namely, the deﬁnition of descriptions, this

problem, whether these are absolute names or only relative names,

may be ignored, since it concerns diﬀerent stages in the hierarchy of

“types,” whereas we have to compare such couples as “Scott” and “the

author of Waverley,” which both apply to the same object, and do not

raise the problem of types. We may, therefore, for the moment, treat

names as capable of being absolute; nothing that we shall have to say

will depend upon this assumption, but the wording may be a little

shortened by it.

We have, then, two things to compare: (1) a name, which is a sim-

ple symbol, directly designating an individual which is its meaning,

and having this meaning in its own right, independently of the mean-

ings of all other words; (z) a description, which consists of several

words, whose meanings are already ﬁxed, and from which results

whatever is to be taken as the “meaning” of the description.

A proposition containing a description is not identical with what

that proposition becomes when a name is substituted, even if the

name names the same object as the description describes. “Scott is the

author of Waverley” is obviously a diﬀerent proposition from “Scott is

Scott”: the ﬁrst is a fact in literary history, the second a trivial truism.

And if we put anyone other than Scott in place of “the author of

Waverley,” our proposition would become false, and would therefore

certainly no longer be the same proposition. But, it may be said, our

proposition is essentially of the same form as (say) “Scott is Sir Wal-

ter,” in which two names are said to apply to the same person. The

reply is that, if “Scott is Sir Walter” really means “the person named

‘Scott’ is the person named ‘Sir Walter,’” then the names are being

used as descriptions: i.e. the individual, instead of being named, is

being described as the person having that name. This is a way in

which names are frequently used | 1;¸ in practice, and there will, as a

rule, be nothing in the phraseology to show whether they are being

used in this way or as names. When a name is used directly, merely to

indicate what we are speaking about, it is no part of the fact asserted,

or of the falsehood if our assertion happens to be false: it is merely

part of the symbolism by which we express our thought. What we

want to express is something which might (for example) be translated

into a foreign language; it is something for which the actual words

are a vehicle, but of which they are no part. On the other hand, when

we make a proposition about “the person called ‘Scott,’” the actual

name “Scott” enters into what we are asserting, and not merely into

the language used in making the assertion. Our proposition will now

Chap. XVI. Descriptions 1j1

be a diﬀerent one if we substitute “the person called ‘Sir Walter.’” But

so long as we are using names as names, whether we say “Scott” or

whether we say “Sir Walter” is as irrelevant to what we are asserting

as whether we speak English or French. Thus so long as names are

used as names, “Scott is Sir Walter” is the same trivial proposition

as “Scott is Scott.” This completes the proof that “Scott is the author

of Waverley” is not the same proposition as results from substituting

a name for “the author of Waverley,” no matter what name may be

substituted.

When we use a variable, and speak of a propositional function, φx

say, the process of applying general statements about φx to particular

cases will consist in substituting a name for the letter “x,” assuming

that φ is a function which has individuals for its arguments. Suppose,

for example, that φx is “always true”; let it be, say, the “law of iden-

tity,” x = x. Then we may substitute for “x” any name we choose, and

we shall obtain a true proposition. Assuming for the moment that

“Socrates,” “Plato,” and “Aristotle” are names (a very rash assump-

tion), we can infer from the law of identity that Socrates is Socrates,

Plato is Plato, and Aristotle is Aristotle. But we shall commit a fallacy

if we attempt to infer, without further premisses, that the author of

Waverley is the author of Waverley. This results | 1;6 from what we have

just proved, that, if we substitute a name for “the author of Waverley”

in a proposition, the proposition we obtain is a diﬀerent one. That is

to say, applying the result to our present case: If “x” is a name, “x = x”

is not the same proposition as “the author of Waverley is the author of

Waverley,” no matter what name “x” may be. Thus from the fact that

all propositions of the form “x = x” are true we cannot infer, without

more ado, that the author of Waverley is the author of Waverley. In

fact, propositions of the form “the so-and-so is the so-and-so” are not

always true: it is necessary that the so-and-so should exist (a term

which will be explained shortly). It is false that the present King

of France is the present King of France, or that the round square

is the round square. When we substitute a description for a name,

propositional functions which are “always true” may become false,

if the description describes nothing. There is no mystery in this as

soon as we realise (what was proved in the preceding paragraph)

that when we substitute a description the result is not a value of the

propositional function in question.

We are now in a position to deﬁne propositions in which a deﬁnite

description occurs. The only thing that distinguishes “the so-and-

so” from “a so-and-so” is the implication of uniqueness. We cannot

speak of “the inhabitant of London,” because inhabiting London is an

attribute which is not unique. We cannot speak about “the present

King of France,” because there is none; but we can speak about “the

present King of England.” Thus propositions about “the so-and-so”

Chap. XVI. Descriptions 1jz

always imply the corresponding propositions about “a so-and-so,”

with the addendum that there is not more than one so-and-so. Such

a proposition as “Scott is the author of Waverley” could not be true

if Waverley had never been written, or if several people had written

it; and no more could any other proposition resulting from a propo-

sitional function φx by the substitution of “the author of Waverley”

for “x.” We may say that “the author of Waverley” means “the value

of x for which ‘x wrote | 1;; Waverley’ is true.” Thus the proposition “the

author of Waverley was Scotch,” for example, involves:

(1) “x wrote Waverley” is not always false;

(z) “if x and y wrote Waverley, x and y are identical” is always true;

(j) “if x wrote Waverley, x was Scotch” is always true.

These three propositions, translated into ordinary language, state:

(1) at least one person wrote Waverley;

(z) at most one person wrote Waverley;

(j) whoever wrote Waverley was Scotch.

All these three are implied by “the author of Waverley was Scotch.”

Conversely, the three together (but no two of them) imply that the

author of Waverley was Scotch. Hence the three together may be taken

as deﬁning what is meant by the proposition “the author of Waverley

was Scotch.”

We may somewhat simplify these three propositions. The ﬁrst

and second together are equivalent to: “There is a term c such that

‘x wrote Waverley’ is true when x is c and is false when x is not c.” In

other words, “There is a term c such that ‘x wrote Waverley’ is always

equivalent to ‘x is c.’” (Two propositions are “equivalent” when both

are true or both are false.) We have here, to begin with, two functions

of x, “x wrote Waverley” and “x is c,” and we form a function of c

by considering the equivalence of these two functions of x for all

values of x; we then proceed to assert that the resulting function of c

is “sometimes true,” i.e. that it is true for at least one value of c. (It

obviously cannot be true for more than one value of c.) These two

conditions together are deﬁned as giving the meaning of “the author

of Waverley exists.”

We may now deﬁne “the term satisfying the function φx exists.”

This is the general form of which the above is a particular case. “The

author of Waverley” is “the term satisfying the function ‘x wrote

Waverley.’” And “the so-and-so” will | 1;8 always involve reference to

some propositional function, namely, that which deﬁnes the property

that makes a thing a so-and-so. Our deﬁnition is as follows:—

“The term satisfying the function φx exists” means:

“There is a term c such that φx is always equivalent to ‘x is c.’”

Chap. XVI. Descriptions 1jj

In order to deﬁne “the author of Waverley was Scotch,” we have

still to take account of the third of our three propositions, namely,

“Whoever wrote Waverley was Scotch.” This will be satisﬁed by merely

adding that the c in question is to be Scotch. Thus “the author of

Waverley was Scotch” is:

“There is a term c such that (1) ‘x wrote Waverley’ is always equiv-

alent to ‘x is c,’ (z) c is Scotch.”

And generally: “the term satisfying φx satisﬁes ψx” is deﬁned as

meaning:

“There is a term c such that (1) φx is always equivalent to ‘x is c,’

(z) ψc is true.”

This is the deﬁnition of propositions in which descriptions occur.

It is possible to have much knowledge concerning a term de-

scribed, i.e. to know many propositions concerning “the so-and-so,”

without actually knowing what the so-and-so is, i.e. without know-

ing any proposition of the form “x is the so-and-so,” where “x” is a

name. In a detective story propositions about “the man who did the

deed” are accumulated, in the hope that ultimately they will suﬃce to

demonstrate that it was A who did the deed. We may even go so far as

to say that, in all such knowledge as can be expressed in words—with

the exception of “this” and “that” and a few other words of which the

meaning varies on diﬀerent occasions—no names, in the strict sense,

occur, but what seem like names are really descriptions. We may

inquire signiﬁcantly whether Homer existed, which we could not do

if “Homer” were a name. The proposition “the so-and-so exists” is sig-

niﬁcant, whether true or false; but if a is the so-and-so (where “a” is a

name), the words “a exists” are meaningless. It is only of descriptions

| 1;o —deﬁnite or indeﬁnite—that existence can be signiﬁcantly asserted;

for, if “a” is a name, it must name something: what does not name

anything is not a name, and therefore, if intended to be a name, is a

symbol devoid of meaning, whereas a description, like “the present

King of France,” does not become incapable of occurring signiﬁcantly

merely on the ground that it describes nothing, the reason being that

it is a complex symbol, of which the meaning is derived from that of

its constituent symbols. And so, when we ask whether Homer existed,

we are using the word “Homer” as an abbreviated description: we

may replace it by (say) “the author of the Iliad and the Odyssey.” The

same considerations apply to almost all uses of what look like proper

names.

When descriptions occur in propositions, it is necessary to distin-

guish what may be called “primary” and “secondary” occurrences.

The abstract distinction is as follows. A description has a “primary”

Chap. XVI. Descriptions 1j±

occurrence when the proposition in which it occurs results from sub-

stituting the description for “x” in some propositional function φx; a

description has a “secondary” occurrence when the result of substi-

tuting the description for x in φx gives only part of the proposition

concerned. An instance will make this clearer. Consider “the present

King of France is bald.” Here “the present King of France” has a

primary occurrence, and the proposition is false. Every proposition

in which a description which describes nothing has a primary occur-

rence is false. But now consider “the present King of France is not

bald.” This is ambiguous. If we are ﬁrst to take “x is bald,” then

substitute “the present King of France” for “x,” and then deny the

result, the occurrence of “the present King of France” is secondary

and our proposition is true; but if we are to take “x is not bald” and

substitute “the present King of France” for “x,” then “the present

King of France” has a primary occurrence and the proposition is false.

Confusion of primary and secondary occurrences is a ready source of

fallacies where descriptions are concerned. |

Descriptions 18o occur in mathematics chieﬂy in the form of descrip-

tive functions, i.e. “the term having the relation R to y,” or “the R

of y” as we may say, on the analogy of “the father of y” and similar

phrases. To say “the father of y is rich,” for example, is to say that

the following propositional function of c: “c is rich, and ‘x begat y’ is

always equivalent to ‘x is c,’” is “sometimes true,” i.e. is true for at

least one value of c. It obviously cannot be true for more than one

value.

The theory of descriptions, brieﬂy outlined in the present chapter,

is of the utmost importance both in logic and in theory of knowledge.

But for purposes of mathematics, the more philosophical parts of

the theory are not essential, and have therefore been omitted in the

above account, which has conﬁned itself to the barest mathematical

requisites.

CHAPTER XVII

CLASSES

181 In the present chapter we shall be concerned with the in the plural:

the inhabitants of London, the sons of rich men, and so on. In other

words, we shall be concerned with classes. We saw in Chapter II.

that a cardinal number is to be deﬁned as a class of classes, and in

Chapter III. that the number 1 is to be deﬁned as the class of all

unit classes, i.e. of all that have just one member, as we should say

but for the vicious circle. Of course, when the number 1 is deﬁned

as the class of all unit classes, “unit classes” must be deﬁned so as

not to assume that we know what is meant by “one”; in fact, they

are deﬁned in a way closely analogous to that used for descriptions,

namely: A class α is said to be a “unit” class if the propositional

function “‘x is an α’ is always equivalent to ‘x is c’” (regarded as a

function of c) is not always false, i.e., in more ordinary language, if

there is a term c such that x will be a member of α when x is c but

not otherwise. This gives us a deﬁnition of a unit class if we already

know what a class is in general. Hitherto we have, in dealing with

arithmetic, treated “class” as a primitive idea. But, for the reasons set

forth in Chapter XIII., if for no others, we cannot accept “class” as

a primitive idea. We must seek a deﬁnition on the same lines as the

deﬁnition of descriptions, i.e. a deﬁnition which will assign a meaning

to propositions in whose verbal or symbolic expression words or

symbols apparently representing classes occur, but which will assign

a meaning that altogether eliminates all mention of classes from a

right analysis | 18z of such propositions. We shall then be able to say

that the symbols for classes are mere conveniences, not representing

objects called “classes,” and that classes are in fact, like descriptions,

logical ﬁctions, or (as we say) “incomplete symbols.”

The theory of classes is less complete than the theory of descrip-

tions, and there are reasons (which we shall give in outline) for re-

garding the deﬁnition of classes that will be suggested as not ﬁnally

satisfactory. Some further subtlety appears to be required; but the

reasons for regarding the deﬁnition which will be oﬀered as being

approximately correct and on the right lines are overwhelming.

1j¸

Chap. XVII. Classes 1j6

The ﬁrst thing is to realise why classes cannot be regarded as part

of the ultimate furniture of the world. It is diﬃcult to explain pre-

cisely what one means by this statement, but one consequence which

it implies may be used to elucidate its meaning. If we had a com-

plete symbolic language, with a deﬁnition for everything deﬁnable,

and an undeﬁned symbol for everything indeﬁnable, the undeﬁned

symbols in this language would represent symbolically what I mean

by “the ultimate furniture of the world.” I am maintaining that no

symbols either for “class” in general or for particular classes would

be included in this apparatus of undeﬁned symbols. On the other

hand, all the particular things there are in the world would have to

have names which would be included among undeﬁned symbols. We

might try to avoid this conclusion by the use of descriptions. Take

(say) “the last thing Cæsar saw before he died.” This is a description

of some particular; we might use it as (in one perfectly legitimate

sense) a deﬁnition of that particular. But if “a” is a name for the same

particular, a proposition in which “a” occurs is not (as we saw in

the preceding chapter) identical with what this proposition becomes

when for “a” we substitute “the last thing Cæsar saw before he died.”

If our language does not contain the name “a,” or some other name

for the same particular, we shall have no means of expressing the

proposition which we expressed by means of “a” as opposed to the

one that | 18j we expressed by means of the description. Thus descrip-

tions would not enable a perfect language to dispense with names for

all particulars. In this respect, we are maintaining, classes diﬀer from

particulars, and need not be represented by undeﬁned symbols. Our

ﬁrst business is to give the reasons for this opinion.

We have already seen that classes cannot be regarded as a species

of individuals, on account of the contradiction about classes which

are not members of themselves (explained in Chapter XIII.), and

because we can prove that the number of classes is greater than the

number of individuals.

We cannot take classes in the pure extensional way as simply heaps

or conglomerations. If we were to attempt to do that, we should ﬁnd

it impossible to understand how there can be such a class as the

null-class, which has no members at all and cannot be regarded as a

“heap”; we should also ﬁnd it very hard to understand how it comes

about that a class which has only one member is not identical with

that one member. I do not mean to assert, or to deny, that there are

such entities as “heaps.” As a mathematical logician, I am not called

upon to have an opinion on this point. All that I am maintaining is

that, if there are such things as heaps, we cannot identify them with

the classes composed of their constituents.

We shall come much nearer to a satisfactory theory if we try to

identify classes with propositional functions. Every class, as we

Chap. XVII. Classes 1j;

explained in Chapter II., is deﬁned by some propositional function

which is true of the members of the class and false of other things.

But if a class can be deﬁned by one propositional function, it can

equally well be deﬁned by any other which is true whenever the ﬁrst

is true and false whenever the ﬁrst is false. For this reason the class

cannot be identiﬁed with any one such propositional function rather

than with any other—and given a propositional function, there are

always many others which are true when it is true and false when

it is false. We say that two propositional functions are “formally

equivalent” when this happens. Two propositions are | 18± “equivalent”

when both are true or both false; two propositional functions φx, ψx

are “formally equivalent” when φx is always equivalent to ψx. It is

the fact that there are other functions formally equivalent to a given

function that makes it impossible to identify a class with a function;

for we wish classes to be such that no two distinct classes have exactly

the same members, and therefore two formally equivalent functions

will have to determine the same class.

When we have decided that classes cannot be things of the same

sort as their members, that they cannot be just heaps or aggregates,

and also that they cannot be identiﬁed with propositional functions,

it becomes very diﬃcult to see what they can be, if they are to be

more than symbolic ﬁctions. And if we can ﬁnd any way of dealing

with them as symbolic ﬁctions, we increase the logical security of our

position, since we avoid the need of assuming that there are classes

without being compelled to make the opposite assumption that there

are no classes. We merely abstain from both assumptions. This is an

example of Occam’s razor, namely, “entities are not to be multiplied

without necessity.” But when we refuse to assert that there are classes,

we must not be supposed to be asserting dogmatically that there are

none. We are merely agnostic as regards them: like Laplace, we can

say, “je n’ai pas besoin de cette hypoth` ese.”

Let us set forth the conditions that a symbol must fulﬁl if it is

to serve as a class. I think the following conditions will be found

necessary and suﬃcient:—

(1) Every propositional function must determine a class, con-

sisting of those arguments for which the function is true. Given

any proposition (true or false), say about Socrates, we can imagine

Socrates replaced by Plato or Aristotle or a gorilla or the man in the

moon or any other individual in the world. In general, some of these

substitutions will give a true proposition and some a false one. The

class determined will consist of all those substitutions that give a

true one. Of course, we have still to decide what we mean by “all

those which, etc.” All that | 18¸ we are observing at present is that a class

is rendered determinate by a propositional function, and that every

propositional function determines an appropriate class.

Chap. XVII. Classes 1j8

(z) Two formally equivalent propositional functions must deter-

mine the same class, and two which are not formally equivalent must

determine diﬀerent classes. That is, a class is determined by its mem-

bership, and no two diﬀerent classes can have the same membership.

(If a class is determined by a function φx, we say that a is a “member”

of the class if φa is true.)

(j) We must ﬁnd some way of deﬁning not only classes, but classes

of classes. We saw in Chapter II. that cardinal numbers are to be

deﬁned as classes of classes. The ordinary phrase of elementary

mathematics, “The combinations of n things m at a time” represents

a class of classes, namely, the class of all classes of m terms that can

be selected out of a given class of n terms. Without some symbolic

method of dealing with classes of classes, mathematical logic would

break down.

(±) It must under all circumstances be meaningless (not false) to

suppose a class a member of itself or not a member of itself. This

results from the contradiction which we discussed in Chapter XIII.

(¸) Lastly—and this is the condition which is most diﬃcult of

fulﬁlment—it must be possible to make propositions about all the

classes that are composed of individuals, or about all the classes

that are composed of objects of any one logical “type.” If this were

not the case, many uses of classes would go astray—for example,

mathematical induction. In deﬁning the posterity of a given term, we

need to be able to say that a member of the posterity belongs to all

hereditary classes to which the given term belongs, and this requires

the sort of totality that is in question. The reason there is a diﬃculty

about this condition is that it can be proved to be impossible to speak

of all the propositional functions that can have arguments of a given

type.

We will, to begin with, ignore this last condition and the problems

which it raises. The ﬁrst two conditions may be | 186 taken together. They

state that there is to be one class, no more and no less, for each group

of formally equivalent propositional functions; e.g. the class of men

is to be the same as that of featherless bipeds or rational animals or

Yahoos or whatever other characteristic may be preferred for deﬁning

a human being. Now, when we say that two formally equivalent

propositional functions may be not identical, although they deﬁne

the same class, we may prove the truth of the assertion by pointing

out that a statement may be true of the one function and false of the

other; e.g. “I believe that all men are mortal” may be true, while “I

believe that all rational animals are mortal” may be false, since I may

believe falsely that the Phœnix is an immortal rational animal. Thus

we are led to consider statements about functions, or (more correctly)

functions of functions.

Some of the things that may be said about a function may be re-

garded as said about the class deﬁned by the function, whereas others

Chap. XVII. Classes 1jo

cannot. The statement “all men are mortal” involves the functions

“x is human” and “x is mortal”; or, if we choose, we can say that it

involves the classes men and mortals. We can interpret the statement

in either way, because its truth-value is unchanged if we substitute

for “x is human” or for “x is mortal” any formally equivalent func-

tion. But, as we have just seen, the statement “I believe that all men

are mortal” cannot be regarded as being about the class determined

by either function, because its truth-value may be changed by the

substitution of a formally equivalent function (which leaves the class

unchanged). We will call a statement involving a function φx an

“extensional” function of the function φx, if it is like “all men are

mortal,” i.e. if its truth-value is unchanged by the substitution of any

formally equivalent function; and when a function of a function is

not extensional, we will call it “intensional,” so that “I believe that

all men are mortal” is an intensional function of “x is human” or

“x is mortal.” Thus extensional functions of a function φx may, for

practical | 18; purposes, be regarded as functions of the class determined

by φx, while intensional functions cannot be so regarded.

It is to be observed that all the speciﬁc functions of functions that

we have occasion to introduce in mathematical logic are extensional.

Thus, for example, the two fundamental functions of functions are:

“φx is always true” and “φx is sometimes true.” Each of these has its

truth-value unchanged if any formally equivalent function is substi-

tuted for φx. In the language of classes, if α is the class determined

by φx, “φx is always true” is equivalent to “everything is a member

of α,” and “φx is sometimes true” is equivalent to “α has members”

or (better) “α has at least one member.” Take, again, the condition,

dealt with in the preceding chapter, for the existence of “the term

satisfying φx.” The condition is that there is a term c such that φx

is always equivalent to “x is c.” This is obviously extensional. It is

equivalent to the assertion that the class deﬁned by the function φx

is a unit class, i.e. a class having one member; in other words, a class

which is a member of 1.

Given a function of a function which may or may not be exten-

sional, we can always derive from it a connected and certainly ex-

tensional function of the same function, by the following plan: Let

our original function of a function be one which attributes to φx the

property f ; then consider the assertion “there is a function having

the property f and formally equivalent to φx.” This is an extensional

function of φx; it is true when our original statement is true, and it

is formally equivalent to the original function of φx if this original

function is extensional; but when the original function is intensional,

the new one is more often true than the old one. For example, con-

sider again “I believe that all men are mortal,” regarded as a function

of “x is human.” The derived extensional function is: “There is a func-

Chap. XVII. Classes 1±o

tion formally equivalent to ‘x is human’ and such that I believe that

whatever satisﬁes it is mortal.” This remains true when we substitute

“x is a rational animal” | 188 for “x is human,” even if I believe falsely that

the Phœnix is rational and immortal.

We give the name of “derived extensional function” to the function

constructed as above, namely, to the function: “There is a function

having the property f and formally equivalent to φx,” where the

original function was “the function φx has the property f.”

We may regard the derived extensional function as having for its

argument the class determined by the function φx, and as asserting

f of this class. This may be taken as the deﬁnition of a proposition

about a class. I.e. we may deﬁne:

To assert that “the class determined by the function φx has the

property f ” is to assert that φx satisﬁes the extensional function

derived from f.

This gives a meaning to any statement about a class which can

be made signiﬁcantly about a function; and it will be found that

technically it yields the results which are required in order to make a

theory symbolically satisfactory.

1

What we have said just now as regards the deﬁnition of classes

is suﬃcient to satisfy our ﬁrst four conditions. The way in which

it secures the third and fourth, namely, the possibility of classes of

classes, and the impossibility of a class being or not being a member

of itself, is somewhat technical; it is explained in Principia Mathe-

matica, but may be taken for granted here. It results that, but for

our ﬁfth condition, we might regard our task as completed. But this

condition—at once the most important and the most diﬃcult—is not

fulﬁlled in virtue of anything we have said as yet. The diﬃculty is

connected with the theory of types, and must be brieﬂy discussed.

z

We saw in Chapter XIII. that there is a hierarchy of logical types,

and that it is a fallacy to allow an object belonging to one of these

to be substituted for an object belonging to another. | 18o Now it is not

diﬃcult to show that the various functions which can take a given

object a as argument are not all of one type. Let us call them all a-

functions. We may take ﬁrst those among them which do not involve

reference to any collection of functions; these we will call “predica-

tive a-functions.” If we now proceed to functions involving reference

to the totality of predicative a-functions, we shall incur a fallacy if we

regard these as of the same type as the predicative a-functions. Take

such an every-day statement as “a is a typical Frenchman.” How shall

we deﬁne a “typical Frenchman”? We may deﬁne him as one “possess-

ing all qualities that are possessed by most Frenchmen.” But unless

1

See Principia Mathematica, vol. i. pp. ;¸–8± and ∗zo.

z

The reader who desires a fuller discussion should consult Principia Mathematica,

Introduction, chap. ii.; also ∗1z.

Chap. XVII. Classes 1±1

we conﬁne “all qualities” to such as do not involve a reference to any

totality of qualities, we shall have to observe that most Frenchmen

are not typical in the above sense, and therefore the deﬁnition shows

that to be not typical is essential to a typical Frenchman. This is not

a logical contradiction, since there is no reason why there should be

any typical Frenchmen; but it illustrates the need for separating oﬀ

qualities that involve reference to a totality of qualities from those

that do not.

Whenever, by statements about “all” or “some” of the values that

a variable can signiﬁcantly take, we generate a new object, this new

object must not be among the values which our previous variable

could take, since, if it were, the totality of values over which the vari-

able could range would only be deﬁnable in terms of itself, and we

should be involved in a vicious circle. For example, if I say “Napoleon

had all the qualities that make a great general,” I must deﬁne “quali-

ties” in such a way that it will not include what I am now saying, i.e.

“having all the qualities that make a great general” must not be itself

a quality in the sense supposed. This is fairly obvious, and is the

principle which leads to the theory of types by which vicious-circle

paradoxes are avoided. As applied to a-functions, we may suppose

that “qualities” is to mean “predicative functions.” Then when I say

“Napoleon had all the qualities, etc.,” I mean | 1oo “Napoleon satisﬁed all

the predicative functions, etc.” This statement attributes a property

to Napoleon, but not a predicative property; thus we escape the vi-

cious circle. But wherever “all functions which” occurs, the functions

in question must be limited to one type if a vicious circle is to be

avoided; and, as Napoleon and the typical Frenchman have shown,

the type is not rendered determinate by that of the argument. It

would require a much fuller discussion to set forth this point fully,

but what has been said may suﬃce to make it clear that the functions

which can take a given argument are of an inﬁnite series of types. We

could, by various technical devices, construct a variable which would

run through the ﬁrst n of these types, where n is ﬁnite, but we cannot

construct a variable which will run through them all, and, if we could,

that mere fact would at once generate a new type of function with

the same arguments, and would set the whole process going again.

We call predicative a-functions the ﬁrst type of a-functions; a-

functions involving reference to the totality of the ﬁrst type we call

the second type; and so on. No variable a-function can run through

all these diﬀerent types: it must stop short at some deﬁnite one.

These considerations are relevant to our deﬁnition of the derived

extensional function. We there spoke of “a function formally equiva-

lent to φx.” It is necessary to decide upon the type of our function.

Any decision will do, but some decision is unavoidable. Let us call

the supposed formally equivalent function ψ. Then ψ appears as a

Chap. XVII. Classes 1±z

variable, and must be of some determinate type. All that we know

necessarily about the type of φ is that it takes arguments of a given

type—that it is (say) an a-function. But this, as we have just seen,

does not determine its type. If we are to be able (as our ﬁfth requisite

demands) to deal with all classes whose members are of the same type

as a, we must be able to deﬁne all such classes by means of functions

of some one type; that is to say, there must be some type of a-function,

say the n

th

, such that any a-function is formally | 1o1 equivalent to some

a-function of the n

th

type. If this is the case, then any extensional

function which holds of all a-functions of the n

th

type will hold of

any a-function whatever. It is chieﬂy as a technical means of embody-

ing an assumption leading to this result that classes are useful. The

assumption is called the “axiom of reducibility,” and may be stated

as follows:—

“There is a type (τ say) of a-functions such that, given any a-

function, it is formally equivalent to some function of the type in

question.”

If this axiom is assumed, we use functions of this type in deﬁning

our associated extensional function. Statements about all a-classes (i.e.

all classes deﬁned by a-functions) can be reduced to statements about

all a-functions of the type τ. So long as only extensional functions of

functions are involved, this gives us in practice results which would

otherwise have required the impossible notion of “all a-functions.”

One particular region where this is vital is mathematical induction.

The axiom of reducibility involves all that is really essential in the

theory of classes. It is therefore worth while to ask whether there is

any reason to suppose it true.

This axiom, like the multiplicative axiom and the axiom of inﬁn-

ity, is necessary for certain results, but not for the bare existence of

deductive reasoning. The theory of deduction, as explained in Chap-

ter XIV., and the laws for propositions involving “all” and “some,”

are of the very texture of mathematical reasoning: without them,

or something like them, we should not merely not obtain the same

results, but we should not obtain any results at all. We cannot use

them as hypotheses, and deduce hypothetical consequences, for they

are rules of deduction as well as premisses. They must be absolutely

true, or else what we deduce according to them does not even follow

from the premisses. On the other hand, the axiom of reducibility, like

our two previous mathematical axioms, could perfectly well be stated

as an hypothesis whenever it is used, instead of being assumed to be

actually true. We can deduce | 1oz its consequences hypothetically; we

can also deduce the consequences of supposing it false. It is therefore

only convenient, not necessary. And in view of the complication

of the theory of types, and of the uncertainty of all except its most

general principles, it is impossible as yet to say whether there may

Chap. XVII. Classes 1±j

not be some way of dispensing with the axiom of reducibility alto-

gether. However, assuming the correctness of the theory outlined

above, what can we say as to the truth or falsehood of the axiom?

The axiom, we may observe, is a generalised form of Leibniz’s

identity of indiscernibles. Leibniz assumed, as a logical principle, that

two diﬀerent subjects must diﬀer as to predicates. Now predicates are

only some among what we called “predicative functions,” which will

include also relations to given terms, and various properties not to be

reckoned as predicates. Thus Leibniz’s assumption is a much stricter

and narrower one than ours. (Not, of course, according to his logic,

which regarded all propositions as reducible to the subject-predicate

form.) But there is no good reason for believing his form, so far as

I can see. There might quite well, as a matter of abstract logical

possibility, be two things which had exactly the same predicates, in

the narrow sense in which we have been using the word “predicate.”

How does our axiom look when we pass beyond predicates in this

narrow sense? In the actual world there seems no way of doubting

its empirical truth as regards particulars, owing to spatio-temporal

diﬀerentiation: no two particulars have exactly the same spatial and

temporal relations to all other particulars. But this is, as it were, an

accident, a fact about the world in which we happen to ﬁnd ourselves.

Pure logic, and pure mathematics (which is the same thing), aims

at being true, in Leibnizian phraseology, in all possible worlds, not

only in this higgledy-piggledy job-lot of a world in which chance

has imprisoned us. There is a certain lordliness which the logician

should preserve: he must not condescend to derive arguments from

the things he sees about him. |

Viewed 1oj from this strictly logical point of view, I do not see any

reason to believe that the axiom of reducibility is logically necessary,

which is what would be meant by saying that it is true in all possible

worlds. The admission of this axiom into a system of logic is therefore

a defect, even if the axiom is empirically true. It is for this reason

that the theory of classes cannot be regarded as being as complete

as the theory of descriptions. There is need of further work on the

theory of types, in the hope of arriving at a doctrine of classes which

does not require such a dubious assumption. But it is reasonable to

regard the theory outlined in the present chapter as right in its main

lines, i.e. in its reduction of propositions nominally about classes to

propositions about their deﬁning functions. The avoidance of classes

as entities by this method must, it would seem, be sound in principle,

however the detail may still require adjustment. It is because this

seems indubitable that we have included the theory of classes, in

spite of our desire to exclude, as far as possible, whatever seemed

open to serious doubt.

The theory of classes, as above outlined, reduces itself to one

Chap. XVII. Classes 1±±

axiom and one deﬁnition. For the sake of deﬁniteness, we will here

repeat them. The axiom is:

There is a type τ such that if φ is a function which can take a given

object a as argument, then there is a function ψ of the type τ which is

formally equivalent to φ.

The deﬁnition is:

If φ is a function which can take a given object a as argument, and

τ the type mentioned in the above axiom, then to say that the class

determined by φ has the property f is to say that there is a function of type

τ, formally equivalent to φ, and having the property f.

CHAPTER XVIII

MATHEMATICS ANDLOGIC

1o± Mathematics and logic, historically speaking, have been entirely

distinct studies. Mathematics has been connected with science, logic

with Greek. But both have developed in modern times: logic has

become more mathematical and mathematics has become more log-

ical. The consequence is that it has now become wholly impossible

to draw a line between the two; in fact, the two are one. They diﬀer

as boy and man: logic is the youth of mathematics and mathemat-

ics is the manhood of logic. This view is resented by logicians who,

having spent their time in the study of classical texts, are incapable

of following a piece of symbolic reasoning, and by mathematicians

who have learnt a technique without troubling to inquire into its

meaning or justiﬁcation. Both types are now fortunately growing

rarer. So much of modern mathematical work is obviously on the

border-line of logic, so much of modern logic is symbolic and formal,

that the very close relationship of logic and mathematics has become

obvious to every instructed student. The proof of their identity is, of

course, a matter of detail: starting with premisses which would be

universally admitted to belong to logic, and arriving by deduction

at results which as obviously belong to mathematics, we ﬁnd that

there is no point at which a sharp line can be drawn, with logic to

the left and mathematics to the right. If there are still those who do

not admit the identity of logic and mathematics, we may challenge

them to indicate at what point, in the successive deﬁnitions and |

1o¸ deductions of Principia Mathematica, they consider that logic ends

and mathematics begins. It will then be obvious that any answer

must be quite arbitrary.

In the earlier chapters of this book, starting from the natural

numbers, we have ﬁrst deﬁned “cardinal number” and shown how

to generalise the conception of number, and have then analysed

the conceptions involved in the deﬁnition, until we found ourselves

dealing with the fundamentals of logic. In a synthetic, deductive

treatment these fundamentals come ﬁrst, and the natural numbers are

only reached after a long journey. Such treatment, though formally

1±¸

Chap. XVIII. Mathematics and Logic 1±6

more correct than that which we have adopted, is more diﬃcult for

the reader, because the ultimate logical concepts and propositions

with which it starts are remote and unfamiliar as compared with

the natural numbers. Also they represent the present frontier of

knowledge, beyond which is the still unknown; and the dominion of

knowledge over them is not as yet very secure.

It used to be said that mathematics is the science of “quantity.”

“Quantity” is a vague word, but for the sake of argument we may re-

place it by the word “number.” The statement that mathematics is the

science of number would be untrue in two diﬀerent ways. On the one

hand, there are recognised branches of mathematics which have noth-

ing to do with number—all geometry that does not use co-ordinates

or measurement, for example: projective and descriptive geometry,

down to the point at which co-ordinates are introduced, does not have

to do with number, or even with quantity in the sense of greater and

less. On the other hand, through the deﬁnition of cardinals, through

the theory of induction and ancestral relations, through the general

theory of series, and through the deﬁnitions of the arithmetical op-

erations, it has become possible to generalise much that used to be

proved only in connection with numbers. The result is that what

was formerly the single study of Arithmetic has now become divided

into a number of separate studies, no one of which is specially con-

cerned with numbers. The most | 1o6 elementary properties of numbers

are concerned with one-one relations, and similarity between classes.

Addition is concerned with the construction of mutually exclusive

classes respectively similar to a set of classes which are not known

to be mutually exclusive. Multiplication is merged in the theory of

“selections,” i.e. of a certain kind of one-many relations. Finitude is

merged in the general study of ancestral relations, which yields the

whole theory of mathematical induction. The ordinal properties of

the various kinds of number-series, and the elements of the theory of

continuity of functions and the limits of functions, can be generalised

so as no longer to involve any essential reference to numbers. It is a

principle, in all formal reasoning, to generalise to the utmost, since

we thereby secure that a given process of deduction shall have more

widely applicable results; we are, therefore, in thus generalising the

reasoning of arithmetic, merely following a precept which is univer-

sally admitted in mathematics. And in thus generalising we have, in

eﬀect, created a set of new deductive systems, in which traditional

arithmetic is at once dissolved and enlarged; but whether any one of

these new deductive systems—for example, the theory of selections—

is to be said to belong to logic or to arithmetic is entirely arbitrary,

and incapable of being decided rationally.

We are thus brought face to face with the question: What is this

subject, which may be called indiﬀerently either mathematics or

logic? Is there any way in which we can deﬁne it?

Chap. XVIII. Mathematics and Logic 1±;

Certain characteristics of the subject are clear. To begin with, we

do not, in this subject, deal with particular things or particular prop-

erties: we deal formally with what can be said about any thing or any

property. We are prepared to say that one and one are two, but not

that Socrates and Plato are two, because, in our capacity of logicians

or pure mathematicians, we have never heard of Socrates and Plato. A

world in which there were no such individuals would still be a world

in which one and one are two. It is not open to us, as pure mathemati-

cians or logicians, to mention anything at all, because, if we do so,

| 1o; we introduce something irrelevant and not formal. We may make

this clear by applying it to the case of the syllogism. Traditional logic

says: “All men are mortal, Socrates is a man, therefore Socrates is

mortal.” Now it is clear that what we mean to assert, to begin with, is

only that the premisses imply the conclusion, not that premisses and

conclusion are actually true; even the most traditional logic points

out that the actual truth of the premisses is irrelevant to logic. Thus

the ﬁrst change to be made in the above traditional syllogism is to

state it in the form: “If all men are mortal and Socrates is a man, then

Socrates is mortal.” We may now observe that it is intended to convey

that this argument is valid in virtue of its form, not in virtue of the

particular terms occurring in it. If we had omitted “Socrates is a man”

from our premisses, we should have had a non-formal argument, only

admissible because Socrates is in fact a man; in that case we could not

have generalised the argument. But when, as above, the argument is

formal, nothing depends upon the terms that occur in it. Thus we may

substitute α for men, β for mortals, and x for Socrates, where α and β

are any classes whatever, and x is any individual. We then arrive at

the statement: “No matter what possible values x and α and β may

have, if all α’s are β’s and x is an α, then x is a β”; in other words, “the

propositional function ‘if all α’s are β’s and x is an α, then x is a β’

is always true.” Here at last we have a proposition of logic—the one

which is only suggested by the traditional statement about Socrates

and men and mortals.

It is clear that, if formal reasoning is what we are aiming at,

we shall always arrive ultimately at statements like the above, in

which no actual things or properties are mentioned; this will happen

through the mere desire not to waste our time proving in a particu-

lar case what can be proved generally. It would be ridiculous to go

through a long argument about Socrates, and then go through pre-

cisely the same argument again about Plato. If our argument is one

(say) which holds of all men, we shall prove it concerning “x,” with

the hypothesis “if x is a man.” With | 1o8 this hypothesis, the argument

will retain its hypothetical validity even when x is not a man. But

now we shall ﬁnd that our argument would still be valid if, instead

of supposing x to be a man, we were to suppose him to be a monkey

Chap. XVIII. Mathematics and Logic 1±8

or a goose or a Prime Minister. We shall therefore not waste our time

taking as our premiss “x is a man” but shall take “x is an α,” where

α is any class of individuals, or “φx” where φ is any propositional

function of some assigned type. Thus the absence of all mention

of particular things or properties in logic or pure mathematics is

a necessary result of the fact that this study is, as we say, “purely

formal.”

At this point we ﬁnd ourselves faced with a problem which is eas-

ier to state than to solve. The problem is: “What are the constituents

of a logical proposition?” I do not know the answer, but I propose to

explain how the problem arises.

Take (say) the proposition “Socrates was before Aristotle.” Here it

seems obvious that we have a relation between two terms, and that

the constituents of the proposition (as well as of the corresponding

fact) are simply the two terms and the relation, i.e. Socrates, Aristotle,

and before. (I ignore the fact that Socrates and Aristotle are not simple;

also the fact that what appear to be their names are really truncated

descriptions. Neither of these facts is relevant to the present issue.)

We may represent the general form of such propositions by “xRy,”

which may be read “x has the relation R to y.” This general form

may occur in logical propositions, but no particular instance of it can

occur. Are we to infer that the general form itself is a constituent of

such logical propositions?

Given a proposition, such as “Socrates is before Aristotle,” we

have certain constituents and also a certain form. But the form is

not itself a new constituent; if it were, we should need a new form

to embrace both it and the other constituents. We can, in fact, turn

all the constituents of a proposition into variables, while keeping the

form unchanged. This is what we do when we use such a schema as

“xRy,” which stands for any | 1oo one of a certain class of propositions,

namely, those asserting relations between two terms. We can proceed

to general assertions, such as “xRy is sometimes true”—i.e. there

are cases where dual relations hold. This assertion will belong to

logic (or mathematics) in the sense in which we are using the word.

But in this assertion we do not mention any particular things or

particular relations; no particular things or relations can ever enter

into a proposition of pure logic. We are left with pure forms as the

only possible constituents of logical propositions.

I do not wish to assert positively that pure forms—e.g. the form

“xRy”—do actually enter into propositions of the kind we are consid-

ering. The question of the analysis of such propositions is a diﬃcult

one, with conﬂicting considerations on the one side and on the other.

We cannot embark upon this question now, but we may accept, as a

ﬁrst approximation, the view that forms are what enter into logical

propositions as their constituents. And we may explain (though not

Chap. XVIII. Mathematics and Logic 1±o

formally deﬁne) what we mean by the “form” of a proposition as

follows:—

The “form” of a proposition is that, in it, that remains unchanged

when every constituent of the proposition is replaced by another.

Thus “Socrates is earlier than Aristotle” has the same form as

“Napoleon is greater than Wellington,” though every constituent of

the two propositions is diﬀerent.

We may thus lay down, as a necessary (though not suﬃcient) char-

acteristic of logical or mathematical propositions, that they are to

be such as can be obtained from a proposition containing no vari-

ables (i.e. no such words as all, some, a, the, etc.) by turning every

constituent into a variable and asserting that the result is always true

or sometimes true, or that it is always true in respect of some of the

variables that the result is sometimes true in respect of the others,

or any variant of these forms. And another way of stating the same

thing is to say that logic (or mathematics) is concerned only with

forms, and is concerned with them only in the way of stating that they

are always or | zoo sometimes true—with all the permutations of “always”

and “sometimes” that may occur.

There are in every language some words whose sole function is

to indicate form. These words, broadly speaking, are commonest

in languages having fewest inﬂections. Take “Socrates is human.”

Here “is” is not a constituent of the proposition, but merely indicates

the subject-predicate form. Similarly in “Socrates is earlier than

Aristotle,” “is” and “than” merely indicate form; the proposition is

the same as “Socrates precedes Aristotle,” in which these words have

disappeared and the form is otherwise indicated. Form, as a rule,

can be indicated otherwise than by speciﬁc words: the order of the

words can do most of what is wanted. But this principle must not be

pressed. For example, it is diﬃcult to see how we could conveniently

express molecular forms of propositions (i.e. what we call “truth-

functions”) without any word at all. We saw in Chapter XIV. that

one word or symbol is enough for this purpose, namely, a word or

symbol expressing incompatibility. But without even one we should

ﬁnd ourselves in diﬃculties. This, however, is not the point that is

important for our present purpose. What is important for us is to

observe that form may be the one concern of a general proposition,

even when no word or symbol in that proposition designates the form.

If we wish to speak about the form itself, we must have a word for

it; but if, as in mathematics, we wish to speak about all propositions

that have the form, a word for the form will usually be found not

indispensable; probably in theory it is never indispensable.

Assuming—as I think we may—that the forms of propositions can

be represented by the forms of the propositions in which they are

expressed without any special words for forms, we should arrive at a

Chap. XVIII. Mathematics and Logic 1¸o

language in which everything formal belonged to syntax and not to

vocabulary. In such a language we could express all the propositions

of mathematics even if we did not know one single word of the

language. The language of | zo1 mathematical logic, if it were perfected,

would be such a language. We should have symbols for variables,

such as “x” and “R” and “y,” arranged in various ways; and the way

of arrangement would indicate that something was being said to be

true of all values or some values of the variables. We should not

need to know any words, because they would only be needed for

giving values to the variables, which is the business of the applied

mathematician, not of the pure mathematician or logician. It is one

of the marks of a proposition of logic that, given a suitable language,

such a proposition can be asserted in such a language by a person who

knows the syntax without knowing a single word of the vocabulary.

But, after all, there are words that express form, such as “is” and

“than.” And in every symbolism hitherto invented for mathematical

logic there are symbols having constant formal meanings. We may

take as an example the symbol for incompatibility which is employed

in building up truth-functions. Such words or symbols may occur in

logic. The question is: How are we to deﬁne them?

Such words or symbols express what are called “logical constants.”

Logical constants may be deﬁned exactly as we deﬁned forms; in fact,

they are in essence the same thing. A fundamental logical constant

will be that which is in common among a number of propositions,

any one of which can result from any other by substitution of terms

one for another. For example, “Napoleon is greater than Wellington”

results from “Socrates is earlier than Aristotle” by the substitution of

“Napoleon” for “Socrates,” “Wellington” for “Aristotle,” and “greater”

for “earlier.” Some propositions can be obtained in this way from the

prototype “Socrates is earlier than Aristotle” and some cannot; those

that can are those that are of the form “xRy,” i.e. express dual rela-

tions. We cannot obtain from the above prototype by term-for-term

substitution such propositions as “Socrates is human” or “the Atheni-

ans gave the hemlock to Socrates,” because the ﬁrst is of the subject- |

zoz predicate form and the second expresses a three-term relation. If we

are to have any words in our pure logical language, they must be such

as express “logical constants,” and “logical constants” will always

either be, or be derived from, what is in common among a group

of propositions derivable from each other, in the above manner, by

term-for-term substitution. And this which is in common is what we

call “form.”

In this sense all the “constants” that occur in pure mathematics

are logical constants. The number 1, for example, is derivative from

propositions of the form: “There is a term c such that φx is true when,

and only when, x is c.” This is a function of φ, and various diﬀerent

Chap. XVIII. Mathematics and Logic 1¸1

propositions result from giving diﬀerent values to φ. We may (with

a little omission of intermediate steps not relevant to our present

purpose) take the above function of φ as what is meant by “the class

determined by φ is a unit class” or “the class determined by φ is a

member of 1” (1 being a class of classes). In this way, propositions

in which 1 occurs acquire a meaning which is derived from a certain

constant logical form. And the same will be found to be the case

with all mathematical constants: all are logical constants, or symbolic

abbreviations whose full use in a proper context is deﬁned by means

of logical constants.

But although all logical (or mathematical) propositions can be

expressed wholly in terms of logical constants together with vari-

ables, it is not the case that, conversely, all propositions that can be

expressed in this way are logical. We have found so far a necessary

but not a suﬃcient criterion of mathematical propositions. We have

suﬃciently deﬁned the character of the primitive ideas in terms of

which all the ideas of mathematics can be deﬁned, but not of the

primitive propositions from which all the propositions of mathematics

can be deduced. This is a more diﬃcult matter, as to which it is not

yet known what the full answer is.

We may take the axiom of inﬁnity as an example of a proposi-

tion which, though it can be enunciated in logical terms, | zoj cannot

be asserted by logic to be true. All the propositions of logic have a

characteristic which used to be expressed by saying that they were

analytic, or that their contradictories were self-contradictory. This

mode of statement, however, is not satisfactory. The law of contra-

diction is merely one among logical propositions; it has no special

pre-eminence; and the proof that the contradictory of some propo-

sition is self-contradictory is likely to require other principles of

deduction besides the law of contradiction. Nevertheless, the char-

acteristic of logical propositions that we are in search of is the one

which was felt, and intended to be deﬁned, by those who said that

it consisted in deducibility from the law of contradiction. This char-

acteristic, which, for the moment, we may call tautology, obviously

does not belong to the assertion that the number of individuals in

the universe is n, whatever number n may be. But for the diversity of

types, it would be possible to prove logically that there are classes of

n terms, where n is any ﬁnite integer; or even that there are classes

of ℵ

o

terms. But, owing to types, such proofs, as we saw in Chapter

XIII., are fallacious. We are left to empirical observation to determine

whether there are as many as n individuals in the world. Among

“possible” worlds, in the Leibnizian sense, there will be worlds having

one, two, three, . . . individuals. There does not even seem any logical

Chap. XVIII. Mathematics and Logic 1¸z

necessity why there should be even one individual

1

—why, in fact,

there should be any world at all. The ontological proof of the exis-

tence of God, if it were valid, would establish the logical necessity of

at least one individual. But it is generally recognised as invalid, and

in fact rests upon a mistaken view of existence—i.e. it fails to realise

that existence can only be asserted of something described, not of

something named, so that it is meaningless to argue from “this is the

so-and-so” and “the so-and-so exists” to “this exists.” If we reject the

ontological | zo± argument, we seem driven to conclude that the existence

of a world is an accident—i.e. it is not logically necessary. If that be so,

no principle of logic can assert “existence” except under a hypothesis,

i.e. none can be of the form “the propositional function so-and-so is

sometimes true.” Propositions of this form, when they occur in logic,

will have to occur as hypotheses or consequences of hypotheses, not as

complete asserted propositions. The complete asserted propositions

of logic will all be such as aﬃrm that some propositional function

is always true. For example, it is always true that if p implies q and

q implies r then p implies r, or that, if all α’s are β’s and x is an α

then x is a β. Such propositions may occur in logic, and their truth

is independent of the existence of the universe. We may lay it down

that, if there were no universe, all general propositions would be true;

for the contradictory of a general proposition (as we saw in Chapter

XV.) is a proposition asserting existence, and would therefore always

be false if no universe existed.

Logical propositions are such as can be known a priori, without

study of the actual world. We only know from a study of empirical

facts that Socrates is a man, but we know the correctness of the syllo-

gism in its abstract form (i.e. when it is stated in terms of variables)

without needing any appeal to experience. This is a characteristic,

not of logical propositions in themselves, but of the way in which we

know them. It has, however, a bearing upon the question what their

nature may be, since there are some kinds of propositions which it

would be very diﬃcult to suppose we could know without experi-

ence.

It is clear that the deﬁnition of “logic” or “mathematics” must be

sought by trying to give a newdeﬁnition of the old notion of “analytic”

propositions. Although we can no longer be satisﬁed to deﬁne logical

propositions as those that follow from the law of contradiction, we

can and must still admit that they are a wholly diﬀerent class of

propositions from those that we come to know empirically. They

all have the characteristic which, a moment ago, we agreed to call

“tautology.” This, | zo¸ combined with the fact that they can be expressed

1

The primitive propositions in Principia Mathematica are such as to allow the

inference that at least one individual exists. But I now view this as a defect in

logical purity.

Chap. XVIII. Mathematics and Logic 1¸j

wholly in terms of variables and logical constants (a logical constant

being something which remains constant in a proposition even when

all its constituents are changed)—will give the deﬁnition of logic or

pure mathematics. For the moment, I do not know how to deﬁne

“tautology.”

z

It would be easy to oﬀer a deﬁnition which might seem

satisfactory for a while; but I knowof none that I feel to be satisfactory,

in spite of feeling thoroughly familiar with the characteristic of which

a deﬁnition is wanted. At this point, therefore, for the moment, we

reach the frontier of knowledge on our backward journey into the

logical foundations of mathematics.

We have now come to an end of our somewhat summary intro-

duction to mathematical philosophy. It is impossible to convey ad-

equately the ideas that are concerned in this subject so long as we

abstain from the use of logical symbols. Since ordinary language

has no words that naturally express exactly what we wish to express,

it is necessary, so long as we adhere to ordinary language, to strain

words into unusual meanings; and the reader is sure, after a time if

not at ﬁrst, to lapse into attaching the usual meanings to words, thus

arriving at wrong notions as to what is intended to be said. Moreover,

ordinary grammar and syntax is extraordinarily misleading. This is

the case, e.g., as regards numbers; “ten men” is grammatically the

same form as “white men,” so that 1o might be thought to be an ad-

jective qualifying “men.” It is the case, again, wherever propositional

functions are involved, and in particular as regards existence and

descriptions. Because language is misleading, as well as because it

is diﬀuse and inexact when applied to logic (for which it was never

intended), logical symbolism is absolutely necessary to any exact or

thorough treatment of our subject. Those readers, | zo6 therefore, who

wish to acquire a mastery of the principles of mathematics, will, it is

to be hoped, not shrink from the labour of mastering the symbols—a

labour which is, in fact, much less than might be thought. As the

above hasty survey must have made evident, there are innumerable

unsolved problems in the subject, and much work needs to be done.

If any student is led into a serious study of mathematical logic by this

little book, it will have served the chief purpose for which it has been

written.

z

The importance of “tautology” for a deﬁnition of mathematics was pointed out

to me by my former pupil Ludwig Wittgenstein, who was working on the problem.

I do not know whether he has solved it, or even whether he is alive or dead.

INDEX

zo; [Online edition note: This is a hyperlinked recreation of the original

index. The page numbers listed are for the original edition, i.e., those

marked in the margins of this edition.]

Aggregates, 1z.

Alephs, 8j, oz, o;, 1z¸.

Aliorelatives, jz.

All, 1¸8ﬀ.

Analysis, ±.

Ancestors, z¸, jj.

Argument of a function, ±;,

1o8.

Arithmetising of mathematics,

±.

Associative law, ¸8, o±.

Axioms, 1.

Between, j8ﬀ., ¸8.

Bolzano, 1j8n.

Boots and socks, 1z6.

Boundary, ;o, o8, oo.

Cantor, Georg, ;;, ;o, 8¸n., 86,

8o, o¸, 1oz, 1j6.

Classes, 1z, 1j;, 181ﬀ.; reﬂexive,

8o, 1z;, 1j8; similar, 1¸, 16.

Cliﬀord, W. K., ;6.

Collections, inﬁnite, 1j.

Commutative law, ¸8, o±.

Conjunction, 1±;.

Consecutiveness, j;, j8, 81.

Constants, zoz.

Construction, method of, ;j.

Continuity, 86, o;ﬀ.; Cantorian,

1ozﬀ.; Dedekindian, 1o1; in

philosophy, 1o¸; of functions,

1o6ﬀ.

Contradictions, 1j¸ﬀ.

Convergence, 11¸.

Converse, 16, jz, ±o.

Correlators, ¸±.

Counterparts, objective, 61.

Counting, 1±, 16.

Dedekind, 6o, oo, 1j8n.

Deduction, 1±±ﬀ.

Deﬁnition, j; extensional and

intensional, 1z.

Derivatives, 1oo.

Descriptions, 1jo, 1±±, 16;ﬀ.

Dimensions, zo.

Disjunction, 1±;.

Distributive law, ¸8, o±.

Diversity, 8;.

Domain, 16, jz, ±o.

Equivalence, 18j.

Euclid, 6;.

Existence, 16±, 1;1, 1;;.

Exponentiation, o±, 1zo.

Extension of a relation, 6o.

Fictions, logical, 1±n., ±¸, 1j;.

1¸±

Index 1¸¸

Field of a relation, jz, ¸j.

Finite, z;.

Flux, 1o¸.

Form, 1o8.

Fractions, j;, 6±.

Frege, ;, 1o [11], z¸n., ;;, o¸,

1±6n.

Functions, ±6; descriptive, ±6,

18o; intensional and

extensional, 186; predicative,

18o; propositional, ±6, 1±±,

1¸¸ﬀ.

Gap, Dedekindian, ;oﬀ., oo.

Generalisation, 1¸6.

Geometry, zo, ¸o, 6;, ;±, 1oo,

1±¸; analytical, ±, 86.

Greater and less, 6¸, oo.

Hegel, 1o;.

Hereditary properties, z1.

Implication, 1±6, 1¸j; formal,

16j.

Incommensurables, ±, 66.

Incompatibility, 1±;ﬀ., zoo.

Incomplete symbols, 18z.

Indiscernibles, 1oz.

Individuals, 1jz, 1±1, 1;j.

Induction, mathematical, zoﬀ.,

8;, oj, 18¸.

Inductive properties, z1.

Inference, 1±8ﬀ.

Inﬁnite, z8; of rationals, 6¸;

Cantorian, 6¸; of cardinals,

;;ﬀ.; and series and ordinals,

8oﬀ.

Inﬁnity, axiom of, 66n., ;;,

1j1ﬀ., zoz.

Instances, 1¸6.

Integers, positive and negative,

6±.

Intervals, 11¸.

Intuition, 1±¸.

Irrationals, 66, ;z. |

Kant, zo8 1±¸.

Leibniz, 8o, 1o;, 1oz.

Lewis, C. I., 1¸j, 1¸±.

Likeness, ¸z.

Limit, zo, 6oﬀ., o;ﬀ.; of

functions, 1o6ﬀ.

Limiting points, oo.

Logic, 1¸o, 16o, 1o±ﬀ.;

mathematical, v, zo1, zo6.

Logicising of mathematics, ;.

Maps, ¸z, 6oﬀ., 8o.

Mathematics, 1o±ﬀ.

Maximum, ;o, o8.

Median class, 1o±.

Meinong, 16o.

Method, vi.

Minimum, ;o, o8.

Modality, 16¸.

Multiplication, 118ﬀ.

Multiplicative axiom, oz, 11;ﬀ.

Names, 1;j, 18z.

Necessity, 16¸.

Neighbourhood, 1oo.

Nicod, 1±8, 1±o, 1¸1n.

Null-class, zj, 1jz.

Number, cardinal, 1oﬀ., ¸6, ;;ﬀ.,

o¸; complex, ;±ﬀ.; ﬁnite,

zoﬀ.; inductive, z;, ;8, 1j1;

inﬁnite, ;;ﬀ.; irrational, 66,

;z; maximum ? 1j¸;

multipliable, 1jo; natural,

zﬀ., zz; non-inductive, 88,

1z;; real, 66, ;z, 8±; reﬂexive,

8o, 1z;; relation, ¸6, o±;

serial, ¸;.

Occam, 18±.

Occurrences, primary and

secondary, 1;o.

Ontological proof, zoj.

Order, zoﬀ.; cyclic, ±o.

Oscillation, ultimate, 111.

Index 1¸6

Parmenides, 1j8.

Particulars, 1±oﬀ., 1;j.

Peano, ¸ﬀ., zj, z±, ;8, 81, 1j1,

16j.

Peirce, jzn.

Permutations, ¸o.

Philosophy, mathematical, v, 1.

Plato, 1j8.

Plurality, 1o.

Poincar´ e, z;.

Points, ¸o.

Posterity, zzﬀ.; proper, j6.

Postulates, ;1, ;j.

Precedent, o8.

Premisses of arithmetic, ¸.

Primitive ideas and propositions,

¸, zoz.

Progressions, 8, 81ﬀ.

Propositions, 1¸¸; analytic, zo±;

elementary, 161.

Pythagoras, ±, 6;.

Quantity, o;, 1o¸.

Ratios, 6±, ;1, 8±, 1jj.

Reducibility, axiom of, 1o1.

Referent, ±8.

Relation-numbers, ¸6ﬀ.

Relations, asymmetrical, j1, ±z;

connected, jz; many-one, 1¸;

one-many, 1¸, ±¸; one-one,

1¸, ±;, ;o; reﬂexive, 16; serial,

j±; similar, ¸zﬀ.; squares of,

jz; symmetrical, 16, ±±;

transitive, 16, jz.

Relatum, ±8.

Representatives, 1zo.

Rigour, 1±±.

Royce, 8o.

Section, Dedekindian, 6oﬀ.;

ultimate, 111.

Segments, ;z, o8.

Selections, 11;ﬀ.

Sequent, o8.

Series, zoﬀ.; closed, 1oj;

compact, 66, oj, 1oo;

condensed in itself, 1oz;

Dedekindian, ;1, ;j, 1o1;

generation of, ±1; inﬁnite,

8oﬀ.; perfect, 1oz, 1oj;

well-ordered, oz, 1zj.

Sheﬀer, 1±8.

Similarity, of classes, 1¸ﬀ.; of

relations, ¸zﬀ., 8j.

Some, 1¸8ﬀ.

Space, 61, 86, 1±o.

Structure, 6oﬀ.

Sub-classes, 8±ﬀ.

Subjects, 1±z.

Subtraction, 8;.

Successor of a number, zj, j¸.

Syllogism, 1o;.

Tautology, zoj, zo¸.

The, 16;, 1;zﬀ.

Time, 61, 86, 1±o.

Truth-function, 1±;.

Truth-value, 1±6.

Types, logical, ¸j, 1j¸ﬀ., 18¸,

188.

Unreality, 168.

Value of a function, ±;, 1o8.

Variables, 1o, 161, 1oo.

Veblen, ¸8.

Verbs, 1±1.

Weierstrass, o;, 1o;.

Wells, H. G., 11±.

Whitehead, 6±, ;6, 1o;, 11o.

Wittgenstein, zo¸n.

Zermelo, 1zj, 1zo.

Zero, 6¸.

CHANGES TOONLINE EDITION

This Online Corrected Edition was created by Kevin C. Klement;

this is version 1.o (February ¸, zo1o). It is based on the April 1ozo

so-called “second edition” published by Allen & Unwin, which, by

contemporary standards, was simply a second printing of the original

1o1o edition but incorporating various, mostly minor, ﬁxes. This

edition incorporates ﬁxes from later printings as well, and some new

ﬁxes, mentioned below. The pagination of the Allen & Unwin edition

is given in the margins, with page breaks marked with the sign “|”.

These are in red, as are other additions to the text not penned by

Russell.

Thanks to members of the Russell-l and HEAPS-l mailing lists

for help in checking and proofreading the version, including Adam

Killian, Pierre Grenon, David Blitz, Brandon Young, Rosalind Car-

ey, and, especially, John Ongley. A tremendous debt of thanks is

owed to Kenneth Blackwell of the Bertrand Russell Archives/Re-

search Centre, McMaster University, for proofreading the bulk of the

edition, checking it against Russell’s handwritten manuscript, and

providing other valuable advice and assistance. Another large debt

of gratitude is owed to Christof Gr¨ aber who compared this version

to the print versions and showed remarkable aptitude in spotting

discrepancies. I take full responsibility for any remaining errors. If

you discover any, please email me at klement@philos.umass.edu.

The online edition diﬀers from the 1ozo Allen & Unwin edition,

and reprintings thereof, in certain respects. Some are mere stylistic

diﬀerences. Others represent corrections based on discrepancies

between Russell’s manuscript and the print edition, or ﬁx small

grammatical or typographical errors. The stylistic diﬀerences are

these:

• In the original, footnote numbering begins anew with each page.

Since this version uses diﬀerent pagination, it was necessary to

number footnotes sequentially through each chapter. Thus, for

example, the footnote listed as note z on page ± of this edition

was listed as note 1 on page ¸ of the original.

1¸;

Changes to Online Edition 1¸8

• With some exceptions, the Allen & Unwin edition uses linear

fractions of the style “x/y” mid-paragraph, but vertical fractions

of the form “

x

y

” in displays. Contrary to this usual practice,

those in the display on page 8± of the original (page 6j of

this edition) were linear, but have been converted to vertical

fractions in this edition. Similarly, the mid-paragraph fractions

on pages 1;, z6, oo and 116 of the original (pages 1z, 1o, ;±

and 86 here) were printed vertically in the original, but here

are horizontal.

The following more signiﬁcant changes and revisions are marked

in green in this edition. Most of these result from Ken Blackwell’s

comparison with Russell’s manuscript. A fewwere originally noted in

an early review of the book by G. A. Pfeiﬀer (Bulletin of the American

Mathematical Society z;:z (1ozo), pp. 81–oo).

1. (page zn. / original page zn.) Russell wrote the wrong publica-

tion date (1o11) for the second volume of Principia Mathematica;

this has been ﬁxed to 1o1z.

z. (page 16 / original page z1) “. . . or all that are less than 1ooo . . . ”

is changed to “. . . or all that are not less than 1ooo . . . ” to match

Russell’s manuscript and the obviously intended meaning of the

passage. This error was noted by Pfeiﬀer in 1ozo but unﬁxed in

Russell’s lifetime.

j. (page jz / original page ±j) “. . . either by limiting the domain

to males or by limiting the converse to females” is changed to

“. . . either by limiting the domain to males or by limiting the

converse domain to females”, which is how it read in Russell’s

manuscript, and seems better to ﬁt the context.

±. (page ±8 / original page 6±) “. . . provided neither mor n is zero.”

is ﬁxed to “. . . provided neither m nor n is zero.” Thanks to John

Ongley for spotting this error, which exists even in Russell’s

manuscript.

¸. (page 6±n. / original page 8¸n.) The word “deutschen” in the

original’s (and the manuscript’s) “Jahresbericht der deutschen

Mathematiker-Vereinigung” has been capitalized.

6. (page ;± / original page o8) “. . . of a class α, i.e. its limits or

maximum, and then . . . ” is changed to “. . . of a class α, i.e. its

limit or maximum, and then . . . ” to match Russell’s manuscript,

and the apparent meaning of the passage.

;. (page 8z / original page 11o) “. . . the limit of its value for ap-

proaches either from . . . ” is changed to “. . . the limit of its

Changes to Online Edition 1¸o

values for approaches either from . . . ”, which matches Russell’s

manuscript, and is more appropriate for the meaning of the

passage.

8. (page 8¸ / original page 11j) The ungrammatical “. . . advan-

tages of this form of deﬁnition is that it analyses . . . ” is changed

to “. . . advantage of this form of deﬁnition is that it analyses

. . . ” to match Russell’s manuscript.

o. (page 8¸ / original page 11¸) “. . . all terms z such that x has

the relation P to x and z has the relation P to y . . . ” is ﬁxed to

“. . . all terms z such that x has the relation P to z and z has the

relation P to y . . . ” Russell himself hand-corrected this in his

manuscript, but not in a clear way, and at his request, it was

changed in the 1o6; printing.

1o. (page oj / original page 1z±) The words “correlator of α with

β, and similarly for every other pair. This requires a”, which

constitute exactly one line of Russell’s manuscript, were omit-

ted, thereby amalgamating two sentences into one. The missing

words are now restored.

11. (page o; / original page 1zo) The passage “. . . if x

1

is the mem-

ber of y

1

, x

z

is a member of y

z

, x

j

is a member of y

j

, and so on;

then . . . ” is changed to “. . . if x

1

is the member of γ

1

, x

z

is a

member of γ

z

, x

j

is a member of γ

j

, and so on; then . . . ” to

match Russell’s manuscript, and the obviously intended mean-

ing of the passage.

1z. (page 1o¸ / original page 1jo) The words “and then the idea

of the idea of Socrates” although present in Russell’s manu-

script, were left out of previous print editions. Note that Russell

mentions “all these ideas” in the next sentence.

1j. (page 1zo / original page 16o) The two footnotes on this page

were misplaced. The second, the reference to Principia Math-

ematica ∗o, was attached in previous versions to the sentence

that now refers to the ﬁrst footnote in the chapter. That foot-

note was placed three sentences below. The footnote references

have been returned to where they had been placed in Russell’s

manuscript.

1±. (page 1zo / original page 161) “. . . the negation of propositions

of the type to which x belongs . . . ” is changed to “. . . the

negation of propositions of the type to which φx belongs . . . ”

to match Russell’s manuscript. This is another error noted by

Pfeiﬀer.

Changes to Online Edition 16o

1¸. (page 1z1 / original page 16z) “Suppose we are considering all

“men are mortal”: we will . . . ” is changed to “Suppose we are

considering “all men are mortal”: we will . . . ” to match the ob-

viously intended meaning of the passage, and the placement of

the opening quotation mark in Russell’s manuscript (although

he here used single quotation marks, as he did sporadically

throughout). Thanks to Christof Gr¨ aber for spotting this error.

16. (page 1zo / original page 1;j) “. . . as opposed to speciﬁc man.”

is ﬁxed to “. . . as opposed to speciﬁc men.” Russell sent this

change to Unwin in 1oj;, and it was made in the 1oj8 printing.

1;. (page 1j1 / original page 1;¸) The “φ” in “. . . the process of

applying general statements about φx to particular cases . . . ”,

present in Russell’s manuscript, was excluded from the Allen &

Unwin printings, and has been restored.

18. (page 1jz / original page 1;6) The “φ” in “. . . resulting from

a propositional function φx by the substitution of . . . ” was

excluded from previous published versions, though it does

appear in Russell’s manuscript, and seems necessary for the

passage to make sense. Thanks to John Ongley for spotting this

error, which had also been noted by Pfeiﬀer.

1o. (page 1jo / original pages 186–8;) The two occurrences of “φ”

in “. . . extensional functions of a function φx may, for practical

purposes, be regarded as functions of the class determined

by φx, while intensional functions cannot . . . ” were omitted

from previous published versions, but do appear in Russell’s

manuscript. Again thanks to John Ongley.

zo. (page 1±o / original page 18o) The Allen & Unwin printings

have the sentence as “How shall we deﬁne a “typical” French-

man?” Here, the closing quotation mark has been moved to

make it “How shall we deﬁne a “typical Frenchman”?” Al-

though Russell’s manuscript is not entirely clear here, it appears

the latter was intended, and it also seems to make more sense

in context.

z1. (page 1±z / original page 1o1) “There is a type (r say) . . . ” has

been changed to “There is a type (τ say) . . . ” to match Russell’s

manuscript, and conventions followed elsewhere in the chapter.

zz. (page 1±6 / original page 1o¸) “. . . divided into numbers of

separate studies . . . ” has been changed to “. . . divided into a

number of separate studies . . . ” Russell’s manuscript just had

“number”, in the singular, without the indeﬁnite article. Some

Changes to Online Edition 161

emendation was necessary to make the passage grammatical,

but the ﬁx adopted here seems more likely what was meant.

zj. (page 1±; / original page 1o;) The passage “the propositional

function ‘if all α’s are β and x is an α, then x is a β’ is always true”

has been changed to “the propositional function ‘if all α’s are

β’s and x is an α, then x is a β’ is always true” to match Russell’s

manuscript, as well as to make it consistent with the other

paraphrase given earlier in the sentence. Thanks to Christof

Gr¨ aber for noticing this error.

z±. (page 1±o / original page zoo) “. . . without any special word for

forms . . . ” has been changed to “. . . without any special words

for forms . . . ”, which matches Russell’s manuscript and seems

to ﬁt better in the context.

z¸. (page 1¸¸ / original page zo;) The original index listed a ref-

erence to Frege on page 1o, but in fact, the discussion of Frege

occurs on page 11. Here, “1o” is crossed out, and “[11]” in-

serted.

Some very minor corrections to punctuation have been made to the

Allen & Unwin 1ozo printing, but not marked in green.

a) Ellipses have been regularized to three closed dots throughout.

b) (page ±o / original page ¸j) “We may deﬁne two relations . . . ”

did not start a new paragraph in previous editions, but does in

Russell’s manuscript, and is changed to do so.

c) (page ±o / original page ¸j) What appears in the 1ozo and later

printings as “. . . is the ﬁeld of Q. and which is . . . ” is changed

to “. . . is the ﬁeld of Q, and which is . . . ”

d) (page ±z / original page ¸6) “. . . a relation number is a class of

. . . ” is changed to “. . . a relation-number is a class of . . . ” to

match the hyphenation in the rest of the book (and in Russell’s

manuscript). A similar change is made in the index.

e) (page ±± / original page 6o) “. . . and “featherless biped,”—so

two . . . ” is changed to “. . . and “featherless biped”—so two . . . ”

f) (pages 61–6z / original pages 8z–8j) One misprint of “proges-

sion” for “progression”, and one misprint of “progessions” for

“progressions”, have been corrected. (Thanks to Christof Gr¨ aber

for noticing these errors in the original.)

g) (page 86 / original page 11¸) In the Allen & Unwin printing,

the “s” in “y’s” in what appears here as “Form all such sections

for all y’s . . . ” was italicized along with the “y”. Nothing in

Russell’s manuscript suggests it should be italicized, however.

(Again thanks to Christof Gr¨ aber.)

Changes to Online Edition 16z

h) (page o1 / original page 1z1) In the Allen & Unwin printing,

“Let y be a member of β . . . ” begins a newparagraph, but it does

not in Russell’s manuscript, and clearly should not.

i) (page o; / original pages 1zo–1jo) The phrase “well ordered”

has twice been changed to “well-ordered” to match Russell’s

manuscript (in the ﬁrst case) and the rest of the book (in the

second).

j) (page oo / original page 1j1) “The way in which the need for

this axiom arises may be explained as follows:—One of Peano’s

. . . ” is changed to “The way in which the need for this axiom

arises may be explained as follows. One of Peano’s . . . ” and

has been made to start a new paragraph, as it did in Russell’s

manuscript.

k) (page 1oj / original page 1j;) The accent on “M´ etaphysique”,

included in Russell’s manuscript but left oﬀ in print, has been

restored.

l) (page 11o / original page 1¸o) “. . . or what not,—and clearly

. . . ” is changed to “. . . or what not—and clearly . . . ”

m) (page 1jz / original page 1;6) Italics have been added to one

occurrence of “Waverley” to make it consistent with the others.

n) (page 1j8 / original page 18¸) “. . . most diﬃcult of fulﬁlment,—

it must . . . ” is changed to “. . . most diﬃcult of fulﬁlment—it

must . . . ”

o) (page 1±; / original page 1o;) In the Allen & Unwin printings,

“Socrates” was not italicized in “. . . we may substitute α for men,

β for mortals, and x for Socrates, where . . . ” Russell had marked

it for italicizing in the manuscript, and it seems natural to do

so for the sake of consistency, so it has been italicized.

p) (page 1¸j / original page zo¸) The word “seem” was not ital-

icized in “. . . a deﬁnition which might seem satisfactory for a

while . . . ” in the Allen & Unwin editions, but was marked to be

in Russell’s manuscript; it is italicized here.

q) (page 1¸6 / original page zo8) Under “Relations” in the index,

“similar, ¸zﬀ;” has been changed to “similar, ¸zﬀ.;” to match the

punctuation elsewhere.

There are, however, a number of other places where the previous

print editions diﬀer from Russell’s manuscript in minor ways that

were left unchanged in this edition. For a detailed examination of

the diﬀerences between Russell’s manuscript and the print editions,

and between the various printings themselves (including the changes

from the 1o1o to the 1ozo printings not documented here), see Ken-

neth Blackwell, “Variants, Misprints and a Bibliographical Index for

Introduction to Mathematical Philosophy”, Russell n.s. zo (zooo): ¸;–6z.

Changes to Online Edition 16j

pBertrand Russell’s Introduction to Mathematical Philosophy is in

the Public Domain.

See http://creativecommons.org/licenses/publicdomain/

cbaThis typesetting (including L

A

T

E

X code) and list of changes

are licensed under a Creative Commons Attribution—Share Alike

j.o United States License.

See http://creativecommons.org/licenses/by-sa/./us/

ii

[Russell’s blurb from the original dustcover:] This book is intended for those who have no previous acquaintance with the topics of which it treats, and no more knowledge of mathematics than can be acquired at a primary school or even at Eton. It sets forth in elementary form the logical deﬁnition of number, the analysis of the notion of order, the modern doctrine of the inﬁnite, and the theory of descriptions and classes as symbolic ﬁctions. The more controversial and uncertain aspects of the subject are subordinated to those which can by now be regarded as acquired scientiﬁc knowledge. These are explained without the use of symbols, but in such a way as to give readers a general understanding of the methods and purposes of mathematical logic, which, it is hoped, will be of interest not only to those who wish to proceed to a more serious study of the subject, but also to that wider circle who feel a desire to know the bearings of this important modern science.

CONTENTS

Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Editor’s Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I. The Series of Natural Numbers . . . . . . . . . . . . . II. Deﬁnition of Number . . . . . . . . . . . . . . . . . . .

iii iv vi

III. Finitude and Mathematical Induction . . . . . . . . . IV. The Deﬁnition of Order . . . . . . . . . . . . . . . . . V. Kinds of Relations . . . . . . . . . . . . . . . . . . . . . VI. Similarity of Relations . . . . . . . . . . . . . . . . . . VII. Rational, Real, and Complex Numbers . . . . . . . . . VIII. Inﬁnite Cardinal Numbers . . . . . . . . . . . . . . . . IX. Inﬁnite Series and Ordinals . . . . . . . . . . . . . . . X. Limits and Continuity . . . . . . . . . . . . . . . . . . XI. Limits and Continuity of Functions . . . . . . . . . . . XII. Selections and the Multiplicative Axiom . . . . . . . . XIII. The Axiom of Inﬁnity and Logical Types . . . . . . . . XIV. Incompatibility and the Theory of Deduction . . . . . XV. Propositional Functions . . . . . . . . . . . . . . . . . XVI. Descriptions . . . . . . . . . . . . . . . . . . . . . . . . XVII. Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . XVIII. Mathematics and Logic . . . . . . . . . . . . . . . . . . Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix: Changes to Online Edition . . . . . . . . . . . . . .

iii

mathematical logic is relevant to philosophy. The utmost endeavour has been made to avoid dogmatism on such questions as are still open to serious doubt. though it can hardly claim. It does deal. but belongs now to mathematics. hitherto only available to those who have mastered logical symbolism. therefore. For this reason.” and does not v aim at giving an exhaustive discussion of the problems with which it deals. the philosophy of mathematics will naturally be expected to deal with questions on the frontier of knowledge. claim to be an introduction to mathematical philosophy. A book dealing with those parts may. for example. as well as by its bearing on still unsolved problems. but are of at least equal philosophical interest. appears to invalidate much traditional philosophy. as to which comparative certainty is not yet attained. perhaps. to be actually dealing with a part of philosophy. The nature of inﬁnity and continuity. cannot. In this way. But speculation on such questions is hardly likely to be fruitful unless the more scientiﬁc parts of the principles of mathematics are known. be held to include such deﬁnite scientiﬁc results as have been obtained in this region. to those who accept it. vi with a body of knowledge which.PREFACE This book is intended essentially as an “Introduction.” though the matters concerned were included in philosophy so long as no satisfactory science of them existed. except where it steps outside its province. in the strict sense. It seemed desirable to set forth certain results. some purpose may be served by a succinct account of the main results of mathematical logic in a form requiring neither a knowledge of mathematics nor an aptitude for mathematical symiv . belonged in former days to philosophy. The beginnings of mathematical logic are less deﬁnitely known than its later portions. as well as on account of the intrinsic importance of the subject. Much of what is set forth in the following chapters is not properly to be called “philosophy. and even a good deal of what is current in the present day. and this endeavour has to some extent dominated the choice of topics considered. | however. in a form oﬀering the minimum of diﬃculty to the beginner. Mathematical philosophy.

Here. It is to be hoped that some readers may be suﬃciently interested to advance to a study of the method by which mathematical logic can be made helpful in investigating the traditional problems of philosophy.Preface v bolism. the method is more important than the results. . and the method cannot well be explained within the framework of such a book as the following. as elsewhere. from the point of view of further research. BERTRAND RUSSELL. But that is a topic with which the following pages have not attempted to deal. however.

it may be well to remind them from the side of philosophy that here. in order to perceive the bearing of the deﬁnitions and discussions that follow on the work of “traditional philosophy..D. in which these concepts play so large a part. that they should know the precise meaning that the science of mathematics. assigns to them. inﬁnity. H.] Those who. vi . at any rate. think that this book is out of place in the present Library. may be referred to what the author himself says on this head in the Preface. on the other hand. It is not necessary to agree with what he there suggests as to the readjustment of the ﬁeld of philosophy by the transference from it to mathematics of such problems as those of class. relying on the distinction between Mathematical Philosophy and the Philosophy of Mathematics. there be mathematicians to whom these deﬁnitions and discussions seem to be an elaboration and complication of the simple. both in one. LL. apparent simplicity may conceal a complexity which it is the business of somebody. If. continuity.” If philosophers cannot consent to relegate the criticism of these categories to any of the special sciences. like the author of this volume. it is essential. Muirhead.EDITOR’S NOTE [The note below was written by J. or. editor of the vii Library of Philosophy series in which Introduction to Mathematical Philosophy was originally published. to unravel. whether philosopher or mathematician. as elsewhere.

CHAPTER I

**THE SERIES OF NATURAL NUMBERS
**

Mathematics is a study which, when we start from its most familiar portions, may be pursued in either of two opposite directions. The more familiar direction is constructive, towards gradually increasing complexity: from integers to fractions, real numbers, complex numbers; from addition and multiplication to diﬀerentiation and integration, and on to higher mathematics. The other direction, which is less familiar, proceeds, by analysing, to greater and greater abstractness and logical simplicity; instead of asking what can be deﬁned and deduced from what is assumed to begin with, we ask instead what more general ideas and principles can be found, in terms of which what was our starting-point can be deﬁned or deduced. It is the fact of pursuing this opposite direction that characterises mathematical philosophy as opposed to ordinary mathematics. But it should be understood that the distinction is one, not in the subject matter, but in the state of mind of the investigator. Early Greek geometers, passing from the empirical rules of Egyptian land-surveying to the general propositions by which those rules were found to be justiﬁable, and thence to Euclid’s axioms and postulates, were engaged in mathematical philosophy, according to the above deﬁnition; but when once the axioms and postulates had been reached, their deductive employment, as we ﬁnd it in Euclid, belonged to mathematics in the | ordinary sense. The distinction between mathematics and mathe- matical philosophy is one which depends upon the interest inspiring the research, and upon the stage which the research has reached; not upon the propositions with which the research is concerned. We may state the same distinction in another way. The most obvious and easy things in mathematics are not those that come logically at the beginning; they are things that, from the point of view of logical deduction, come somewhere in the middle. Just as the easiest bodies to see are those that are neither very near nor very far, neither very small nor very great, so the easiest conceptions to grasp are those that are neither very complex nor very simple (using “simple” in a logical sense). And as we need two sorts of instruments, the telescope and

Chap. I. The Series of Natural Numbers

the microscope, for the enlargement of our visual powers, so we need two sorts of instruments for the enlargement of our logical powers, one to take us forward to the higher mathematics, the other to take us backward to the logical foundations of the things that we are inclined to take for granted in mathematics. We shall ﬁnd that by analysing our ordinary mathematical notions we acquire fresh insight, new powers, and the means of reaching whole new mathematical subjects by adopting fresh lines of advance after our backward journey. It is the purpose of this book to explain mathematical philosophy simply and untechnically, without enlarging upon those portions which are so doubtful or diﬃcult that an elementary treatment is scarcely possible. A full treatment will be found in Principia Mathematica; the treatment in the present volume is intended merely as an introduction. To the average educated person of the present day, the obvious starting-point of mathematics would be the series of whole numbers, , , , , . . . etc. | Probably only a person with some mathematical knowledge would think of beginning with instead of with , but we will presume this degree of knowledge; we will take as our starting-point the series: , , , , . . . n, n + , . . . and it is this series that we shall mean when we speak of the “series of natural numbers.” It is only at a high stage of civilisation that we could take this series as our starting-point. It must have required many ages to discover that a brace of pheasants and a couple of days were both instances of the number : the degree of abstraction involved is far from easy. And the discovery that is a number must have been diﬃcult. As for , it is a very recent addition; the Greeks and Romans had no such digit. If we had been embarking upon mathematical philosophy in earlier days, we should have had to start with something less abstract than the series of natural numbers, which we should reach as a stage on our backward journey. When the logical foundations of mathematics have grown more familiar, we shall be able to start further back, at what is now a late stage in our analysis. But for the moment the natural numbers seem to represent what is easiest and most familiar in mathematics. But though familiar, they are not understood. Very few people are prepared with a deﬁnition of what is meant by “number,” or “,” or “.” It is not very diﬃcult to see that, starting from , any other of the

University Press, vol. i., ; vol. ii., ; vol. iii., . By Whitehead and Russell.

Cambridge

Chap. I. The Series of Natural Numbers

natural numbers can be reached by repeated additions of , but we shall have to deﬁne what we mean by “adding ,” and what we mean by “repeated.” These questions are by no means easy. It was believed until recently that some, at least, of these ﬁrst notions of arithmetic must be accepted as too simple and primitive to be deﬁned. Since all terms that are deﬁned are deﬁned by means of other terms, it is clear that human knowledge must always be content to accept some terms as intelligible without deﬁnition, in order | to have a starting-point for its deﬁnitions. It is not clear that there must be terms which are incapable of deﬁnition: it is possible that, however far back we go in deﬁning, we always might go further still. On the other hand, it is also possible that, when analysis has been pushed far enough, we can reach terms that really are simple, and therefore logically incapable of the sort of deﬁnition that consists in analysing. This is a question which it is not necessary for us to decide; for our purposes it is suﬃcient to observe that, since human powers are ﬁnite, the deﬁnitions known to us must always begin somewhere, with terms undeﬁned for the moment, though perhaps not permanently. All traditional pure mathematics, including analytical geometry, may be regarded as consisting wholly of propositions about the natural numbers. That is to say, the terms which occur can be deﬁned by means of the natural numbers, and the propositions can be deduced from the properties of the natural numbers—with the addition, in each case, of the ideas and propositions of pure logic. That all traditional pure mathematics can be derived from the natural numbers is a fairly recent discovery, though it had long been suspected. Pythagoras, who believed that not only mathematics, but everything else could be deduced from numbers, was the discoverer of the most serious obstacle in the way of what is called the “arithmetising” of mathematics. It was Pythagoras who discovered the existence of incommensurables, and, in particular, the incommensurability of the side of a square and the diagonal. If the length of the side is inch, the number of inches in the diagonal is the square root of , which appeared not to be a number at all. The problem thus raised was solved only in our own day, and was only solved completely by the help of the reduction of arithmetic to logic, which will be explained in following chapters. For the present, we shall take for granted the arithmetisation of mathematics, though this was a feat of the very greatest importance. | Having reduced all traditional pure mathematics to the theory of the natural numbers, the next step in logical analysis was to reduce this theory itself to the smallest set of premisses and undeﬁned terms from which it could be derived. This work was accomplished by Peano. He showed that the entire theory of the natural numbers could be derived from three primitive ideas and ﬁve primitive propo-

The Series of Natural Numbers sitions in addition to those of pure logic. and so on. for the present. We shall have much to say concerning mathematical induction in the sequel. Afterwards the word will be used in a more general sense.Chap. the truth of this whole series is assured if the truth of the ﬁve primitive propositions is guaranteed. () No two numbers have the same successor. in virtue of (). Let us consider brieﬂy the kind of way in which the theory of the natural numbers results from these three ideas and ﬁve propositions. By “number” he means. These three ideas and ﬁve propositions thus became. the successor of is . By “successor” he means the next number in the natural order. that there is nothing erroneous in the purely logical apparatus which is also involved. He is not assuming that we know all the members of this class. if it were. () Any property which belongs to . The work of analysing mathematics is extraordinarily facilitated by this work of Peano’s. the class of the natural numbers. two diﬀerent numbers would have the same successor. of course.” and so on. and.” if one may use such an expression. and also to the successor of every number which has the property. We can obviously go on as long as we like with these deﬁnitions. we are concerned with it only as it occurs in Peano’s analysis of arithmetic.” though we do not know all men individually. and in virtue of () none of the shall use “number” in this sense in the present chapter. belongs to all numbers. To begin with. The three primitive ideas in Peano’s arithmetic are: . since. hostages for the whole of traditional pure mathematics. just as we know what we mean when we say “Jones is a man. but only that we know what we mean when we say that this or that is a number. I. so could all pure mathematics. That is to say. The ﬁve primitive propositions which Peano assumes are: () is a number. If they could be deﬁned and proved in terms of others. in virtue of (). is equal to that of the whole series of sciences that have been deduced from the theory of the natural numbers. every number that we reach will have a successor. in this connection. We . | () is not the successor of any number. the successor of is . because. () The successor of any number is a number. provided. this cannot be any of the numbers already deﬁned. Their logical “weight. number. successor. as it were. we deﬁne as “the successor of . The last of these is the principle of mathematical induction.” as “the successor of .

is not a “number” in the sense which we are now giving to the word “number. even the fourth.” and “successor”—are capable of an inﬁnite number of diﬀerent interpretations. all of which will satisfy the ﬁve primitive propositions. “. . i. whatever number n may be.” and let the “successor” of a number be what results from adding two to it. . In virtue of () all numbers come in this series. The reader can easily convince himself that any ordinary elementary proposition of arithmetic can be proved by means of our ﬁve premisses. Then all our primitive propositions are satisﬁed. Taking any number m. . to that of Frege. but we shall give some of the reasons why Peano’s treatment is less ﬁnal than it appears to be. and so on.Chap. () Let “” be taken to mean .e. () Let “” have its usual meaning. . by mathematical induction. and (b) if a number n belongs to it. two. and let “number” be taken to mean the numbers from onward in the series of natural numbers. . who | represents the last perfection of the “arithmetisation” of mathematics. let “number” mean the set . so does its successor. () Let “” mean the number one. in this chapter.” It is obvious that any number may be substituted for in this example. I. Then “” will stand for the number two.” “number. In virtue of () this gives a deﬁnition of the sum of m and n. and m + (n + ) as the successor of m + n. the series of “numbers” now will be . eight . Suppose we wish to deﬁne the sum of two numbers. six. In the ﬁrst place. We will give some examples. The Series of Natural Numbers numbers we reach in the series of successors can be . . Similarly we can deﬁne the product of any two numbers. for. . who ﬁrst succeeded in “logicising” mathematics.. whence. in reducing to logic the arithmetical notions which his predecessors had shown to be suﬃcient for mathematics. We shall not. we deﬁne m + as m. which begins with and travels on through successive successors: for (a) belongs to this series. It is time now to turn to the considerations which make it necessary to advance beyond the standpoint of Peano. . though is the successor of . actually give Frege’s deﬁnition of number and of particular numbers. four. All Peano’s ﬁve premisses are satisﬁed still. but let “number” mean what we usually call “even numbers. and if he has any diﬃculty he can ﬁnd the proof in Peano.. Thus the series of successors gives us an endless series of continually new numbers. Peano’s three primitive ideas—namely. “” will stand for the number four. every number belongs to the series.

x . all these possible interpretations will be equally true. () “ is not the successor of any number. the name “number” to the whole set of its terms. It can be proved.e. xn . and the name “successor” to the next in the progression. has a beginning. a successor to each term (so that there is no last term). . contains no repetitions. The Series of Natural Numbers and let “successor” mean “half.” i. conversely. or any other terms of which there is an inﬁnite supply.e. belongs to all the x’s. we have a set of terms verifying Peano’s axioms. | which is endless. I.” Then all Peano’s ﬁve axioms will be true of this set.” i. Hence these ﬁve axioms may be used to deﬁne the class of progressions: “progressions” are “those series which verify these ﬁve axioms. . every progression veriﬁes Peano’s ﬁve axioms. . xm+ and xn+ are diﬀerent. and let the “successor” of xn mean xn+ . and every term can be reached from the start in a ﬁnite number of steps. . () “No two numbers have the same successor. . in which there is a ﬁrst term. . and belongs to xn+ provided it belongs to xn . though the formal proof is somewhat long. xn . () This becomes: Any property which belongs to x . Let “” mean x . A series of the form x . x . . xn+ is also in the set. that every series which veriﬁes Peano’s ﬁve axioms is a progression. () “The successor of any number is a number. . no repetitions. no term in the set comes before x . x . this results from the fact that (by hypothesis) there are no repetitions in the set. This is easily seen. let “number” mean the whole set of terms. Progressions are of great importance in the principles of mathematics.” Any progression may be taken as the basis of pure mathematics: we may give the name “” to its ﬁrst term.Chap. The progression need not be composed of numbers: it may be | composed of points in space. As we have just seen. x . if xm and xn are two diﬀerent members of the set. . given any series x .e. x . It is clear that such examples might be multiplied indeﬁnitely. x is a member of the set. .e. .” i.” i. Each diﬀerent progression will give rise to a diﬀerent interpretation of all the propositions of traditional pure mathematics. In fact. . and has no terms that cannot be reached from the beginning in a ﬁnite number of steps. This follows from the corresponding property for numbers. is called a progression. . or moments of time. taking any term xn in the set. Then () “ is a number.

it does not enable us to know whether there are any sets of terms verifying Peano’s axioms. . and this requires that our numbers should have a deﬁnite meaning. namely. We have already some knowledge (though not suﬃciently articulate or analytic) of what we mean by “” and “” and so on. is to say “we know what we mean by ‘’ and ‘number’ and ‘successor.” terms concerning which we make certain hypotheses. It might be suggested that. In the second place. our theorems will not be proved concerning an ascertained set of terms called “the natural numbers. is important. and “” meant . but must be independently understood. and our use of numbers in arithmetic must conform to this knowledge. I.’ though we cannot explain what we mean in terms of other simpler concepts. all that we can do. but to apply in the right way to common objects. Such a procedure is not fallacious. If we adopt this plan. but it is the object of mathematical philosophy to put oﬀ saying it as long as possible. In the ﬁrst place.” but concerning all sets of terms having certain properties. indeed for certain purposes it represents a valuable generalisation. we want our numbers to be such as can be used for counting common objects. as already observed. We want to have ten ﬁngers and two eyes and one nose. This deﬁnite meaning is deﬁned by the logical theory of arithmetic. We cannot secure that this shall be the case by Peano’s method. They will then no longer be terms which have a meaning that is deﬁnite though undeﬁned: they will be “variables. might be all right for pure mathematics. We want our numbers not merely to verify mathematical formulæ. This point.” It is quite legitimate to say this when we must. The Series of Natural Numbers In Peano’s system there is nothing to enable us to distinguish between these diﬀerent interpretations of his primitive ideas. We want “” and “number” and “successor” to have meanings which will give us the right allowance of ﬁngers and eyes and noses. instead of setting up “” and “number” and “successor” as terms of which we know the meaning although we cannot deﬁne them. it does not even give the faintest suggestion of any way of discovering whether there are such sets.Chap. and at some point we all must. if we adopt his method. those stated in the ﬁve axioms.” and that we shall not suppose that this symbol means or Cleopatra’s Needle or any of the other things that it might mean. not merely that they should have certain formal properties. we might let them | stand for any three terms that verify Peano’s ﬁve axioms. and so on. A system in which “” meant . It is assumed that we know what is meant by “. By the logical theory of arithmetic we are able to put it oﬀ for a very long time. but which are otherwise undetermined. but would not suit daily life. that “” and “number” and “successor” cannot be deﬁned by means of Peano’s ﬁve axioms. But from two points of view it fails to give an adequate basis for arithmetic.

it attracted almost no attention. we will say as little as possible. and of the very highest importance.. The answer was given by Frege in . vol. In seeking a deﬁnition of number. namely. which is quite a diﬀerent thing. Jones.” or sometimes a “set. as man is what is characteristic of men. Many philosophers. and which distinguishes them from other collections. For the present. But there are some remarks that must be made immediately. We may enumerate its members. A number is something that characterises certain collections. but has only been correctly answered in our own time. are really setting to work to deﬁne plurality. as when we say.” Other words used in mathematics for the same thing are “aggregate” and “manifold. when attempting to deﬁne number. Instead of speaking of a “collection. A trio of men. in his Grundlagen der Arithmetik.” we shall as a rule speak of a “class. those that have that number.CHAPTER II DEFINITION OF NUMBER The question “What is a number?” is one which has been often asked. same answer is given more fully and with more development in his Grundgesetze der Arithmetik. This point may seem elementary and scarcely worth mentioning. the ﬁrst thing to be clear about is what we may call the grammar of our inquiry. The . A particular number is not identical with any collection of terms having that number: the number is not identical with | the trio consisting of Brown. not diﬃcult. A plurality is not an instance of number. but of some particular number. i. and the deﬁnition of number which it contains remained practically unknown until it was rediscovered by the present author in . and the number is an instance of number. for example. .” We shall have much to say later on about classes. Number is what is characteristic of numbers. but the trio is not an instance of number. The number is something which all trios have in common. is an instance of the number . A class or collection may be deﬁned in two ways that at ﬁrst sight seem quite distinct. and Robinson. yet it has proved too subtle for the philosophers. with few exceptions. Although this book is quite short.

() Brown. and Robinson. we ﬁnd that enumeration is not even theoretically possible for beings who only live for a ﬁnite time. In the third place.” Or we may mention a deﬁning property. Consider such a formula as “x is Brown or x is Jones or x is Robinson. It may be taken as assigning a property common to the members of the class consisting of these three | men. Each of these points needs a word of explanation. Brown and Jones and Robinson. in three diﬀerent ways. namely.Chap. numbers themselves form an inﬁnite collection. () that the intensional one often cannot even theoretically be reduced to the extensional one.” and the one which mentions a deﬁning property is called a deﬁnition by “intension. which. that there are an inﬁnite collection of trios in the world. and cannot therefore be deﬁned by enumeration. A similar treatment can obviously be applied to any other class given in extension. At some point we must content ourselves with “and so on. for example.” We cannot enumerate all fractions or all irrational numbers. though possible. We cannot enumerate all the natural numbers: they are . Jones. . But when we come to consider inﬁnite classes. In the ﬁrst place. the collections having a given number of terms themselves presumably form an inﬁnite collection: it is to be presumed.” The deﬁnition which enumerates is called a deﬁnition by “extension. In this respect it resembles a cubic equation with its three roots. or even all the inhabitants of London. . In the second place. This is enough to show that deﬁnition by extension is not necessary to knowledge about a class. yet a great deal is known about each of these classes. These remarks are relevant.” This formula will be true for just three x’s. we wish to deﬁne “number” in such a way that inﬁnite numbers may be possible. seems unlikely. and Robinson all of them possess a certain property which is possessed by nothing else in the whole universe. Deﬁnition of Number “The collection I mean is Brown. as when we speak of “mankind” or “the inhabitants of London. Jones. or all of any other inﬁnite collection. for if this were not the case the total number of things in the world would be ﬁnite. II. and peculiar to them. No one man could actually enumerate all men. the one by intension is logically more fundamental. This property can be used to give a deﬁnition by intension of the class consisting of Brown and Jones and Robinson.” Of these two kinds of deﬁnition. when we are seeking the deﬁnition of number. . the property of being either Brown or Jones or Robinson. namely. and so on. () It is obvious that in practice we can often know a great deal about a class without being able to enumerate its members. thus we must be able to . This is shown by two considerations: () that the extensional deﬁnition can always be reduced to an intensional one. Thus our knowledge in regard to all such collections can only be derived from a deﬁnition by intension.

Deﬁnition of Number speak of the number of terms in an inﬁnite collection. it is simpler logically to ﬁnd out whether two collections have the same number of terms than it is to deﬁne what will be explained later. Men | may be deﬁned as featherless bipeds. however. thus each is a class of classes. though familiar. otherwise we could be content with the properties common and peculiar to their members. a class and a deﬁning characteristic of it are practically interchangeable. manufactured out of deﬁning characteristics. Any one of these properties can be used in place of the class whenever uniqueness is not important. As . each bundle consisting of all the collections that have a certain number of terms. and put them in the same bundle if they have the same number of members. How shall we decide whether two collections are to belong to the same bundle? The answer that suggests itself is: “Find out how many members each has. We need. But for the present it will simplify our exposition to treat classes as if they were real. namely. each of which is a class of two members. Our deﬁnition of number must not assume in advance that all numbers are ﬁnite. it is clear that number is a way of bringing together certain collections.” But this presupposes that we have deﬁned numbers. classes. counting. We can suppose all couples in one bundle. all trios in another. The vital diﬀerence between the two consists in the fact that there is only one class having a given set of members. without a vicious circle. whereas there are always many diﬀerent characteristics by which a given class may be deﬁned. for example. therefore. i.Chap. In actual fact. when the collection is ﬁnite.e. and that we know how to discover how many terms a collection has. We are so used to the operation of counting that such a presupposition might easily pass unnoticed. as a means of discovering how many terms a collection has. and so on. and we cannot in any case. moreover it is only available. In fact.e. or (more correctly) by the traits by which Swift delineates the Yahoos. Returning now to the deﬁnition of number. Each bundle is a class whose members are collections. and such a collection must be deﬁned by intension. In this way we obtain various bundles of collections. or as rational animals. For many purposes. is a class of classes: each couple is a class with two members. classes may be regarded as logical ﬁctions. by a property common to all its members and peculiar to them. i. and the whole bundle of couples is a class with an inﬁnite number of members. It is this fact that a deﬁning characteristic is never unique which makes classes useful. The bundle consisting of all couples. is logically a very complex operation. because numbers are used in counting. those that have a given number of terms. | use counting to deﬁne numbers. some other method of deciding when two collections have the same number of terms. II.

” It should be observed that the number is not used in these deﬁnitions. II. but that of eldest son to father is one-one. and many-one relations. and x does not have the same relation to any term y other than y. it is called “many-one. Similarly less is the converse of greater. Two classes are said to be “similar” when there is a one-one | relation which correlates the terms of the one class each with one term of the other class. in Mahometan countries it is one-many.” A relation is said to be “one-one” when. The converse domain of a relation is the domain of its converse: thus the class of wives is the converse domain of the relation of husband to wife. and husbands and wives together are the domain of the relation of marriage. the relation of husband to wife is one-one. We do not need a census to assure us of this. We know the number must be the same in both collections. because each husband has one wife and each wife has one husband. it is clear that the number of husbands living at any moment would be exactly the same as the number of wives. in the same manner in which the relation of marriage correlates husbands with wives. We may now state our deﬁnition of similarity as follows:— One class is said to be “similar” to another when there is a one-one . the converse of a given relation is that relation which holds between y and x whenever the given relation holds between x and y. The relation of husband and wife is what is called “one-one. and so on. no other term x has the same relation to y. In Christian countries. so is the relation of n to n or to n. later is the converse of earlier. in Tibet it is many-one. it becomes two-one. that of son to father is many-one. when only the second is fulﬁlled. the relation of n to n + is one-one. husbands are the domain of the relation of husband to wife. The relation of father to son is one-many. The class of those terms that have a given relation to something or other is called the domain of that relation: thus fathers are the domain of the relation of father to child. An illustration will make this clear. the relation of n to n is one-one. A few preliminary deﬁnitions will help us to state this deﬁnition more precisely. but in many other connections. The relation of wife to husband is called the converse of the relation of husband to wife. If n is any number. if x has the relation in question to y. Deﬁnition of Number that number is. which play a great part in the principles of mathematics. but when negative numbers are admitted. nor do we need to know what is the actual number of husbands and of wives. not only in relation to the deﬁnition of numbers. If there were no polygamy or polyandry anywhere in the world. since n and −n have the same square. When only the ﬁrst of these two conditions is fulﬁlled. When we are considering only positive numbers. These instances should suﬃce to make clear the notions of one-one. the relation is called “onemany”. one-many.Chap. Generally. wives are the domain of the relation of wife to husband.

and so on. provided the collection is ﬁnite. The notion of similarity is logically presupposed in the operation of counting. But this result. thus proving that the two classes are similar. In counting. and transitive when it possesses the third. while the other is the converse domain. without having to establish an order of precedence among them.Chap. () that if a class α is similar to a class β. so long as we conﬁne ourselves to ﬁnite numbers. The act of counting consists in establishing a one-one correlation | between the set of objects counted and the natural numbers (excluding ) that are used up in the process. but order is not of the essence of number: it is an irrelevant addition. The notion of similarity also does not require that the classes which are similar should be ﬁnite. as ﬁrst. It is obvious to common sense that two ﬁnite classes have the same number of terms if they are similar. A relation is said to be reﬂexive when it possesses the ﬁrst of these properties. II. besides being only applicable to ﬁnite collections. Take. depends upon and assumes the fact that two classes which are similar have the same number of terms. an unnecessary complication from the logical point of view.. And we also know that. in the sense in which we were asking this question earlier in this chapter. but not otherwise. we saw that the number of husbands is the same as the number of wives. The notion of similarity does not demand an order: for example. for what we do when we count (say) objects is to show that the set of these objects is similar to the set of numbers to . We may thus use the notion of “similarity” to decide when two collections are to belong to the same bundle. then β is similar to α. there are just n numbers from up to n. for example. and is logically simpler though less familiar. Deﬁnition of Number relation of which the one class is the domain. Relations which possess these properties are an important kind. etc. with /. Then we want a bundle of all the classes that have one . symmetrical when it possesses the second. () that if α is similar to β and β to γ. Accordingly common sense concludes that there are as many objects in the set to be counted as there are numbers up to the last number used in the counting. We want to make one bundle containing the class that has no members: this will be for the number . the natural numbers (excluding ) on the one hand. then α is similar to γ. it is necessary to take the objects counted in a certain order. and it is worth while to note that similarity is one of this kind of relations. It is easy to prove () that every class is similar to itself. and the fractions which have for their numerator on the other hand: it is obvious that we can correlate with /. It is obvious that a relation which is symmetrical and transitive must be reﬂexive throughout its domain. Hence it follows that the last number used in counting a collection is the number of terms in the collection. second. third.

those collections that are “similar” to it will have the same number of terms. A number will be a set of classes such as that any two are similar to each | other. Thus the number of a couple will be the class of all couples. though this impression will soon wear oﬀ. for example. We naturally think that the class of couples (for example) is something diﬀerent from the number . or. Accordingly we set up the following deﬁnition:— The number of a class is the class of all those classes that are similar to it. Deﬁnitions of this sort are in fact very common. Given any collection. We may now go on to deﬁne numbers in general as any one of the bundles into which similarity collects classes. this deﬁnition secures deﬁniteness and indubitableness. At the expense of a little oddity. is a metaphysical entity about which we can never feel sure that it exists or that we have tracked it down. Then. The class of fathers. In fact. So far we have not suggested anything in the slightest degree paradoxical. and none outside the set are similar to any inside the set. the class of all couples will be the number . and so on. therefore we may deﬁne number in general in terms of “the number of a given class” without committing any logical error. more simply still: A number is anything which is the number of some class. the class of all those collections that are similar to it will be the class of trios. we want a bundle consisting | of all couples. Such a deﬁnition has a verbal appearance of being circular. a number (in general) is any collection which is the number of one of its members. Deﬁnition of Number member: this will be for the number . which we are sure of. II. according to our deﬁnition. than to hunt for a problematical number which must always remain elusive.Chap. It is very easy to see that if (for example) a collection has three members. for the number . in any other sense.” It is obvious that it gives results conformable to usage so long as we conﬁne ourselves to ﬁnite collections. We deﬁne “the number of a given class” without using the notion of number in general. whereas the number . we can deﬁne the bundle it is to belong to as being the class of all those collections that are “similar” to it. But there is no doubt about the class of couples: it is indubitable and not diﬃcult to deﬁne. And whatever number of terms a collection may have. We may take this as a deﬁnition of “having the same number of terms. then one of all trios. would have to be deﬁned by ﬁrst deﬁning what . But when we come to the actual deﬁnition of numbers we cannot avoid what must at ﬁrst sight seem a paradox. but in fact it is not. and it is not diﬃcult to prove that numbers so deﬁned have all the properties that we expect numbers to have. It is therefore more prudent to content ourselves with the class of couples. In other words.

Chap. Deﬁnition of Number it is to be the father of somebody. This kind of procedure is very common. and it is important to realise that it is legitimate and even often necessary. . But ﬁrst we must decide what we mean by “ﬁnite” and “inﬁnite. Similarly if we want to deﬁne square numbers (say). II. It remains to be seen how it will serve for inﬁnite collections. and then deﬁne square numbers as those that are the squares of other numbers. then the class of fathers will be all those who are somebody’s father.” which cannot be done within the limits of the present chapter. We have now given a deﬁnition of numbers which will serve for ﬁnite collections. we must ﬁrst deﬁne what we mean by saying that one number is the square of another.

such as . the proof that we can reach it by proceeding step by step in this fashion may be made. . It is this “and so on” that we wish to replace by something less vague and indeﬁnite. We might be tempted to say that “and so on” means that the process of proceeding to the successor may be repeated any ﬁnite number of times. ..” “number. as the successor of . and so on.CHAPTER III FINITUDE AND MATHEMATICAL INDUCTION The series of natural numbers. by actual experiment: we can go on until we actually arrive at . Our deﬁnition must not assume that we know what a ﬁnite number is. It is easy to see that we can reach any assigned number.” and therefore we must not use this notion in our deﬁnition.. if we have the patience. say .” then we deﬁne “” as “the successor of . What are the numbers that can be reached.” But we may go a step farther: we can deﬁne all the natural numbers if we know what we mean by “” and “successor. as the successor of . it is not available for proving the general proposition that all such numbers can be reached in this way. given the terms “” and | “successor”? Is there any way by which we can deﬁne the whole class of such numbers? We reach . We will not yet consider how “” and “successor” are to be deﬁned: we will for the moment assume that we know what these terms mean. as we saw in Chapter I.. and why the method by which it is done cannot be extended beyond the ﬁnite. by proceeding from step by step from each number to its successor.e.” and “successor. as the successor of . But although the method of experiment is available for each particular natural number.” and so on. i. can all be deﬁned if we know what we mean by the three terms “. .. but the problem upon which we are engaged is the problem of deﬁning “ﬁnite number. In the case of an assigned number. and show how thence all other natural numbers can be obtained. Is there any other way by which this can be proved? Let us consider the question the other way round. We ﬁrst deﬁne “” as “the successor of .” It will help us to understand the diﬀerence between ﬁnite and inﬁnite to see how this can be done.

It stated that any property which belongs to . in Chapter I. e. It is easy to see. It is again easy to see that the posterity of a natural number consists of itself and all greater natural numbers. It will be remembered that. This was then presented as a principle. We shall do well to begin with some deﬁnitions. It is now not diﬃcult to make it obvious that the posterity of is the same set as those terms that can be reached from by successive steps from next to next. this was the ﬁfth of the ﬁve primitive propositions which we laid down about the natural numbers.g. We will deﬁne the “posterity” of a given natural number with respect to the relation “immediate predecessor” (which is the converse of “successor”) as all those terms that belong to every hereditary class to which the given number belongs. and is the successor of . Finitude and Mathematical Induction The key to our problem lies in mathematical induction. but we shall now adopt it as a deﬁnition. the posterity of will consist of those terms which belong to every inductive class. the comparison of a relatively vague idea . III. it follows that is a member of it. and to the successor of any number which has the property. A property is said to be “inductive” when it is a hereditary | property which belongs to . belongs to both these sets (in the sense in which we have deﬁned our terms). so is n + . to all without exception.Chap. Similarly a class is “inductive” when it is a hereditary class of which is a member. whenever it belongs to a number n. because a hereditary class contains the successors of its members. that to say a property is hereditary is equivalent to saying that it belongs to all the natural numbers not less than some one of them. which will be useful in other connections also. but as the point is important we will set forth the matter in some detail. if n belongs to both sets. For. or it may be that it belongs to all that are not less than . in the ﬁrst place. whenever n is a member of the class. it also belongs to n + . but this also we do not yet oﬃcially know. Similarly a class is said to be “hereditary” if. It is to be observed that we are dealing here with the kind of matter that does not admit of precise proof. Given a hereditary class of which is a member.e. namely. By the above deﬁnitions. so does n + ... A property is said to be “hereditary” in the natural-number series if. it must belong to all that are not less than . Similarly. given a hereditary class of which is a member. and so on. i. Thus we can prove by a step-by-step procedure that any assigned natural number. or all that are not less than . though we are not yet supposed to know. It is not diﬃcult to see that the terms obeying it are the same as the numbers that can be reached from by successive steps from next to next. say . the successor of n. belongs to all the natural numbers. it follows that is a member of it. in the second place. is a member of every inductive class.

Certain niceties are required to make this deﬁnition perfect. It remains to deﬁne “successor. though it seems as if it conveyed a deﬁnite meaning. “the posterity of ” is precise and explicit just where the other idea is hazy. ∗ . vol. See Principia Mathematica.” By the general deﬁnition of number. namely. namely. The notion of “those terms that can be reached from by successive steps from next to next” is vague. on the other hand. We have thus reduced Peano’s three primitive ideas to ideas of logic: we have given deﬁnitions of them which make them deﬁnite. the one asserting that is a number and the one asserting mathematical induction—become unnecessary.) Thus we have the following purely logical deﬁnition:— is the class whose only member is the null-class. Thus we have the following deﬁnition:— The successor of the number of terms in the class α is the number of terms in the class consisting of α together with x. but they need not concern us. We have thus arrived at a deﬁnition of one of Peano’s three primitive ideas in terms of the other two. the null-class. i. since they result from the deﬁnition. ii.e. the number of terms in the null-class is the set of all classes similar to the null-class. i. The one asserting that the successor of a natural number is a natural number is only needed in the weakened form “every natural number has a successor. III. whereas the null-class itself has no members. let α be a class which has n members.) a logical deﬁnition of the number of terms in a class. (This is not identical with the null-class: it has one member. It will be remembered that we | have already given (in Chapter II. Finitude and Mathematical Induction with a relatively precise one. Then the class consisting of α with x added on will have n + members. (as is easily proved) the set consisting of the null-class all alone. i.e.” Given any number n. It may be taken as giving what we meant to mean when we spoke of the terms that can be reached from by successive steps. as we shall explain when we come to the theory of classes. As a result of this deﬁnition. where x is any term not belonging to the class.Chap.” We can. the class whose only member is the null-class. sor”). in the class which is called the “null-class. of course. and let x be a term which is not a member of α. two of his primitive propositions—namely. we deﬁned it as the set of all classes that are similar to the given class. easily deﬁne “” and “successor” by means of the deﬁnition of number in general which we arrived at in Chapter II. We now lay down the following deﬁnition:— The “natural numbers” are the posterity of with respect to the | relation “immediate predecessor” (which is the converse of “succes. A class which has one member is never identical with that one member.e. . The number is the number of terms in a class which has no members.

This failure of the third axiom cannot arise. we have already succeeded in making two of them demonstrable by our deﬁnition of “natural number. for given two numbers m and n.. in so far as it is deducible from the theory of the natural numbers. III.” The diﬃculty does not arise unless the total number of individuals in the universe is ﬁnite. it is easy to prove that we cannot have m + = n + unless we have m = n. but in seeing how to prove his ﬁve primitive propositions. We shall return to this topic at a later stage. neither of which is the total number of individuals in the universe. by means of which we deﬁned the natural numbers. is capable of generalisation. then there would be no class of individuals. It follows that all pure mathematics. as they were when they were only determinate to the extent of obeying Peano’s ﬁve axioms. Thus we should have = . . for rational dynamics. namely. part vii. is only a prolongation of logic. But there is a diﬃculty about the remaining primitive proposition. We have removed them from the fundamental apparatus of terms that must be merely apprehended. The process of mathematical induction.” if.” How stands it with the remaining three? It is very easy to prove that is not the successor of any number. Thus we should have two diﬀerent numbers with the same successor. and have thus increased the deductive articulation of mathematics. i. to the number to which m has the relation N.e. we have now succeeded not only in deﬁning Peano’s | three primitive ideas. if the number of individuals in the world is not ﬁnite. But let us suppose that the total number of individuals in the universe were (say) . see Principles of Mathematics. Assuming that the number of individuals in the universe is not ﬁnite. Chapter XIII. So would the number . As regards the ﬁve primitive propositions. Finitude and Mathematical Induction no longer capable of an inﬁnity of diﬀerent meanings. And a number n will be said to belong to the “posterity” of m with respect to the relation N if n has See For geometry. We deﬁned the natural numbers as the “posterity” of with respect to the relation of a number to its immediate successor. and that the successor of any number is a number. “no two numbers have the same successor. and the number would be the null-class. whenever the property belongs to a number m. part vi. If we call this relation N. ibid. as we have shown elsewhere. A property is “hereditary with respect to N. however. although would not be the same as .Chap.” or simply “N-hereditary.. therefore the successor of would be the same as the successor of . The extension of this result to those modern branches of mathematics which are not deducible from the theory of the natural numbers oﬀers no diﬃculty of principle. it also belongs to m + . any number m will have this relation to m + . by means of primitive ideas and propositions belonging to logic. in so far as it is not purely analytical.

. Here we have ﬁrst a series of negative fractions with no end. We have framed the above deﬁnitions so that if a term is the ancestor of anything it is its own ancestor and belongs to its own posterity. − . These . . . . if it belongs to a term x. . C. . and were published so long ago as in his Begriﬀsschrift. . then it belongs to y. Shall we say that. we can lay down the following deﬁnitions: — A property is called “R-hereditary” when.” “ancestor” and “posterity” will have the usual meanings. as we saw. . each is a parent of the next. the ﬁrst person who ever read it—more than twenty years after its publication. − . . But this deﬁnition is not adequate unless we add that the number of intermediate terms is to be ﬁnite. provided x is a term which has the relation R to something or to which something has the relation R. for example.” A brief consideration of this point will serve to show the importance of the theory. Finitude and Mathematical Induction every N-hereditary property belonging to m. there are a certain number of people. who is a parent of Z. and then a series of positive fractions with no beginning. −/ is an ancestor of /? It will be so according to the beginner’s deﬁnition suggested above. “ﬁnite” is to be deﬁned by means of mathematical induction. In spite of the great value of this work. in this series. It is. For this purpose. It will be observed that if we take for R the relation “parent. But. . no one could have deﬁned “ancestor” precisely in terms of “parent. except that a person will be included among his own ancestors and posterity. I believe. (This is only to exclude trivial cases. A term x is said to be an “R-ancestor” of the term y if y has every R-hereditary property that x has.) | The “R-posterity” of x is all the terms of which x is an R-ancestor. it is essential that the number of intermediaries should be ﬁnite.Chap. . A person confronted for the ﬁrst time with the problem of deﬁning “ancestor” in terms of “parent” would naturally say that A is an ancestor of Z if. I was. A class is R-hereditary when its deﬁning property is R-hereditary. and it is simpler to deﬁne the ancestral relation generally at once than to deﬁne it ﬁrst deﬁnitions. and x has the relation R to y. between A and Z. B. until the last. and the generalised theory of induction. This is merely for convenience. Take. These deﬁnitions can all be applied to any other relation just as well as to N. are due to Frege. Thus if R is any relation whatever.” but until Frege developed his generalised theory of induction. of course. − . of whom B is a child of A. III. obvious at once that “ancestor” must be capable of deﬁnition in terms of “parent. but it will not be so according to any deﬁnition which will give the kind of idea that we wish to deﬁne. such a series as the following:— −.

It follows that such proofs can be applied to the natural numbers. Some believed it to be really a case of induction. but as a purely verbal proposition. not a principle. will be found in the long run to economise thought and increase logical power.e. We deﬁne the “natural numbers” as those to which proofs by mathematical induction can be applied. in the sense in which that word is used in logic. something of a mystery. Mathematical induction aﬀords. When the train is very long. though there Science and Method. the essential characteristic by which the ﬁnite is distinguished from the inﬁnite. III. it is a very long time before the last truck moves. and then extend it to other cases. every truck would begin to move sooner or later if the engine persevered. If “quadrupeds” are deﬁned as animals having four legs. i. We now know that all such views are mistaken. The principle of mathematical induction might be stated popularly in some such form as “what can be inferred from next to next can be inferred from ﬁrst to last. but no one quite knew why it was valid. as constantly elsewhere. and the time would never come when the whole train would be in motion. The use of mathematical induction in demonstrations was. not in virtue of any mysterious intuition or axiom or principle.) to which it cannot be applied. not otherwise.” This is true when the number of intermediate steps between ﬁrst and last is ﬁnite. iv. There seemed no reasonable doubt that it was a valid method of proof. more than anything else. We shall use the phrase “inductive numbers” to mean the same set as we have hitherto spoken of as the “natural numbers. as those that possess all inductive properties. there would be an inﬁnite succession of jerks. Poincar´ considered it to be a principle e of the utmost importance. There are some numbers to which it can be applied. is an instance of the smallest of inﬁnites). Here. Finitude and Mathematical Induction only for the case of the relation of n to n + . Nevertheless. and the case of numbers that obey mathematical induction is exactly similar. and that mathematical induction is a deﬁnition. . Anyone who has ever | watched a goods train beginning to move will have noticed how the impulse is communicated with a jerk from each truck to the next. and there are others (as we shall see in Chapter VIII.Chap. if there were a series of trucks no longer than the series of inductive numbers (which. by means of which an inﬁnite number of syllogisms could be condensed into one argument.” The phrase “inductive numbers” is preferable as aﬀording a reminder that the deﬁnition of this set of numbers is obtained from mathematical induction. until at last even the hindmost truck is in motion. generality from the ﬁrst. as we shall see. it will follow that animals that have four legs are quadrupeds. If the train were inﬁnitely long. chap. in the past. though it may | require more thought at the start.

Finitude and Mathematical Induction would always be other trucks further back which had not yet begun to move. When we come to inﬁnite numbers. where arguments from mathematical induction will be no longer valid. the almost unconscious use that is made of mathematical induction where ﬁnite numbers are concerned.Chap. . by contrast. and its connection with ﬁnitude. This image will help to elucidate the argument from next to next. III. the properties of such numbers will help to make clear.

then all the multiples of but not of or or . then all the even numbers. the ﬁrst thing to realise is that no set of terms has just one order to the exclusion of others. We might. and this is essential to most of their mathematical properties. Dimensions. which underlies all higher mathematics. . In seeking a deﬁnition of order. but also rational fractions and all real numbers have an order of magnitude. of the whole class of its members. The conception of a limit. and of the relation of a number to its immediate successor. that is an inaccurate expression: . . but they are very few in comparison with the parts in which this notion is involved. in geometry. The notion of order is one which has enormous importance in mathematics. are a development of order. but they are capable of an inﬁnite number of other arrangements. as we shall also call them—occur to us most readily in order of magnitude. then all the multiples of but not of or . We must now consider the serial character of the natural numbers in the order . There are parts of mathematics which do not depend upon the notion of order. . . and it is an essential part of the work of analysing our data to seek a deﬁnition of “order” or “series” in logical terms. We ordinarily think of the numbers as in this order. When we say that we “arrange” the numbers in these various orders. consider ﬁrst all the odd numbers and then all the even numbers. for example. The order of points on a line is essential to geometry. or of planes through a line. is a serial conception. or ﬁrst . The natural numbers—or the “inductive” numbers. and so on through the whole series of primes. so is the slightly more complicated order of lines through a point in a plane. . A set of terms has all the orders of which it is capable.CHAPTER IV THE DEFINITION OF ORDER We have now carried our analysis of the series of natural numbers to the point where we have obtained logical deﬁnitions of the members of this series. then all the odd multiples of . Sometimes one order is so much more familiar and natural to our | thoughts that we are inclined to regard it as the order of that set of terms. but this is a mistake. Not only the integers. .

left of. The Deﬁnition of Order what we really do is to turn our attention to certain relations between the natural numbers. whether we choose to notice it or not. which themselves generate such-and-such an arrangement. earlier. then all those that have algebraic non-rational co-ordinates. One important result of this consideration is that we must not look for the deﬁnition of order in the nature of the set of terms to be ordered. when the relation holds between x and y. x must precede z. y is of a diﬀerent height from x. A relation having this ﬁrst property is called asymmetrical. so there are various relations among numbers which may be observed. y is not also earlier than x. not in the class of terms. In all these cases. If x is earlier in time than y. but others are equally valid. of any two terms in the class which is to be ordered. If x is a brother or sister of y. relations which do not give rise to series often do not have this property. it also holds between y and x. But with serial relations such a thing cannot happen. But . This is an obvious characteristic of the kind of relations that lead to series. but in a relation among | the members of the class. on a line.Chap. y is not also less than x. y is a brother or sister of x. This may be illustrated by the same instances as before: less. for example. On the other hand.” Now. y is of the same height as x. If x is to the left of y. that one “precedes” and the other “follows. () If x precedes y and y precedes z. through any set of complications we please. all the points that have integral co-ordinates. If x is less than y. What properties must a relation have in order to give rise to an order? The essential characteristics of a relation which is to give rise to order may be discovered by considering that in respect of such a relation we must be able to say. IV. we require that the ordering relation should have three properties:— () If x precedes y. If x is of a diﬀerent height from y. in order that we may be able to use these words in the way in which we should naturally understand them. y must not also precede x. in respect of which some appear as earlier and some as later. and which give rise to various diﬀerent orders among numbers. the only thing that is arbitrary about the various orders of a set of terms is our attention. then all those that have non-integral rational co-ordinates. for the terms themselves have always all the orders of which they are capable. The fact that a class may have many orders is due to the fact that there can be many relations holding among the members of one single class. take ﬁrst. If x is of the same height as y. We might. since one set of terms has many orders. We can no more “arrange” the natural numbers than we can the starry heavens. all equally legitimate. but just as we may notice among the ﬁxed stars either their order of brightness or their distribution in the sky. and so on. And what is true of numbers is equally true of points on a line or of the moments of time: one order is more familiar. The order lies. The resulting order will be one which the points of the line certainly have. y is not to the left of x.

” “diﬀerent in size.” “dear friend” are not. The Deﬁnition of Order as instances of relations which do not have this property only two of our previous three instances will serve. of any two integers.” and so on. Of two points on a line. and wherever an order exists.” “born of the same parents. or to be contained in or imply diversity. there must be one which precedes and the other which follows.” “greater by ” is the square of “greater by . S. and the converse domain consists of all those terms to which something or other has the relation.” “father” are aliorelatives. but are recalled here for the sake of the following deﬁnition:— () The ﬁeld of a relation consists of its domain and converse domain together. | () One relation is said to contain or be implied by another if it holds whenever the other holds. () The square of a relation is that relation which holds between two terms x and z when there is an intermediate term y such that the given relation holds between x and y and between y and z. Thus “paternal grandfather” is the square of “father. though an This term is due to C. but of any two complex numbers this is not true. x may not be brother or sister of z. one is smaller and the other greater. The same applies to diﬀerence of height. Of any two moments in time.” “husband. () Given any two terms of the class which is to be ordered. but of events. which has our second property but not our ﬁrst. A relation having this third property is called connected.Chap. has our ﬁrst property but not | our second.” on the other hand. A relation having our second property is called transitive. we will introduce a few deﬁnitions. These words have been already deﬁned. if no term has this relation to itself. this cannot be said. for example. Before illustrating this thesis. Peirce. IV. and y of z. one must be to the left of the other. For example. It often happens that a relation is an aliorelative without being asymmetrical. . () A relation is said to be an aliorelative. The relation “father. one must be earlier than the other. but not to sameness of height.” “brother. If x is brother or sister of y. since x and z may be the same person. It will be seen that an asymmetrical relation is the same thing as a relation whose square is an aliorelative. but “equal. or fractions. When a relation possesses these three properties. some relation having these three properties can be found generating it. or real numbers. () The domain of a relation consists of all those terms that have the relation to something or other. it is of the sort to give rise to an order among the terms between which it holds. Thus. which may be simultaneous. “greater.

. when it is asymmetrical. one whose square implies both it and diversity—because. tive. but not asymmetrical or an aliorelative. since a relation may have any two without having the third. . IV. “contains” its square. as we also say.” for example. It will be seen that the relation “ancestor. . . but not connected. is transitive and connected. are six diﬀerent series which all have the same ﬁeld. A transitive aliorelative is one which contains its square and is contained in diversity. . . . but is not transitive. but is symmetrical. but “father” is not transitive. . What distinguishes the above six series is simply the diﬀerent ordering relations in the six cases. But this would be an error. The relation “less than or equal to. it is because it is not connected that it does not suﬃce to arrange the human race in a series. . given any two diﬀerent terms of its ﬁeld. but the ﬁeld cannot be so taken. transitive. . because an ancestor’s ancestor is an ancestor. () | transi. not the serial relation itself. . From the deﬁnitions it will be seen that a transitive relation is one which is implied by its square.Chap. all aliorelatives are asymmetrical as well as vice versa. . The relation “greater or less” among numbers is an aliorelative and is connected. A series is the same thing as a serial relation. transitive. though both cannot happen if the relation is asymmetrical). The Deﬁnition of Order asymmetrical relation is always an aliorelative.” among numbers. is an aliorelative and transitive. it may happen that x and z are the same number. when a relation is transitive. and connected. what is equivalent. It might have been thought that a series should be the ﬁeld of a serial relation. Thus “ancestor” is transitive. and y is greater or less than z. and connected. or. the ﬁeld and the order are both determinate. A relation is connected when. what comes to the same thing. . or. For example. Thus the three properties of being () an aliorelative. the relation holds between the ﬁrst and the second or between the second and the ﬁrst (not excluding the possibility that both may happen. But among transitive relations. . “spouse” is an aliorelative. Thus the ordering relation may be taken to be the series. . or. are mutually independent. Given the ordering relation. . . for if x is greater or less than y. y is the spouse of x. and () connected. We now lay down the following deﬁnition:— A relation is serial when it is an aliorelative. asymmetry is equivalent to being an aliorelative. For example. . If the ﬁeld were the series. since if x is the spouse of y. because a father’s father is not a father. there could only be one series with a given ﬁeld.

the relation between consecutive integers. () P must imply P. The reader can easily convince himself that. Although a transitive asymmetrical connected relation always exists wherever there is a series. The generation of series by means of relations more or less resembling that of n to n + is very common. It is easy to see. we shall lay down the following deﬁnition:— An inductive number m is said to be less than another number n when n possesses every hereditary property possessed by the successor of m. We are therefore justiﬁed in taking the above as a deﬁnition of order | or series. x must precede z. This is probably the easiest way. i. In this method we pass on from . that the relation “less than. and vice versa. For purposes of generating the series of natural numbers. where it is applicable. is generated by relations of each to his successor. the characteristics we expect of series will also be found. This relation is asymmetrical. The three characteristics which P must have in order to be serial are: () We must never have xPx. by the method of mathematical induction. or order of magnitude. where these three properties are found in an ordering relation. however.e. The natural-number series may serve as an illustration.e. we shall have xPy or yPx.Chap. and connected. transitive.e. and has the inductive numbers for its ﬁeld. of conceiving the generation of a series. The Deﬁnition of Order Given any serial relation. for example. and not diﬃcult to prove. i. we shall say that.e. IV. We can. in respect of this relation. but not transitive or connected. the “ancestral” relation which we considered in the preceding chapter. i. we want the relation “less than.” and this order is the so-called “natural” order. which we shall write “xPy” for short. This relation will be the same as “less than or equal to” among inductive integers. The series of the Kings of England. say P. i. it is not always the relation which would most naturally be regarded as generating the series. one of the two must precede the other. Thus by means of this relation the inductive numbers acquire an order in the sense in which we deﬁned the term “order. And it will be observed that the deﬁnition is eﬀected in purely logical terms.” so deﬁned.” excluding “equal to. or (what comes to the same thing) when the successor of m is an ancestor of n in the sense in which a number is its own ancestor. no term must precede itself. x “precedes” y if x has the relation P to y. if x precedes y and y precedes z. The relation we assumed in considering the natural numbers was the relation of immediate succession.” This is the relation of m to n when m is an ancestor of n but not identical with n. is asymmetrical. derive from it. () If x and y are two diﬀerent terms in the ﬁeld of P. That is to say.

g. We deﬁne further: A term x is a “proper ancestor” of y with respect to R if y belongs to the proper posterity of x with respect to R.” let us give the name “proper posterity of x with respect to R” to the class of those terms that belong to the R-posterity of some term to which x has the relation R. Under what circumstances will this occur? It will always be transitive: no matter what sort of relation R may be. if this method is to be possible. transitive. Reverting now to the generation of series by the relation R between consecutive terms. This includes everybody at the table. We shall speak for short of “R-posterity” and “R-ancestors” when these terms seem more convenient. we see that. On the analogy of “proper fractions. the relation “proper R-ancestor” must be an aliorelative. If we call this relation R. and connected. since | twelve steps bring us back to our starting-point. The above was an instance in which the ancestral relation was connected but not contained in diversity. This method always requires the generalised form of mathematical induction in order to enable us to deﬁne “earlier” and “later” in a series so generated. but also in cases (as e. The Deﬁnition of Order each term to the next. It is to be observed that this deﬁnition has to be so framed as to be applicable not only when there is only one term to which x has the relation R. and though R itself is an aliorelative. An instance where it is contained in diversity but not connected is derived from the ordinary sense of the word “ancestor. But it is only under certain circumstances that it will be an aliorelative or connected. we ﬁnd that the “proper posterity” may be deﬁned as follows:— The “proper posterity” of x with respect to R consists of all terms that possess every R-hereditary property possessed by every term to which x has the relation R. It is for this reason that we cannot say that one person comes before another with respect to the relation “right of” or to its ancestral derivative.” which includes a term in its own posterity. that of father and child) where there may be many terms to which x has the relation R. in the sense which we gave before to “posterity. Reverting to the fundamental deﬁnitions. Thus in such a case. the proper R-posterity of a person consists of all who can be reached by going round the table from right to left. x and y cannot be the same person. but it is not true that of any two persons one must be an ancestor of the other.” If x is a proper ancestor of y. as long as there | is a next. as long as there is one before. IV. or back to the one before. for example. Consider. including the person himself.Chap. “Rancestor” and “proper R-ancestor” are always both transitive. the relation to one’s left-hand neighbour at a round dinner-table at which there are twelve people. . we do not get a series because “proper R-ancestor” is not an aliorelative. though the relation “proper R-ancestor” is connected.

it may even be extended to certain inﬁnite series. Between any two fractions there are others—for example. If we depended | upon consecutiveness for deﬁning order. but all depend upon the ﬁnding or construction of an asymmetrical transitive connected relation. In all such cases the order must be deﬁned by means of a transitive relation. those in which. space. fractions in order of magnitude.Chap. it is best introduced in connection with elementary geometry. the relation “proper R-ancestor” must be connected. for instance. time. the arithmetic mean of the two. And such series are of vital importance for the understanding of continuity. Here again. IV. It often happens in a series that there are an inﬁnite number of intermediate terms between any two that may be selected. the number of terms between any two is always ﬁnite. and motion. therefore all that remains to ensure its being serial is that it shall be contained in diversity. like that of counting for discovering the number of a collection. though important in its own sphere. But in fact the relations of greater and less among fractions do not demand generation from relations of consecutiveness. though the total number of terms is inﬁnite. the one condition required to secure the generation of a series is that the relation “proper R-ancestor” shall be contained in diversity. but care must be taken to eradicate from the imagination all habits of thought resulting from supposing it general. Another generalisation consists in taking R to be a one-one relation. Some of these ways have considerable importance. is less general than the method which uses a transitive relation to deﬁne the order. The generation of order by means of relations of consecutiveness. is appropriate to the ﬁnite. and the relations of greater and less among fractions have the three characteristics which we need for deﬁning serial relations. Take.” This method is very useful in geometry. Some of the most important cases are the following: Let R be a many-one relation. When so conﬁned. If this is not done. since only such a relation is able to leap over an inﬁnite number of intermediate terms. but it must not be regarded as general. however near together these may be. . and including the ancestry of x as well as the posterity. Not only so. Consequently there is no such thing as a pair of consecutive fractions. and may serve as an introduction to relations having more than two terms. We may take as illustrative the generation of series by means of a three-term relation which we may call “between. and let us conﬁne our attention to the posterity of some term x. we should not be able to deﬁne the order of magnitude among fractions. The method of consecutiveness. series in which there are no consecutive terms will remain diﬃcult and puzzling. This is a generalisation of the instance of the dinner-table. The Deﬁnition of Order The question of the circumstances under which series can be generated by ancestral relations derived from relations of consecutiveness is often important. namely. There are many ways in which series may be generated.

In fact. IV. namely. then either x and y are identical. then either x and y are identical. For the sake of deﬁniteness. iv. because. Rivista di Matematica. on the line (ab).. the notion “between” is characteristic of open series—or series in the strict sense— as opposed to what may be called | “cyclic” series. the following:— () If anything is between a and b. or x is between b and y. or x is between a and y. y.” In order that this relation “between” may arrange the points of the line in an order from left to right. () If b is between a and x and also between a and y. let us assume | that a is to the left of b. () Points x such that a is between x and b. a lies—these we will call to the left of a. (§). the line (ab) consists of three parts (besides a and b themselves): () Points between a and b. Thus the line (ab) can be deﬁned in terms of the relation “between. () If x is between a and b. as may be seen from the following deﬁnitions. but for the present we will only consider its application to a single straight line and to the ordering of the points on a straight line. () those between a and b. we can travel from any one to any other without passing through the third. () b itself. given any three points on a circle. We may now deﬁne generally that of two points x. we need certain assumptions. or y is between b and x. consequently. a suﬃcient journey brings us back to our starting-point. Any three-term relation which veriﬁes them gives rise to series. a and b are not identical. we shall say that x is “to the left of” y in any of the following cases:— Cf. pp. Taking any two points a. () those between which and a lies b—these we will call to the right of b. () Anything between a and b is also between b and a. anything between a and x is also between a and b. This notion of “between” may be chosen as the fundamental notion of ordinary geometry. in virtue of ()). ﬀ.Chap. and b is between x and y. then b is between a and y. with b. as with people at the dinner-table. there must be one of them which is between the other two. where. or x is between y and b. . p. Principles of Mathematics. () a itself. () Points y such that b is between y and a. b. This will not be the case with the points on a circle or any other closed curve. Then the points of the line (ab) are () those between which and b. () If x is between a and b. () If x and y are between a and b. () Anything between a and b is not identical with a (nor. The Deﬁnition of Order Given any three points on a straight line in ordinary space. These seven properties are obviously veriﬁed in the case of points on a straight line in ordinary space.

cannot be generated by means of three-term relations of “between. the generation of them from other relations possessing only some of the properties required for series becomes very important.” When such relations have been deﬁned. and y is between x and b. y). in order to get from a to b one must pass through either x or y.” The point may be illustrated by considering a journey round the world. if we take any four points on a circle.Chap. we cannot | say deﬁnitely that either of these two places is “between” England and New Zealand. and y is between a and b or is b or is to the right of b. Out of this relation a cyclic order can be generated. The Deﬁnition of Order () When x and y are both to the left of a. and y is b or to the right of b.” We need a relation of four terms. in a way resembling that in which we generated an open order from “between. say a and b and x and y. Generalising.” it can be deduced that the relation “to the left of. such as that of the points on a circle. and in order to get from x to y one must pass through either a or b. p. his times in England and New Zealand are separated from each other by his times in Suez and San Francisco. () When x and y are both to the right of b and x is between b and y. (§).” as above deﬁned. But if a man chooses that route to go round the world. Under these circumstances we say that the couple (a. () When x is a. It is important to notice that nothing in the deﬁnitions or the argument depends upon our meaning by “between” the actual relation of that name which occurs in empirical space: any three-term relation having the above seven purely formal properties will serve the purpose of the argument equally well. is a serial relation as we deﬁned that term. Principles of Mathematics. we can separate them into two couples. and references there given. . () When x is to the left of a. and conversely. and y is between x and a. () When x is b and y is to the right of b. especially in the philosophy of geometry Cf. () When x and y are both between a and b. from the seven properties which we have assigned to the relation “between. The purpose of the latter half of this chapter has been to suggest the subject which one may call “generation of serial relations. b) are “separated” by the couple (x. which may be called “separation of couples. One may go from England to New Zealand by way of Suez or by way of San Francisco. whichever way round he goes. and y is a or b or between a and b or to the right of b.” but somewhat more complicated. () When x is between a and b. such that. Cyclic order. IV. It will be found that.

IV.Chap. within the limits of the present volume. But we cannot. The Deﬁnition of Order and physics. do more than make the reader aware that such a subject exists. .

one term of the relation belongs to the class of males and one to the class of females. i. in the case of spouse. without the help of any further relation.e. We see from this instance that. and so is the relation wife. Asymmetry. But the cases where this is possible are rare and exceptional: they are cases where there are two mutually exclusive classes. Suppose now we are given the relation spouse. namely. one of the terms is a member of α and the other is a member of β—as. i. is a characteristic of the very greatest interest and importance. if a is husband of b. as well as from the intrinsic interest of the subject. the property of being incompatible with the converse. and connexity—has its own importance. On the other hand. serial relations. to separate it into two asymmetrical relations. in these cases the reader will not see the bearing of the proposition asserting such a property unless he has in mind the sorts of relations for which it is useful. when a symmetrical relation is given.CHAPTER V KINDS OF RELATIONS A great part of the philosophy of mathematics is concerned with relations. asymmetry. say α and β. For reasons of this description. We dealt in the preceding chapter with a supremely important class.e. thus the relation husband can | be derived from spouse either by limiting the domain to males or by limiting the converse domain to females. such that whenever the relation holds between two terms. it is sometimes possible. We will begin by saying something on each of these three. Each of the three properties which we combined in deﬁning series—namely. It often happens that a property which belongs to all relations is only important as regards relations of certain sorts. then b is spouse of a. and many diﬀerent kinds of relations have diﬀerent kinds of uses. b cannot be husband of a. In order to develop its functions. we will consider various examples. it is well to have in our minds a rough list of the more mathematically serviceable varieties of relations. The relation husband is asymmetrical. and we wish to derive the relation husband. In such a . Husband is the same as male spouse or spouse of a female. the relation “spouse” is symmetrical: if a is spouse of b. and similarly in the case of wife. transitiveness.

Chap. V. Kinds of Relations

case, the relation with its domain conﬁned to α will be asymmetrical, and so will the relation with its domain conﬁned to β. But such cases are not of the sort that occur when we are dealing with series of more than two terms; for in a series, all terms, except the ﬁrst and last (if these exist), belong both to the domain and to the converse domain of the generating relation, so that a relation like husband, where the domain and converse domain do not overlap, is excluded. The question how to construct relations having some useful property by means of operations upon relations which only have rudiments of the property is one of considerable importance. Transitiveness and connexity are easily constructed in many cases where the originally given relation does not possess them: for example, if R is any relation whatever, the ancestral relation derived from R by generalised induction is transitive; and if R is a many-one relation, the ancestral relation will be connected if conﬁned to the posterity of a given term. But asymmetry is a much more diﬃcult property to secure by construction. The method by which we derived husband from spouse is, as we have seen, not available in the most important cases, such as greater, before, to the right of, where domain and converse domain overlap. In all these cases, we can of course obtain a symmetrical relation by adding together the given relation and its converse, but we cannot pass back from this symmetrical relation to the original asymmetrical relation except by the help of some asymmetrical | relation. Take, for example, the relation greater: the relation greater or less—i.e. unequal—is symmetrical, but there is nothing in this relation to show that it is the sum of two asymmetrical relations. Take such a relation as “diﬀering in shape.” This is not the sum of an asymmetrical relation and its converse, since shapes do not form a single series; but there is nothing to show that it diﬀers from “differing in magnitude” if we did not already know that magnitudes have relations of greater and less. This illustrates the fundamental character of asymmetry as a property of relations. From the point of view of the classiﬁcation of relations, being asymmetrical is a much more important characteristic than implying diversity. Asymmetrical relations imply diversity, but the converse is not the case. “Unequal,” for example, implies diversity, but is symmetrical. Broadly speaking, we may say that, if we wished as far as possible to dispense with relational propositions and replace them by such as ascribed predicates to subjects, we could succeed in this so long as we conﬁned ourselves to symmetrical relations: those that do not imply diversity, if they are transitive, may be regarded as asserting a common predicate, while those that do imply diversity may be regarded as asserting incompatible predicates. For example, consider the relation of similarity between classes, by means of which we deﬁned numbers. This relation is symmetrical and transitive and

Chap. V. Kinds of Relations

does not imply diversity. It would be possible, though less simple than the procedure we adopted, to regard the number of a collection as a predicate of the collection: then two similar classes will be two that have the same numerical predicate, while two that are not similar will be two that have diﬀerent numerical predicates. Such a method of replacing relations by predicates is formally possible (though often very inconvenient) so long as the relations concerned are symmetrical; but it is formally impossible when the relations are asymmetrical, because both sameness and diﬀerence of predicates are symmetrical. Asymmetrical relations are, we may | say, the most characteristically relational of relations, and the most important to the philosopher who wishes to study the ultimate logical nature of relations. Another class of relations that is of the greatest use is the class of one-many relations, i.e. relations which at most one term can have to a given term. Such are father, mother, husband (except in Tibet), square of, sine of, and so on. But parent, square root, and so on, are not one-many. It is possible, formally, to replace all relations by one-many relations by means of a device. Take (say) the relation less among the inductive numbers. Given any number n greater than , there will not be only one number having the relation less to n, but we can form the whole class of numbers that are less than n. This is one class, and its relation to n is not shared by any other class. We may call the class of numbers that are less than n the “proper ancestry” of n, in the sense in which we spoke of ancestry and posterity in connection with mathematical induction. Then “proper ancestry” is a one-many relation (one-many will always be used so as to include one-one), since each number determines a single class of numbers as constituting its proper ancestry. Thus the relation less than can be replaced by being a member of the proper ancestry of. In this way a onemany relation in which the one is a class, together with membership of this class, can always formally replace a relation which is not onemany. Peano, who for some reason always instinctively conceives of a relation as one-many, deals in this way with those that are naturally not so. Reduction to one-many relations by this method, however, though possible as a matter of form, does not represent a technical simpliﬁcation, and there is every reason to think that it does not represent a philosophical analysis, if only because classes must be regarded as “logical ﬁctions.” We shall therefore continue to regard one-many relations as a special kind of relations. One-many relations are involved in all phrases of the form “the so-and-so of such-and-such.” “The King of England,” | “the wife of Socrates,” “the father of John Stuart Mill,” and so on, all describe some person by means of a one-many relation to a given term. A person cannot have more than one father, therefore “the father of John Stuart Mill” described some one person, even if we did not know

Chap. V. Kinds of Relations

whom. There is much to say on the subject of descriptions, but for the present it is relations that we are concerned with, and descriptions are only relevant as exemplifying the uses of one-many relations. It should be observed that all mathematical functions result from onemany relations: the logarithm of x, the cosine of x, etc., are, like the father of x, terms described by means of a one-many relation (logarithm, cosine, etc.) to a given term (x). The notion of function need not be conﬁned to numbers, or to the uses to which mathematicians have accustomed us; it can be extended to all cases of one-many relations, and “the father of x” is just as legitimately a function of which x is the argument as is “the logarithm of x.” Functions in this sense are descriptive functions. As we shall see later, there are functions of a still more general and more fundamental sort, namely, propositional functions; but for the present we shall conﬁne our attention to descriptive functions, i.e. “the term having the relation R to x,” or, for short, “the R of x,” where R is any one-many relation. It will be observed that if “the R of x” is to describe a deﬁnite term, x must be a term to which something has the relation R, and there must not be more than one term having the relation R to x, since “the,” correctly used, must imply uniqueness. Thus we may speak of “the father of x” if x is any human being except Adam and Eve; but we cannot speak of “the father of x” if x is a table or a chair or anything else that does not have a father. We shall say that the R of x “exists” when there is just one term, and no more, having the relation R to x. Thus if R is a one-many relation, the R of x exists whenever x belongs to the converse domain of R, and not otherwise. Regarding “the R of x” as a function in the mathematical | sense, we say that x is the “argument” of the function, and if y is the term which has the relation R to x, i.e. if y is the R of x, then y is the “value” of the function for the argument x. If R is a one-many relation, the range of possible arguments to the function is the converse domain of R, and the range of values is the domain. Thus the range of possible arguments to the function “the father of x” is all who have fathers, i.e. the converse domain of the relation father, while the range of possible values for the function is all fathers, i.e. the domain of the relation. Many of the most important notions in the logic of relations are descriptive functions, for example: converse, domain, converse domain, ﬁeld. Other examples will occur as we proceed. Among one-many relations, one-one relations are a specially important class. We have already had occasion to speak of one-one relations in connection with the deﬁnition of number, but it is necessary to be familiar with them, and not merely to know their formal deﬁnition. Their formal deﬁnition may be derived from that of onemany relations: they may be deﬁned as one-many relations which are also the converses of one-many relations, i.e. as relations which are

It is obvious that x and z must be | the same person. V. Such correlations are simplest to grasp when the two classes have no members in common. term for term. for example. we take the relation of parent and child. to think of y as being reached from x by an “R-step” or an “R-vector.” In the same case x will be reached from y by a “backward R-step.Chap. because one may be the father of y and the other the mother. and the term relatum for the term to which it . Kinds of Relations both one-many and many-one. Thus. we can no longer argue that. Or. this is by no means the case. again. such that x is the father of y and y is the son of z. One-one relations give a correlation of two classes. where the “relative product” of two relations R and S is that relation which holds between x and z when there is an intermediate term y.g. x and z must be the same person. such that x has the relation R to y and y has the relation S to z. so that each term in either class has its correlate in the other.” It will be observed that the relative product of two relations is not in general commutative. for in that case we know at once whether a term is to be considered as one from which the correlating relation R goes. One-many relations may be deﬁned as relations such that. This illustrates that it is characteristic of one-many relations when the relative product of a relation and its converse implies identity. they may be deﬁned as follows: Given two terms x and x . With other relations. It is convenient to use the word referent for the term from which the relation goes. or as one to which it goes. E. there is no other term x which also has the relation to y. the relative product of R and S is not in general the same relation as the relative product of S and R. for example. which is not one-many. like the class of husbands and the class of wives. again.” Thus we may state the characteristic of one-many relations with which we have been dealing by saying that an R-step followed by a backward R-step must bring us back to our starting-point. and also the relative product of the converse and the relation implies identity. If. the relative product of R and its converse is the relation “self or brother or sister. they may be deﬁned as relations such that the relative product of one of them and its converse implies identity. In the case of one-one relations this happens. the terms to which x has the given relation and those to which x has it have no member in common. on the other hand. but the relative product of brother and parent is parent.e. if R is the relation of father to son.” and if R is the relation of grandchild to grandparent. if R is the relation of child to parent. if x is a parent of y and y is a child of z. it is convenient. the relative product of R and its converse will be the relation which holds between x and a man z when there is a person y. if x has the relation in question to y. Or. the relative product of R and its converse is “self or brother or sister or ﬁrst cousin. if x has the relation R to y. the relative product of parent and brother is uncle. i. Given a relation R.

Or. . It will be observed that the class of all possible referents to a given relation is its domain. . namely. . . . .” y is referent and x relatum. . . On this occasion only three of our original set are left. We may place the numbers concerned in two rows.Chap. . . . . . V. again. n + . we could have doubled each of them. n . . putting the correlate directly under the number whose correlate it is. . the ﬁrst ten integers (excluding ). . . we had considered the whole of the inductive numbers. The fact that a relation has a “sense” is fundamental. Thus when the correlator is the relation of n to n + . . . we have the two rows: . . of the domain. which is again a one-one relation. But it very often happens that the domain and converse domain of a one-one relation overlap. . then. . . thus obtaining the set . . . . Take. The most interesting case of the above kind is the case where our one-one relation has a converse domain which is part. and the class of all possible relata is its converse domain. thus obtaining the integers . thus the “sense” of a relation that goes from x to y is the opposite of that of the corresponding relation from y to x. . We say that a relation and its converse have opposite “senses”. . . instead of conﬁning the domain to the ﬁrst ten integers. . . . . . . . . the above instances would have illustrated this case. thus instead of the ﬁrst ten integers we now have the integers . except that has been cut oﬀ at the beginning and has been joined on at the end. . instead of adding to each of our original ten integers. . There are still ten integers: they are correlated with the previous ten by the relation of n to n + . . Such processes of correlation may be varied endlessly. for example. . .” x is referent and y relatum. namely. and is part of the reason why order can be generated by suitable relations. The correlating relation in this case is the relation of a number to its double. . . . . with respect to the relation | “husband. . . Or we might have replaced each number by its square. Thus if x and y are husband and wife. . . . Here we still have ﬁve of our previous set of integers. and add to each. . which is a one-one relation. but with respect to the relation “wife. . . Kinds of Relations goes. If. These are the same as those we had before. . but | not the whole.

. Another class of correlations which are often important is the class called “permutations. as we have partly seen already. c b a c b a | Each of these can be obtained from any one of the others by means of a correlation. . a part not the whole).e. . One of its uses will occupy us in our next chapter. n . . Consider. . .e. V. for example. we have the two rows: . . we wish only to note that they exist and demand consideration. c. . . . . . where the converse domain is a “proper part” of the domain (i. the ﬁrst and last. Here a is correlated with c. Kinds of Relations When the correlator is the relation of a number to its double. . a). . the rows are: .” These various kinds of correlations have importance in various connections. . for example. When the correlator is the relation of a number to its square. c. b. b. will occupy us again when we come to deal with inﬁnity. . b. . . Take. a. . . some for one purpose. . (a. . c. n . c) and (c. . For the present. the permutations of a given class form what is called a “group. the six possible arrangements of three letters: a.Chap. . b. and c with a. a. i. . a. . some for another.” where the domain and converse domain are identical. . . b. It is obvious that the combination of two permutations is again a permutation. b. but shall see much more fully as we proceed. . all inductive numbers occur in the top row. c. . and only some in the bottom row. . b with itself. . In all these cases. . n . The general notion of one-one correlations has boundless importance in the philosophy of mathematics. . n . Cases of this sort. . . . . .

since every class is the domain of some relation. to such relations as have “ﬁelds. The structure of the map corresponds with that of | the country of which it is a map.” i. In the present chapter we have to deﬁne a relation between relations. the place on the map corresponding to the one is to the left of the place on the map corresponding to the other. whenever either relation holds between two terms. for example. . when one place is west of another. the relation which the domain of a relation has to the relation. which will play the same part for them that similarity of classes plays for classes. The easiest example of the sort of thing we desire is a map. How is likeness to be deﬁned? We shall employ still the notion of correlation: we shall assume that the domain of the one relation can be correlated with the domain of the other. and the converse domain with the converse domain. when there is a one-one relation whose domain is the one class and whose converse domain is the other.e.CHAPTER VI SIMILARITY OF RELATIONS We saw in Chapter II. but that is not enough for the sort of resemblance which we desire to have between our two relations. to such as permit of the formation of a single class out of the domain and the converse domain. This is not always the case.e. We may. What we desire is that. in deﬁning likeness.e. Take. We will call this relation “similarity of relations.” i. proﬁtably introduce a certain restriction. the place on the map corresponding to the one is above the place on the map corresponding to the other. It is this kind of connection between relations that we wish to deﬁne. In such a case we say that there is a “one-one correlation” between the two classes. and so on. the relation “domain. The space-relations in the map have “likeness” to the space-relations in the country mapped. that two classes have the same number of terms when they are “similar. in the ﬁrst place. We will conﬁne ourselves. the other relation shall hold between the correlates of these two terms.” or “likeness” when it seems desirable to use a diﬀerent word from that which we use for classes. This relation has all classes for its domain.” i. When one place is north of another.

And having been stated. and is such that P is the relative product of S and Q and the converse of S. and vice versa.” We do not need to enter upon the diﬃcult doctrine of types. and the backward S-step from w to y. but it is well to know when we are abstaining from entering upon it. it is clear that for every instance in which the rez w lation P holds there is a corresponding Q instance in which the relation Q holds. Thus we may set up the following deﬁnitions:— A relation S is said to be a “correlator” or an “ordinal correlator” of two relations P and Q if S is one-one. If this happens with every pair of terms such as x and y. in deﬁning likeness.” i. that a relation only has a “ﬁeld” when it is what we call “homogeneous.e. when its domain and converse domain are of the same logical type. without entering upon the grounds for the assertion. the correlate of the one has the relation Q to the correlate of the other. We may deﬁne two relations P and Q as “similar. We may say. Let x and y be two terms having the relation P. relations between individuals. But classes and relations cannot be added together to form a new single class. Now the notion of likeness is not very useful as applied to relations that are not homogeneous. because they are of diﬀerent logical “types. we shall. y has the relation S to w. such that x has the relation x y P S to z. classes of individuals. and which is such that. therefore. by observing that. but the limitation is not of any practical importance. and as a rough-and-ready indication of what we mean by a “type. and vice versa. and z has the relation Q to w. A ﬁgure will make this clearer. it need no longer be remembered. and so on. Similarity of Relations and it has all relations for its converse domain.” or as having “likeness. . and this is what we desire to secure by our deﬁnition.” when there is a one-one relation S whose domain is the ﬁeld of P and whose converse domain is the ﬁeld of Q. simplify our problem by speaking of the “ﬁeld” of one of the relations concerned. the P-step from x to y may be replaced by the succession of the S-step from x to z. the Q-step from z to w.e. the relation P is the same as the relative product of S and Q and the converse of S.” we may say that individuals.Chap. relations of classes to individuals. S S and if the converse happens with every pair of terms such as z and w. w. VI. since every relation has a domain. Then there are to be two terms z. This somewhat limits the generality of our deﬁnition. relations between classes. if one term has the relation P | to another. when the above conditions are realised. has the ﬁeld of Q for its converse domain. We can eliminate some redundancies in the above sketch of a deﬁnition. are diﬀerent types. i.

Chap. so is the other. if one implies diversity. through all the general properties of relations. though they may not be true as they stand when applied to a similar relation. Our problem may be stated as follows:— Given some statement in a language of which we know the grammar and the syntax. the state of our knowledge of nature. if one is connected. so we may . Thus great importance attaches to the question: What are the possible meanings of a law expressed in terms of which we do not know the substantive meaning. For the present we will ignore the general question. their properties are the same except when they depend upon the ﬁelds being composed of just the terms of which they are composed. but not the vocabulary. an old-fashioned pair of terms) about the form of nature than about the matter. in the most advanced sciences. We know much more (to use. what we really know when we enunciate a law of nature is only that there is probably some interpretation of our terms which will make the law approximately true. but only the grammar and syntax? And this question is the one suggested above.” or to have “likeness. when two relations are similar.” when there is at least one correlator of P and Q. For instance. and what are the meanings of the unknown words that would make it true? The reason that this question is important is that it represents. an importance by no means adequately recognised hitherto. so is the other. so does the other. Just as we called the set of those classes that are similar to a given class the “number” of that class. Hence if one is serial. it is desirable to have a nomenclature which collects | together all the relations that are similar to a given relation. These deﬁnitions will be found to yield what we above decided to be necessary. and so on. they share all properties which do not depend upon the actual terms in their ﬁelds. It will be found that. Even statements involving the actual terms of the ﬁeld of a relation. what are the possible meanings of such a statement. the other is one-many | or one-one. are expressed in mathematical symbols—are more or less true of the world. for a moment. if one is transitive. much more nearly than might be supposed. Similarity of Relations Two relations P and Q are said to be “similar. Again. We know that certain scientiﬁc propositions—which. if one is one-many or one-one. We are led by such considerations to a problem which has. but we are very much at sea as to the interpretation to be put upon the terms which occur in these propositions. the subject of likeness itself must ﬁrst be further investigated. when two relations are similar. in mathematical philosophy. will always be capable of translation into statements that are analogous. which will occupy us again at a later stage. so is the other. Accordingly. Owing to the fact that. VI.

we will speak. of a “relation-number. The deﬁnition of a cardinal number. When we speak of “numbers” without qualiﬁcation. as we shall see when we come to consider inﬁnite series. Thus.” i. a relation-number is a class of relations consisting of all those relations that are similar to one member of the class. The relation-numbers applicable to series may be | called “serial numbers” (what are commonly called “ordinal numbers” are a sub-class of these). thus a ﬁnite serial number is determinate when we know the cardinal number of terms in the ﬁeld of a series having the serial number in question.Chap. what comes to the same thing. the relation-number of a series which has n terms is called the “ordinal” number n. it will be remembered. there is parallelism between cardinal and relation-numbers. Two ﬁnite series will have the same relationnumber when their ﬁelds have the same cardinal number of terms. of which we shall speak later. (There are also inﬁnite ordinal numbers. is as follows:— The “cardinal number” of a given class is the set of all those classes that are similar to the given class.) When the cardinal number of terms in the ﬁeld of a series is inﬁnite. but when a series is ﬁnite.” Thus cardinal numbers are the numbers appropriate to classes.” Thus we have the following deﬁnitions:— The “relation-number” of a given relation is the class of all those relations that are similar to the given relation. nor. When a series is inﬁnite. or. in this case. its relation-number. These include the ordinary integers of daily life. a series of (say) terms will have the same relation-number as any other series of ﬁfteen terms. But in order to avoid confusion with the numbers appropriate to classes. indeed an inﬁnite number of relation-numbers exist for one inﬁnite cardinal number. of course. “Relation-numbers” are the set of all those classes of relations that are relation-numbers of various relations. When it is necessary to speak of the numbers of classes in a way which makes it impossible to confuse them with relation-numbers.e.e. this cannot happen. the relationnumber of the series is not determined merely by the cardinal number. The most obvious application of relation-numbers is to series. but of them we shall speak in a later chapter. the same relation-number as a relation which is not serial. Similarity of Relations call the set of all those relations that are similar to a given relation the “number” of that relation. we shall call them “cardinal numbers. what we may call its “length. and also certain inﬁnite numbers. Two series may be regarded as equally long when they have the same relation-number. but will not have the same relation-number as a series of or terms. we are to be understood as meaning cardinal numbers. and only then—i. in the quite special case of ﬁnite series. If n is a ﬁnite cardinal number. VI. and a whole arithmetic of relation- . may vary without change in the cardinal number. We can deﬁne addition and multiplication for relation-numbers as well as for cardinal numbers.

Veblen that we may regard our whole space as the ﬁeld of a three-term between-relation. In the ﬁrst place. being not connected.e. We may illustrate this by the relations of the sort that may be called “between. all the propositions are the same. We there saw that. Thus the serial relation which is to be deﬁned as the sum of P and Q is not “P or Q” simply. We have already mentioned maps. Relation-arithmetic. and we might extend our thoughts from this illustration to geometry generally. Suppose. b. It must not be supposed. But it obeys the associative law. it consists of a and b together with all points x. If the system of relations by which a geometry is applied to a certain set of terms can be brought fully into relations of likeness with a system applying to another set of terms. but as applied to relation-numbers generally. but “P or Q” is not serial. Similarity of Relations numbers can be developed.” Given any two points. except for the fact that they are applied in one case to one set of terms and in the other to another. that we wish to deﬁne the sum of two non-overlapping series in such a way that the relation-number of the sum shall be capable of being deﬁned as the sum of the relation-numbers of the two series. though recent.” Assuming that P and Q do not overlap. It has been shown by O. The resulting arithmetic does not obey the commutative law: the sum or product of two relation-numbers generally depends upon the order in which they are taken. but “P or Q or the relation of any member of the ﬁeld of P to any member of the ﬁeld of Q. The manner in which this is to be done is easily seen by considering the case of series. in fact. and may be called a “betweenrelation. and two of the formal laws for powers. then the geometry of the two sets is indistinguishable from the mathematical point of view. that there are no other applications that are important. since it does not hold between a member of the ﬁeld of P and a member of the ﬁeld of Q. such that the between-relation holds between the three points a. Thus the sum of P and Q. Thus if P and Q are the generating relations of the two series. VI. x in some order or other. every member of the ﬁeld of P will precede every member of the ﬁeld of Q. not only as applied to serial numbers. is what we need in order | to deﬁne the sum of two relation-numbers. we can use the between-relation to deﬁne the straight line determined by those two points. Similar modiﬁcations are needed for products and powers. as above deﬁned. this relation is serial. it will give rise to series.Chap. merely because series aﬀord the most obvious application of the idea of likeness. one form of the distributive law.” which we considered in Chapter IV. in the series which is their sum with P put before Q. for example. and deﬁne our geometry . is a thoroughly respectable branch of mathematics. provided a three-term relation has certain formal logical properties. it is clear that there is an order involved as between the two series: one of them must be placed before the other. i.

out of empirical material. “man” and “featherless biped”—so two relations which are conceptually diﬀerent may hold in the same set of instances.” We must not say that there is nothing else that could legitimately be called a “point”. Just as a class may be deﬁned by various diﬀerent but co-extensive concepts—e. of two similar relations. no matter how complicated. W. We may say. if B is like B in this sense. is not the intrinsic nature of our terms. any of which would be suﬃcient. even when he is speculating as an applied mathematician. in so far as geometry is not a matter of deﬁnition. It has to be something that as nearly as possible satisﬁes our axioms. not its intrinsic nature.Chap. If B and B are two between-relations.” we shall call S a correlator of B and B if it has the ﬁeld of B for its converse domain. but only to spaces in which the straight line is an open series. but the logical nature of their interrelations. that structure may legitimately be called a “point. of course. such that its ﬁrst term has the relation in does not apply to elliptic space. VI. so long as it satisﬁes the axioms. but it does not have to be “very small” or “without parts. Modern Mathematics.” This is only an illustration of the general principle that what matters in mathematics.” Whether or not it is those things is a matter of indiﬀerence. This . that they have the same | “structure. it may be one of many objects. And we shall say that B is like B when there is at least one correlator of B with B . so that one of the terms comes ﬁrst and the other second. we must only say: “This object we have constructed is suﬃcient for the geometer. construct a logical structure. It follows from this that the mathematician need not concern himself with the particular being or intrinsic nature of his points.g. The reader can easily convince himself that. Young. so that “xB(y. lines. the couple is to be. since this object is enough to vindicate the empirical truth of geometry. and planes. If we can.” For mathematical purposes (though not for those of pure philosophy) the only thing of importance about a relation is the cases in which it holds. An “instance” in which a relation holds is to be conceived as a couple of terms. with an order. – (monograph by O. We may say that there is empirical evidence of the approximate truth of such parts of geometry as are not matters of deﬁnition. edited by J. Veblen on “The Foundations of Geometry”). z)” means “x is between y and z with respect to B. Similarity of Relations by the properties we assign to our between-relation. but that is no concern of ours. and to a very great extent in physical science. which will satisfy our geometrical axioms. and is such that the relation B holds between three terms when B holds between their S-correlates. and only then. pp. A. But there is no empirical evidence as to what a “point” is to be. there can be no diﬀerence between the geometry generated by B and that generated by B . Now likeness is just as easily | deﬁnable between three-term relations as between two-term relations.

Speaking generally. bc. Take (say) the relation “father”: we can deﬁne what we may call the “extension” of this relation as the class of all ordered couples (x. For the sake of definiteness.” That is to say. c. if it is a suﬃciently simple one.” Given any relation. but they have objective counterparts. where a. e are ﬁve terms. Thus what we deﬁned as the “relation-number” is the very same thing as is obscurely intended by the word “structure”— a word which. is the very same thing as what we have called “likeness. we can. d The ﬁeld may be changed without changing the structure. when either can be a map for the other (since every relation can be its own map). We may make a “map” of this a b relation by taking ﬁve points on a plane and connecting them by arrows. we say: The “extension” of a relation is the class of those ordered couples (x. From the mathematical point of view. and the structure may be e changed without changing the ﬁeld—for | example. b. In actual .” we shall say. Where such hypotheses are made. and the diﬃculty of getting behind it. two relations have the same structure when they have likeness. There has been a great deal of speculation in traditional philosophy which might have been avoided if the importance of structure. as a moment’s reﬂection shows. but are caused by things in themselves. let us take a relation of which the extension is the following couples: ab. Similarity of Relations question to its second. important as it is. had been realised. y) which are such that x has the relation in question to y. or that phenomena are subjective. it is generally supposed that we can know very little about the objective counterparts. the only thing of importance about the relation “father” is that it deﬁnes this set of ordered couples. ce. i.Chap. d. when the same map will do for both—or. For example. which must have diﬀerences inter se corresponding with the diﬀerences in the phenomena to which they give rise. as in the accompanying ﬁgure. Two relations have the same “structure. what comes to the same thing. y) which are such that x is the father of y.e. it is often said that space and time are subjective. when they have the same relation-number. It is clear that the “structure” of the relation does not depend upon the particular c terms that make up the ﬁeld of the relation. We can now go a step further in the process of abstraction. if we were to add the couple ae in the above illustration we should alter the structure but not the ﬁeld. ad. no matter what. is never (so far as we know) deﬁned in precise terms by those who use it. construct a map of it. What is revealed by the map is what we call the “structure” of the relation. de. and consider what we mean by “structure. And that. VI. dc. ac.

so must the world behind phenomena. but which. We can all sympathise with their wish to prove such a very desirable proposition. it will bring about too much of a rapprochement between the real and the phenomenal world. if the hypotheses as stated were correct. as a rule. Similarity of Relations fact. probably because they feel instinctively that. if the phenomenal world is Euclidean. is irrelevant to science. If the phenomenal world has three dimensions. but we cannot congratulate them on their success. the objective counterparts would form a world having the same structure as the phenomenal world. every proposition having a communicable signiﬁcance must be true of both worlds or of neither: the only diﬀerence must lie in just that essence of individuality which always eludes words and baﬄes description. they could hardly avoid the conclusions which we have been suggesting. however. and these escape from the above argument. . if pursued. Now the only purpose that philosophers | have in view in condemning phenomena is in order to persuade themselves and others that the real world is very diﬀerent from the world of appearance. It is true that many of them do not assert objective counterparts to phenomena. for that very reason. Those who do assert counterparts are. the notion of structure or relation-number is important. and allowing us to infer from phenomena the truth of all propositions that can be stated in abstract terms and are known to be true of phenomena. as well as in many others. and so on.Chap. If they were to pursue the topic. so must the other be. In short. VI. In such ways. very reticent on the subject.

CHAPTER VII

**RATIONAL, REAL, AND COMPLEX NUMBERS
**

We have now seen how to deﬁne cardinal numbers, and also relation- numbers, of which what are commonly called ordinal numbers are a particular species. It will be found that each of these kinds of number may be inﬁnite just as well as ﬁnite. But neither is capable, as it stands, of the more familiar extensions of the idea of number, namely, the extensions to negative, fractional, irrational, and complex numbers. In the present chapter we shall brieﬂy supply logical deﬁnitions of these various extensions. One of the mistakes that have delayed the discovery of correct deﬁnitions in this region is the common idea that each extension of number included the previous sorts as special cases. It was thought that, in dealing with positive and negative integers, the positive integers might be identiﬁed with the original signless integers. Again it was thought that a fraction whose denominator is may be identiﬁed with the natural number which is its numerator. And the irrational numbers, such as the square root of , were supposed to ﬁnd their place among rational fractions, as being greater than some of them and less than the others, so that rational and irrational numbers could be taken together as one class, called “real numbers.” And when the idea of number was further extended so as to include “complex” numbers, i.e. numbers involving the square root of −, it was thought that real numbers could be regarded as those among complex numbers in which the imaginary part (i.e. the part | which was a multiple of the square root of −) was zero. All these suppositions were erroneous, and must be discarded, as we shall ﬁnd, if correct deﬁnitions are to be given. Let us begin with positive and negative integers. It is obvious on a moment’s consideration that + and − must both be relations, and in fact must be each other’s converses. The obvious and suﬃcient deﬁnition is that + is the relation of n + to n, and − is the relation of n to n + . Generally, if m is any inductive number, +m will be the relation of n + m to n (for any n), and −m will be the relation of

Chap. VII. Rational, Real, and Complex Numbers

n to n + m. According to this deﬁnition, +m is a relation which is one-one so long as n is a cardinal number (ﬁnite or inﬁnite) and m is an inductive cardinal number. But +m is under no circumstances capable of being identiﬁed with m, which is not a relation, but a class of classes. Indeed, +m is every bit as distinct from m as −m is. Fractions are more interesting than positive or negative integers. We need fractions for many purposes, but perhaps most obviously for purposes of measurement. My friend and collaborator Dr A. N. Whitehead has developed a theory of fractions specially adapted for their application to measurement, which is set forth in Principia Mathematica. But if all that is needed is to deﬁne objects having the required purely mathematical properties, this purpose can be achieved by a simpler method, which we shall here adopt. We shall deﬁne the fraction m/n as being that relation which holds between two inductive numbers x, y when xn = ym. This deﬁnition enables us to prove that m/n is a one-one relation, provided neither m nor n is zero. And of course n/m is the converse relation to m/n. From the above deﬁnition it is clear that the fraction m/ is that relation between two integers x and y which consists in the fact that x = my. This relation, like the relation +m, is by no means capable of being identiﬁed with the inductive cardinal number m, because a relation and a class of classes are objects | of utterly diﬀerent kinds. It will be seen that /n is always the same relation, whatever inductive number n may be; it is, in short, the relation of to any other inductive cardinal. We may call this the zero of rational numbers; it is not, of course, identical with the cardinal number . Conversely, the relation m/ is always the same, whatever inductive number m may be. There is not any inductive cardinal to correspond to m/. We may call it “the inﬁnity of rationals.” It is an instance of the sort of inﬁnite that is traditional in mathematics, and that is represented by “∞.” This is a totally diﬀerent sort from the true Cantorian inﬁnite, which we shall consider in our next chapter. The inﬁnity of rationals does not demand, for its deﬁnition or use, any inﬁnite classes or inﬁnite integers. It is not, in actual fact, a very important notion, and we could dispense with it altogether if there were any object in doing so. The Cantorian inﬁnite, on the other hand, is of the greatest and most fundamental importance; the understanding of it opens the way to whole new realms of mathematics and philosophy. It will be observed that zero and inﬁnity, alone among ratios, are not one-one. Zero is one-many, and inﬁnity is many-one.

iii. ∗ ﬀ., especially . course in practice we shall continue to speak of a fraction as (say) greater or less than , meaning greater or less than the ratio /. So long as it is understood that the ratio / and the cardinal number are diﬀerent, it is not necessary to be always pedantic in emphasising the diﬀerence.

Of Vol.

Chap. VII. Rational, Real, and Complex Numbers

There is not any diﬃculty in deﬁning greater and less among ratios (or fractions). Given two ratios m/n and p/q, we shall say that m/n is less than p/q if mq is less than pn. There is no diﬃculty in proving that the relation “less than,” so deﬁned, is serial, so that the ratios form a series in order of magnitude. In this series, zero is the smallest term and inﬁnity is the largest. If we omit zero and inﬁnity from our series, there is no longer any smallest or largest ratio; it is obvious that if m/n is any ratio other than zero and inﬁnity, m/n is smaller and m/n is larger, though neither is zero or inﬁnity, so that m/n is neither the smallest | nor the largest ratio, and therefore (when zero and inﬁnity are omitted) there is no smallest or largest, since m/n was chosen arbitrarily. In like manner we can prove that however nearly equal two fractions may be, there are always other fractions between them. For, let m/n and p/q be two fractions, of which p/q is the greater. Then it is easy to see (or to prove) that (m + p)/(n + q) will be greater than m/n and less than p/q. Thus the series of ratios is one in which no two terms are consecutive, but there are always other terms between any two. Since there are other terms between these others, and so on ad inﬁnitum, it is obvious that there are an inﬁnite number of ratios between any two, however nearly equal these two may be. A series having the property that there are always other terms between any two, so that no two are consecutive, is called “compact.” Thus the ratios in order of magnitude form a “compact” series. Such series have many important properties, and it is important to observe that ratios aﬀord an instance of a compact series generated purely logically, without any appeal to space or time or any other empirical datum. Positive and negative ratios can be deﬁned in a way analogous to that in which we deﬁned positive and negative integers. Having ﬁrst deﬁned the sum of two ratios m/n and p/q as (mq + pn)/nq, we deﬁne +p/q as the relation of m/n + p/q to m/n, where m/n is any ratio; and −p/q is of course the converse of +p/q. This is not the only possible way of deﬁning positive and negative ratios, but it is a way which, for our purpose, has the merit of being an obvious adaptation of the way we adopted in the case of integers. We come now to a more interesting extension of the idea of number, i.e. the extension to what are called “real” numbers, which are the kind that embrace irrationals. In Chapter I. we had occasion to mention “incommensurables” and their | discovery by Pythagoras. It was through them, i.e. through geometry, that irrational numbers were ﬁrst thought of. A square of which the side is one inch long will have a diagonal of which the length is the square root of inches. But, as the ancients discovered, there is no fraction of which the square

speaking, this statement, as well as those following to the end of the paragraph, involves what is called the “axiom of inﬁnity,” which will be discussed in a later chapter.

Strictly

We can equally well form a descending series of fractions whose squares are all greater than . If possible. and Complex Numbers is . taken to so-and-so many places. That is to say. But the problem did not remain in this geometrical form. and therefore n/p will also be the square root of . Thus m is an even number. because the square of an odd number is odd. where p is half of m. let m/n be the square root of . they must be. nature seems able to baﬄe him by exhibiting lengths which no numbers can estimate in terms of the unit. though here it took on a wider form. m and n cannot both be even. if we divide a number by . if m /n = . suppose I assign some small amount in advance. sooner or later. through an unending series of numbers that are each half of its predecessor. but we should have reached sooner or later a term in the series. we must reach an odd number after a ﬁnite number of steps. p/q will also be the square root of . have squares that diﬀer from by less than this amount. since it also involved complex numbers. so that m /n = . Thus we shall have p = n . and diﬀering. after which all terms would have had squares diﬀering from by less than this still smaller amount. say the tenth. This seems like a challenge thrown out by nature to arithmetic. say one-billionth. but diﬀering from in their later members by less than any assigned amount. If we set to work to extract the square root of by the usual arithmetical rule. Or we may put the argument even more simply by assuming that the m/n we start with is in its lowest terms. it might have been necessary to go further along the series. in that case. for if m = p. exactly fulﬁls the above conditions. Now if m is even. Rational. This proposition is proved in the tenth book of Euclid. Hence p = n . by less than any assigned . the same problem arose as regards the solution of equations.e. VII. and therefore m must be an even number. i. yet we have seen that. We can form an ascending series of fractions all of which have their squares less than . m must divide by . But then we can repeat the argument: if n = q. and so on. It is clear that fractions can be found which approach nearer | and nearer to having their square equal to . But this is impossible. Thus there cannot be any fraction m/n whose square is . and so on. it will be found that all the terms of our series after a certain one.Chap. then m = p . we shall obtain an unending decimal which. but greater by continually smaller amounts as we come to later terms of the series. Real. Thus no fraction will express exactly the length of the diagonal of a square whose side is one inch long. As soon as algebra was invented. say the twentieth. The proof is extraordinarily simple. which is one of those books that schoolboys supposed to be fortunately lost in the days when Euclid was still used as a text-book. However the arithmetician may boast (as Pythagoras did) about the power of numbers. And if I had assigned a still smaller amount. m = n . and then halve the half.

we say that the maximum of the lower section is the lower limit of the upper section. Thus our | cordon. there is nothing. since it only arises in series where there are consecutive terms. we ﬁnd that. we say that the minimum of the Stetigkeit und irrationale Zahlen. Rational. and in our upper section all ratios whose square is greater than . In this way we seem to be drawing a cordon round the square root of . but a minimum to the other. the ﬁrst is illustrated by any series in which there are consecutive terms: in the series of integers. () there may be a maximum to the one and no minimum to the other. a lower section must end with some number n and the upper section must then begin with n + . Of these four cases. If we divide all ratios into two classes. and Complex Numbers amount. and it may seem diﬃcult to believe that it can permanently escape us. Real. The above method of dividing all the terms of a series into two classes. according as their squares are less than or not. has been drawn in the as wrong place. and has not caught . We can. . In the third of our four cases. is illustrated if we put in our lower section all ratios whose square is less than . There is no lower limit short of zero to the diﬀerence between the numbers whose square is a little less than and the numbers whose square is a little greater than . . there are four possibilities: () there may be a maximum to the lower section and a minimum to the upper section. for instance. where ought to be. though we have drawn it as tight√ possible. The second case will be illustrated in the series of ratios if we take as our lower section all ratios up to and including . of which the one wholly precedes the other. there is no maximum to the one class. as we have seen. divide all ratios into two classes such that all the terms in one class are less than all in the other. and no minimum to those whose square is greater than . We may neglect the ﬁrst of our four cases. or of any set of terms chosen out of the upper section in such a way that no term of the upper section is before all of them. and for our upper section all ratios from upward (including itself). it is not by this method that we shall actually reach the square root of . Between these √ two classes. Brunswick. and there is no minimum to the other. all have their squares greater than . There is no maximum to the ratios whose square is less than . Nevertheless.Chap. among those whose squares are not less than . In the second of our four cases. The fourth case. was brought into prominence by Dedekind.” With respect to what happens at the point of section. () there may be neither a maximum to the one nor a minimum to the other. and in our upper section all ratios greater than . VII. and is therefore called a “Dedekind cut. The third case is illustrated if we take for our lower section all ratios less than . nd edition. () there may be no maximum to the one. in short.

The notion of “limit” is of the utmost importance. we say that | there is a “gap”: neither the upper section nor the lower has a limit or a last term. A term x is said to be an “upper limit” of a class α with respect to a relation P if () α has no maximum in P. in the ﬁrst. What delayed the true theory of irrationals was a mistaken belief that there must be “limits” of series of ratios. Real. and before proceeding further it will be well to deﬁne it.” to which we may give the name “upper boundary. if not. () every member of the ﬁeld of P which precedes x precedes some member of α. These deﬁnitions do not demand that the terms to which they are applied should be quantitative.” since sections of the series of ratios have “gaps” when they correspond to irrationals. If there is neither | a maximum nor a limit. In the fourth case. their “maximum” (if any) will be the last of the moments. In the second and third cases. The “lower boundary” is the lower limit or minimum. The “minimum” of a class with respect to P is its maximum with respect to the converse of P. but. or of any set of terms chosen out of the lower section in such a way that no term of the lower section is after all of them. if there is such a term. but they have few important applications except to cases when the relation is serial or quasi-serial. Rational. Reverting to the four kinds of Dedekind section. there is no upper boundary. we see that in the case of the ﬁrst three kinds each section has a boundary (upper or lower as the case may be). their “maximum” (if any) will be the ﬁrst of the moments. . The notions of limit and maximum do not essentially demand that the relation in respect to which they are deﬁned should be serial. In this case. they are consecutive terms of the series. the two boundaries are identical. (By “precedes” we mean “has the relation P to.”) This presupposes the following deﬁnition of a “maximum”:— A term x is said to be a “maximum” of a class α with respect to a relation P if x is a member of α and of the ﬁeld of P and does not have the relation P to any other member of α. the upper section has a lower boundary. VII. () every member of α which belongs to the ﬁeld of P precedes x.Chap. we may also say that we have an “irrational section. For example. It is also clear that. A notion which is often important is the notion “upper limit or maximum. and Complex Numbers upper section is the upper limit of the lower section. and the “lower limit” with respect to P is the upper limit with respect to the converse of P.” Thus the “upper boundary” of a set of terms chosen out of a series is their last member if they have one. while in the fourth kind neither has a boundary. whenever the lower section has an upper boundary. given a series of moments of time arranged by earlier and later. but if they are arranged by later and earlier. it is the ﬁrst term after all of them.

in the above-mentioned work. in this case we will call the lower section a “segment. We have to deﬁne a new kind of numbers called “real numbers. let us observe that an irrational is represented by an irrational cut. From the habit of being inﬂuenced by spatial imagination.” But there are an inﬁnite number of series for which it is not veriﬁed. hence they can all be arranged in a series by the relation of whole and part. In order to decide what they are to be. and then the segments without boundaries will be extra. people have supposed that series must have limits in cases where it seems odd if they do not.e. must not be identiﬁed with ratios. VII. We have seen that the series of ratios in order of magnitude is not Dedekindian. In order to make use of this. are such that. they are the same as the advantages of theft over honest toil. Real. Let us conﬁne ourselves to cuts in which the lower section has no maximum. will give rise to more segments than it has terms. The method of “postulating” what we want has many advantages. i. upper or lower as the case may be.” Then those segments that correspond to ratios are those that consist of all ratios less than the ratio they correspond to. Thus. we must ﬁnd some way of eliciting from it a precise deﬁnition. in which there are segments that have no boundary. we must disabuse our minds of the notion that an irrational must be the limit of a set of ratios. which to begin with is no more than a vague feeling. or can have irrationals as their limits. It is clear that an irrational Dedekind cut in some way “represents” an irrational. and a cut is represented by its lower section. and in order to do this. in the same kind of way in which the ratio n/ corresponds to the integer n. one must be part of the other. which was to ﬁll the Dedekind gap. Rational. so those rational | numbers which can be greater or less than irrationals. A series in which there are Dedekind gaps. Segments. of any two pertaining to one series. perceiving that there was no rational limit to the ratios whose square is less than . both those that have boundaries and those that do not.” of which some will be rational and some irrational. We are now in a position to deﬁne a real number and an irrational number.Chap. which is their boundary. Let us leave them to others and proceed with our honest toil. Dedekind. but they are not the same as ratios. that every section must have a boundary. It is for this reason that series where his axiom is veriﬁed are called “Dedekindian. they allowed themselves to “postulate” an irrational limit. . set up the axiom that the gap must always be ﬁlled. i. while those that represent irrationals are those that have no boundary. Just as ratios whose denominator is are not identical with integers.e. since each term will deﬁne a segment having that term for boundary. and Complex Numbers A series is called “Dedekindian” when every section has a boundary. Those that are rational “correspond” to ratios.

An “irrational number” is a segment of the series of ratios which has no boundary. is the class of proper fractions. A “rational real number” is a segment of the series of ratios which has a boundary. We may state the deﬁnition more shortly as follows:— The arithmetical sum of two real numbers is the class of the arithmetical sums of a member of the one and a member of the other chosen in all possible ways. vol. given any set of segments. the truth is that it is the limit of the corresponding set of rational real numbers in the series of segments √ ordered by whole and part. For a fuller treatment of real numbers. Rational. Real. and Complex Numbers A “real number” is a segment of the series of ratios in order of magnitude. see ibid.Chap. We deﬁne it as the sum of µ and ν. i. More simply still.) For a fuller treatment of the subject of segments and Dedekindian relations. see Principia Mathematica. is the upper limit of all those segments of the series of ratios that correspond to ratios whose √ square is less than . This gives a new class of ratios. | We can deﬁne the arithmetical product of two real numbers in exactly the same way.. VII. and xxxiv. ∗ ﬀ. their boundary will be their logical sum.e. Form the class of all such sums obtainable by varying the selected members of µ and ν. each being a class of ratios. vol. The great advantage of this method is that it requires no new assumptions. chaps. and it is easy to prove that this new class is a segment of the series of ratios. iii. For. (In all such deﬁnitions. It is easy to prove that the series of segments of any series is Dedekindian. Given two real numbers µ and ν. the series of ratios is to be deﬁned as excluding and inﬁnity. The real number . ii. and it is the rational real number corresponding to that ratio. . There is no diﬃculty in deﬁning addition and multiplication for real numbers as above deﬁned. | In the cases in which we naturally supposed that an irrational must be the limit of a set of ratios.. but enables us to proceed deductively from the original apparatus of logic. ∗ –. Thus a rational real number consists of all ratios less than a certain ratio. the class of all those terms that belong to at least one segment of the set. for instance. is the segment consisting of all those ratios whose square is less than . The class of ratios thus generated is deﬁned as the product of the two real numbers. The above deﬁnition of real numbers is an example of “construction” as against “postulation. For example. xxxiii. and Principles of Mathematics. by multiplying a member of the one by a member of the other in all possible ways. take any member of µ and any member of ν and add them together according to the rule for the addition of ratios.” of which we had another example in the deﬁnition of cardinal numbers.

and that two complex numbers are only identical when the ﬁrst real number involved in the one is equal to the ﬁrst involved in the other. for example. x being the “real” part. A complex number may be regarded and deﬁned as simply an ordered couple of real numbers. Real. to be able to say that every quadratic equation has two roots. In the case of complex numbers. where x and y are real. fractional. that two real numbers are required to determine a complex number. if they are deﬁned as ordered couples of real numbers. The part yi is called the “imaginary” part of this number. such an equation as x + = has no roots. Every generalisation of number has ﬁrst presented itself as needed for some simple problem: negative numbers were needed in order that subtraction might be always possible. and the second to the second. They are required less for geometry than for algebra and analysis. Rational. many deﬁnitions are possible. It has been simply assumed that they would obey the usual arithmetical rules. and so on. and complex numbers are needed in order that extraction of roots and solution of equations may be always possible. or real. fractions were needed | in order that division might be always possible. and on this assumption their employment has been found proﬁtable. Here. and it is to the deﬁnition of complex numbers that we must now turn our attention. since otherwise a−b would be meaningless if a were less than b. We are to have . and Complex Numbers There is no diﬃculty in extending our deﬁnitions to positive and negative real numbers and their addition and multiplication. We desire. are not demanded by geometry in the same imperative way in which irrationals are demanded.”) Complex numbers have been for a long time habitually used by mathematicians. as elsewhere. A “complex” number means a number involving the square root of a negative number. (The reason for the phrase “real numbers” is that they are contrasted with such as are “imaginary. though capable of a geometrical interpretation. and that among these we can distinguish a ﬁrst and a second. But if we are conﬁned to real numbers.Chap. VII. any number involving the square root of a negative number can be expressed in the form x + yi. It remains to give the deﬁnition of complex numbers. Since the square of a negative number is positive. we secure at once some of the properties required. But extensions of number are not created by the mere need for them: they are created by the deﬁnition. whether integral. All that is necessary is that the deﬁnitions adopted shall lead to certain properties. a number whose square is to be negative has to be a new sort of number. Complex numbers. namely. in spite of the absence of any precise deﬁnition. What is needed further can be secured by deﬁning the rules of addition and multiplication. Using the letter i for the square root of −. and every cubic equation has three. and such an equation as x − = has only one.

The deﬁnition of complex numbers of order n is obtained by an obvious extension of the deﬁnition we have given. xy + x y). by the above rule. ). and Complex Numbers (x + yi) + (x + y i) = (x + x ) + (y + y )i (x + yi)(x + y i) = (xx − yy ) + (xy + x y)i.Chap. . And when y is . in Dr Whitehead’s Universal Algebra. By these deﬁnitions we shall secure that our ordered couples shall have the properties we desire. have certain uses that are not without importance in geometry. It is easy to give a geometrical interpretation of complex numbers in the geometry of the plane. Thus the square of the couple (. p. (x. which it is natural to write simply x. x . for example. For example. Just as it is natural (but erroneous) | to identify ratios whose denominator is unity with integers.” provided we remember that the “x” is not really a real number. . §. We deﬁne a complex number of order n as a one-many relation whose domain consists of certain real numbers and whose converse domain consists of the integers from to n. Principles of Mathematics. y ).e. “x + i” may be replaced simply by “x” and “ + yi” by “yi. Now those couples in which the second term is are those which. be the couple (−yy . in the notation x + yi. Cliﬀord in his Common Sense of the Exact Sciences. but a special case of a complex number. y ). as may be seen. and the correlation is one-many. and their product is to be the couple (xx − yy . Although this is an error in theory. “yi” may of course be replaced by “i. Cf. Rational. they are x + i. This will. xn ). Thus our deﬁnitions serve all necessary purposes. though much less useful and important than those what we have been deﬁning. given two ordered couples of real numbers. a book of great merit. VII. y + y ). This subject was agreeably expounded by W.” Thus the couple (. because xr and xs may be equal when r and s are not equal. ) equal to (−. their sum is to be the couple (x + x . according to the usual nomenclature. This is what we desired to secure. ). Real. This is what would ordinarily be indicated by the notation (x . ). Thus we shall deﬁne that. not necessarily one-one. x . The above deﬁnition. it is a convenience in practice. the square of i is −. ) will be the couple (−. ) is represented by −. . Complex numbers of a higher order. with a suitable rule of multiplication. but written before the importance of purely logical deﬁnitions had been realised. where the suﬃxes denote correlation with the integers used as suﬃxes. K. so it is natural (but erroneous) to identify complex numbers whose imaginary part is zero with real numbers. i. Now our rules of multiplication make the square of (. y) and (x . ) is represented by i. . y) and (. have their imaginary part zero. take the product of the two couples (. . and the couple (−.

Chap. Real. VII. The application of number to inﬁnite collections must be our next topic. Rational. and Complex Numbers will serve all purposes for which complex numbers of higher orders are needed. We have now completed our review of those extensions of number which do not involve inﬁnity. .

It cannot be said to be certain that there are in fact any inﬁnite collections in the world.e. Various subtleties arise in identifying this form of our assumption with | the form that asserts the existence of inﬁnite collections. and we are therefore justiﬁed. The assumption that there are is what we call the “axiom of inﬁnity. for our present purposes. In the present chapter we shall attempt to explain the theory of transﬁnite or inﬁnite cardinal numbers as it results from a combination of his discoveries with those of Frege on the logical theory of numbers. for. To these we gave the name “inductive numbers. is the assumption that. if n is an inductive number. if n = n + . there is reason to fear that they are all fallacious. This is a perfectly well-deﬁned class. we come to consider the axiom of inﬁnity on its own account. This is involved in Peano’s assumption that no two inductive numbers have the same successor. The practical form of this hypothesis. which has been solved in our own day. nor have we inquired whether such collections can be said to have a number at all. but we will leave these out of account until. For the present we shall merely assume that. chieﬂy by Georg Cantor.CHAPTER VIII INFINITE CARDINAL NUMBERS The deﬁnition of cardinal numbers which we gave in Chapter II. and that there is no conclusive logical reason for believing it to be true. At the same time. i. in logic. was applied in Chapter III.” because we found that they are to be deﬁned as numbers which obey mathematical induction starting from . In the ﬁrst place. Thus we are assuming nothing that was not involved in Peano’s primitive propositions. But we have not yet considered collections which do not have an inductive number of terms. n is not equal to n + . n is not equal to n + . This is an ancient problem. Let us now consider the collection of the inductive numbers themselves.” Although various ways suggest themselves by which we might hope to prove this axiom. then n − and n have the same successor. a . in a later chapter. in investigating the hypothesis that there are such collections. to ﬁnite numbers. to the ordinary natural numbers. namely n. there is certainly no logical reason against inﬁnite collections. if n is any inductive number.

e. i. To say that a class has a number which is not altered by the addition of is the same thing as to say that. Such diﬀerences might be multiplied ad lib. therefore the total number of inductive numbers is greater than n. this set of classes is the number of the inductive numbers according to our deﬁnitions. the number of terms in the class of inductive numbers is to be deﬁned as “all those classes that are similar to the class of inductive numbers”—i. of the inductive properties were wrongly judged to be such as must belong to all numbers. and that if n has it so has n + . those which possess every property possessed by and by the successors of possessors. Nevertheless. The ﬁrst step in understanding inﬁnite numbers consists in realising the mistakenness of this view.e. Thus the class of “inductive numbers” is perfectly deﬁnite. if we take a term x which does not belong to the class. i. The most noteworthy and astonishing diﬀerence between an inductive number and this new number is that this new number is unchanged by adding or subtracting or doubling or halving or any of a number of other operations which we think of as necessarily making a number larger or smaller. diﬀerent from all of them. at least. no matter which of the inductive numbers n may be. the number of numbers from to n (both included) is n + . every series whose ﬁeld has n terms has a last term. and we must dwell on it for a time. Thus the number of inductive numbers is a new number. the class is similar to the sum . not possessing all inductive properties. we can ﬁnd a one-one relation whose domain is the class and whose converse domain is obtained by adding x to the class. Inﬁnite Cardinal Numbers cardinal number is a set of classes which are all similar to each other and are not similar to anything except each other. the property of being unchanged by the addition of is a very important one. For in that case. The diﬃculties that so long delayed the theory of inﬁnite numbers were largely due to the fact that some. it is better to deﬁne an inﬁnite cardinal number as one which does not possess all inductive properties. Now it is easy to see that this number is not one of the inductive numbers. but if n is an inductive number. and yet that this new number does not have it. We then deﬁne as the “inductive numbers” those among cardinals which belong to the posterity of with respect to the relation of n to n + . meaning by the “successor” of n the number n + . If n is any inductive number.Chap. VIII. but for various reasons. as it is easy to prove. By our general deﬁnition of cardinal numbers. some of which will appear as we proceed. simply as one which is not an inductive number. The fact of being not altered by the addition of is used by Cantor for the deﬁnition of what he calls “transﬁnite” cardinal numbers. It may happen that has a certain | property. If we arrange the inductive numbers in a series in order of magnitude. this series has no last term. indeed it was thought that they could not be denied without contradiction.e.

e. we shall also have n = n − . has a perfect one-one correspondence with its original. has the whole of the inductive numbers for its domain. which is part. is oneone. Thus the whole class of inductive numbers is similar to what the same class becomes when we omit .) A “reﬂexive” cardinal number is the cardinal number of a reﬂexive class. and all except for its converse domain. The relation of n to n + . Royce is interested in the fact that the map. When this can be done. such classes will be called “reﬂexive.” Thus: A “reﬂexive” class is one which is similar to a proper part of itself. the relation of n to n. conﬁned to inductive numbers. must contain a map of the map. if it is correct. One of the most striking instances of a “reﬂexion” is Royce’s illustration of the map: he imagines it decided to make a map of England upon a part of the surface of England. conﬁned to inductive numbers. i. Again. and must contain the same number of points as the whole. is in one-one relation with the whole. This property was used by Leibniz (and many others) as a proof that inﬁnite numbers are impossible. and the even inductive numbers alone for its converse domain. and for this purpose we cannot do better than consider the number-series itself. This point is interesting. Consequently it is a “reﬂexive” class according to the deﬁnition. | the correlator by which it is done may be said to “reﬂect” the whole class into a part of itself.” there is . Hence the total number of inductive numbers is the same as the number of even inductive numbers. thus our map. so that it has the same number as a class with one extra term. n = n + .Chap. has the whole of the inductive numbers for its domain. and so on ad inﬁnitum. In this case. but need not occupy us at this moment. i. and the number of its terms is a “reﬂexive” number. so that if n is this number. is one-one.” But this is one of those phrases that depend for their plausibility upon an unperceived vagueness: the word “equal” has many meanings. It can be shown that the cases in which this happens are the same as the apparently more general cases in which some part (short of the whole) can be put into one-one relation with the whole. to a class having one extra term. which must therefore be a reﬂexive number. if it is accurate. which must in turn contain a map of the map of the map. for this reason. We have now to consider this property of reﬂexiveness. A map. it was thought self-contradictory that | “the part should be equal to the whole. In fact. (A “proper part” is a part short of the whole. VIII. but if it is taken to mean what we have called “similar. Inﬁnite Cardinal Numbers of itself and the term x. there will be one-one relations whose domains consist of the whole class and whose converse domains consist of just one term short of the whole class. we shall do well to pass from picturesque illustrations to such as are more completely deﬁnite.e.

The kind of series which is called a “progression” has already been considered in Chapter I. VIII. since an inﬁnite collection can perfectly well have parts similar to itself. Inﬁnite Cardinal Numbers no contradiction. by the same relation (that of n to n) reﬂect the even numbers into the multiples of . The term belonging to the domain but not to the converse domain will be what he calls “”. and every member of the series is to be in the posterity of this term with respect to the relation “immediate predecessor. and so on.” These characteristics may be summed up in the following deﬁnition:— A “progression” is a one-one relation such that there is just one term belonging to the domain but not to the converse domain. and so on. For this purpose we will ﬁrst deﬁne the kind of series exempliﬁed by the inductive cardinals in order of magnitude. The even numbers are a “map” of all the inductive numbers. in which the air of paradox is much less. the map of the map of the map of all from onward.” Cf. attributed to numbers in general properties which can only be proved by mathematical induction. the term to which a term has the one-one relation will be the “successor” of the term. we have the following translations:— () “ is a number” becomes: “The member of the domain which is not a member of the converse domain is a member of the domain. satisﬁes Peano’s ﬁve axioms. It is easy to see that a progression. It is a series which can be generated by a relation of consecutiveness: | every member of the series is to have a successor. and so on. The chief use of such illustrations is in order to become familiar with the idea of reﬂexive classes. but there is to be just one which has no predecessor. unconsciously as a rule. and the domain is identical with the posterity of this one term. Those who regard this as impossible have. the multiples of are a map of the map. so deﬁned. If we had applied the same process to the relation of n to n + . our “map” would have consisted of all the inductive numbers except . Whenever we can “reﬂect” a class into a part of itself. we can. For example. so that apparently paradoxical arithmetical propositions can be readily translated into the language of reﬂexions and classes.” Taking his ﬁve axioms in turn.Chap. these into the multiples of . the same relation will necessarily reﬂect that part into a smaller part. the map of the map would have consisted of all from onward. the multiples of are a map of the map of the map. mistakenly. Principia Mathematica. ∗ . we can reﬂect. This is an abstract analogue to Royce’s problem of the map. and which only their familiarity makes us regard. and the domain of the one-one relation will be what he calls “number. as true beyond the region of the ﬁnite. . vol. all the inductive numbers into the even numbers. ii. It will be useful to give a deﬁnition of the number which is that of the inductive cardinals. and so on ad inﬁnitum. as we have just seen.

because by the deﬁnition the posterity of the ﬁrst term is the same as the domain. derive a relation which is serial from the one-one relation by which we deﬁne a progression: the method used is that explained in Chapter IV. with cardinal numbers. We will call this member “the ﬁrst term. It is easy to show that two progressions are “similar” in the sense deﬁned for similarity of relations in Chapter VI. () This is mathematical induction. | () “ is not the successor of any number” becomes: “The ﬁrst term is not a member of the converse domain. This cardinal number is the smallest of the inﬁnite cardinal numbers. which are the same as their domains) are similar classes.” () “The successor of any number is a number” becomes: “The term to which a given member of the domain has the relation in question is again a member of the domain. it follows that their domains (or their ﬁelds. Thus progressions as we have deﬁned them have the ﬁve formal properties from which Peano deduces arithmetic. The domains of progressions form a cardinal number. for the moment. Inﬁnite Cardinal Numbers This is equivalent to the existence of such a member. Two transitive asymmetrical relations which generate progressions are similar. every member of the domain is a member of the posterity of the ﬁrst term. it is the one to which Cantor has appropriated the Hebrew Aleph with the suﬃx . Since two progressions are similar relations.. since every class which is similar to the domain of a progression is easily shown to be itself the domain of a progression. which it is by deﬁnition (being one-one). to distinguish it from larger inﬁnite cardinals. and the relation is that of a term to a member of its proper posterity with respect to the original one-one relation. for the same reasons for which the corresponding one-one relations are similar. Thus the name of the smallest of inﬁnite cardinals is ℵ . by the general deﬁnition of posterity). But we are concerned. the number to which Cantor has given the name ω. which have other suﬃxes. () “No two numbers have the same successor. of course. and therefore a member of the domain. by which he has made it famous. which is given in our deﬁnition.” This is only to say that the relation is one-many.” which is again an immediate result of the deﬁnition. We can. and becomes: “Every member of the domain belongs to the posterity of the ﬁrst term.Chap. The class of all such transitive generators of progressions is a “serial number” in the sense of Chapter VI.” This is proved as follows: By the deﬁnition. VIII. hence the successor of a member of the domain must be a member of the posterity of the ﬁrst term (because the posterity of a term always contains its own successors. .. it is in fact the smallest of inﬁnite serial numbers.” which was part of our deﬁnition.

and therefore do not diminish the number of its terms. This is easily seen by arranging ratios in a series on the following plan: If the sum of numerator and denominator in one is less than in the other. . ratios. Take (say) inductive numbers of the form nn . however sparsely it may be distributed. . . ﬁrst. . it is. is greater than ℵ . namely. it contains n subclasses—in other words. Hence we can arrange all ratios in a progression. put ﬁrst the one with the smaller numerator. in fact. and seem to be only an inﬁnitesimal proportion of ratios. which remains ℵ . put the one before the other. and is not hard to extend to inﬁnite numbers. . and it is not hard to prove that n is greater than n even when n is inﬁnite. and secondly. that if a class has n members. and all ratios occur in it sooner or later. ℵ . Inﬁnite Cardinal Numbers To say that a class has ℵ terms is the same thing as to say that it is a member of ℵ . . . . or all except every tenth term or every hundredth term. The easiest way of proving this is to prove. the ﬁrst is familiar in the case of ﬁnite numbers. . that the number of sub-classes contained in a class is always greater than the number of members of the class. ℵ . These methods of thinning out a progression do not make it cease to be a progression. it is clear that the number of sub-classes of a given class (say α) is at least as great as the number of members. and their number is therefore ℵ . The proof of the second is so simple and so instructive that we shall give it: In the ﬁrst place. or n nn . since ratios whose denominator is correspond to the integers. Of these two propositions. One might be inclined to think that there must be many more ratios than integers. or every other term. Hence it follows that. and yet there are just as many of them as there are inductive numbers altogether. . It is obvious that any progression remains a progression if we omit a ﬁnite number of terms from it. VIII. Take. Conversely. The number of real numbers. . for example. It is not the case. that there are n ways | of selecting some of its members (including the extreme cases where we select all or none). and this is the same thing as to say | that the members of the class can be arranged in a progression. if the sum is equal in the two. that all inﬁnite collections have ℵ terms. however. But in actual fact the number of ratios (or fractions) is exactly the same as the number of inductive numbers. namely. In fact. for example. since each member constitutes a sub-class.Chap. ℵ . we can add terms to the inductive numbers without increasing their number. . Such numbers grow very rare in the higher parts of the number series. . This gives us the series . and we thus have a correlation of all the members with some of the sub-classes. This series is a progression. any selection from a progression is a progression if it has no last term.

it must be greater. Combining this with the proposition that. i. Let us form the whole class. and it is not correlated with any member of α. and therefore again is not correlated with β. It follows from this proposition that there is no maximum to the inﬁnite cardinal numbers. if n is the number of members. by what was said earlier. given any one-one relation whose domain is the members and whose converse domain is contained among the set of sub-classes. n will be still greater. Taking next the terms which are not members of β. each of them is (by the deﬁnition of β) correlated with some sub-class of which it is not a member. for example. Inﬁnite Cardinal Numbers if the number of sub-classes is not equal to the number of members. Since R was any one-one correlation of all members | with some sub-classes. . However great an inﬁnite number n may be. (). where n is any inductive number. ℵ + n = ℵ . there must be at least one sub-class not belonging to the converse domain. This . even when n is inﬁnite. This is a sub-class of α. VIII. Now it is easy to prove that the number is not equal. and is therefore not correlated with β. it may happen that x is correlated with a sub-class of which it is not a member. The arithmetic of inﬁnite numbers is somewhat surprising until one becomes accustomed to it. we have the theorem that n is always greater than n. again. it is greater. or. taking ﬁrst the members of β. p. For. It does not matter to the proof if β has no members: all that happens in that case is that the sub-class which is shown to be omitted is the null-class. it follows that there is no correlation of all members with all sub-classes. each of them (by the deﬁnition of β) is correlated with some sub-class of which it is a member. ℵ + = ℵ . The proof is as follows: When a one-one correlation R is established between all the members of α and some of the sub-classes. by showing that. Hence in any case the number of sub-classes is not equal to the number of members. proof is taken from Cantor. β say. of those members x which are correlated with sub-classes of which they are not members. n is the number of sub-classes. with some simpliﬁcations: see Jahresbericht der Deutschen Mathematiker-Vereinigung. Thus no member of α is correlated with β.Chap. We have. ℵ = ℵ . it may happen that a given member x is correlated with a sub-class of which it is a member. and therefore.

very similar results follow from the fact that ℵ is unchanged when multiplied by or or any ﬁnite number n or by ℵ . But it is otherwise when we subtract ℵ from itself. the number of terms in a series which has “continuity” in the sense in which this word is used by Cantor. we may then get any result. Addition. where n is any inductive number. since a ratio is determined by a pair of inductive numbers. the numbers from to n−. () All the odd numbers—remainder.) ℵ n = ℵ . i. numbering n terms in all. and cannot therefore be employed as they are employed in elementary arithmetic. Assuming space and time to be continuous in this sense (as we commonly do in analytical geometry and kinematics). Although addition and multiplication are always possible with inﬁnite cardinals. It follows that ℵ divided by ℵ may have any value from up to ℵ . . if the other number is reﬂexive.Chap. but the inverse operations—subtraction. All these are diﬀerent ways of subtracting ℵ from ℵ . it is ℵ . and exponentiation proceed quite satisfactorily. (This follows from ℵ = ℵ by induction. then ℵ n+ = ℵ = ℵ . whether | line. This is easily seen by examples. multiplication. as we shall see later. if n is ﬁnite. and extraction of roots—are ambiguous. () All the inductive numbers from n onwards—remainder. or volume. numbering ℵ terms. VIII. it is easy to see that the number of ratios is the square of the number of inductive numbers. all the even numbers. take away the following collections of ℵ terms:— () All the inductive numbers—remainder. area. this will be the number of points in space or of instants in time. subtraction gives a perfectly deﬁnite result. and the notions that depend upon them fail when inﬁnite numbers are concerned. from up to ℵ . Take subtraction to begin with: so long as the number subtracted is ﬁnite. Thus ℵ − n = ℵ . and all give diﬀerent results. zero. for if ℵ n = ℵ . After ℵ . it remains unchanged.e. ℵ is a very important number. but we saw that it is also ℵ . From the inductive numbers.) But ℵ > ℵ . so far. ℵ is the most important and interesting of inﬁnite cardinal numbers. for. all goes well. namely. division. it will also be the number of points in any ﬁnite portion of space. subtraction and division no longer give deﬁnite results. Inﬁnite Cardinal Numbers (This follows from the case of the ratios. In fact. From the ambiguity of subtraction and division it results that negative numbers and ratios cannot be extended to inﬁnite numbers. As regards division.

” and would be lacking in some of the inductive properties. we should not have n = n + . All known inﬁnite classes and cardinals are reﬂexive. but for the present it is well to preserve an open mind as to whether there are instances. but it is not known at present whether all inﬁnite classes and cardinals are reﬂexive. and a class as ﬁnite when its number is ﬁnite. we deﬁned a number as ﬁnite when it obeys mathematical induction starting from . . . we adopt the following deﬁnitions:— A ﬁnite class or cardinal is one which is inductive. of classes and cardinals which are neither reﬂexive nor inductive. If n were such a cardinal. but the proofs hitherto oﬀered by Cantor and others (including the present author in former days) are fallacious. that is to say. We shall return to this subject in Chapter XII. VIII.e. . that the ﬁnite | numbers are those that occur in the ordinary number-series . it is not known whether there are classes and cardinals which are neither reﬂexive nor inductive. Inﬁnite Cardinal Numbers The characteristic by which we deﬁned ﬁnitude was mathematical induction. An inﬁnite class or cardinal is one which is not inductive. and may very possibly be capable of proof. the inﬁnite numbers we have discussed have not merely been noninductive: they have also been reﬂexive. he believes that every class and every cardinal is either inductive or reﬂexive. and believes that it is equivalent to non-inductiveness. .Chap. but n would not be one of the “natural numbers. But in the present chapter. This deﬁnition yields the sort of result that a deﬁnition ought to yield.” At present. hitherto unknown. All reﬂexive classes and cardinals are inﬁnite. . namely. . i. Cantor used reﬂexiveness as the deﬁnition of the inﬁnite. Meanwhile. . This may be true. for reasons which will be explained when we come to consider the “multiplicative axiom.

. In order to make matters clear. . . We have already had occasion to consider one kind of inﬁnite series. . . . . n + . . progressions. . . and so on. We start with the series . . . . on the other hand. . . . . the sort that Cantor calls ω. If we imagine this process carried on as long as possible. . . . In this chapter we shall consider the subject more generally. . . Let us ﬁrst consider various diﬀerent kinds of series which can be made out of the inductive numbers arranged on various plans. as we have already seen. . to add terms without altering the serial number: everything depends upon the way in which they are added.. n + . . . it is possible to change the serial number of a series without adding or taking away any terms. n. . . . . . we ﬁnally reach the series . .. . . . . . . . . . Let us proceed to thin out this series by repeatedly performing the | operation of removing to the end the ﬁrst even number that occurs. . . . . It is possible to keep the cardinal number of a reﬂexive class unchanged in spite of adding terms to it. . We thus obtain in succession the various series: . . . . by mere re-arrangement. which. . The most noteworthy characteristic of an inﬁnite series is that its serial number can be altered by merely re-arranging its terms. n. . . . . .CHAPTER IX INFINITE SERIES AND ORDINALS An “inﬁnite series” may be deﬁned as a series of which the ﬁeld is an inﬁnite class. At the same time. . . . n. . . in the case of any inﬁnite series it is also possible. n + . . as with cardinals. . . In this respect there is a certain oppositeness between cardinal and serial numbers. namely. represents the smallest of inﬁnite serial numbers. . it will be best to begin with examples. .

. in accordance with the general deﬁnition of order. . . Multiplication. . . This number is greater than ω or ω + n. . It is to be observed that. y .” The one which puts ﬁrst all the odd numbers and then all the even ones will be deﬁned by: “x and y are ﬁnite integers. It is therefore necessary . ω + . according to the deﬁnition—i. like addition. n + . and either x is odd and y is even or x is less than y and both are odd or both are even. but no series having the second number contains a part having the ﬁrst number. The number which we have called ω. to give these formulæ in future. . . but the fact that they could be given is essential. ω + is greater than ω. This is characteristic of relation-arithmetic generally: if µ and ν are two relation-numbers.g. Inﬁnite Series and Ordinals in which we have ﬁrst all the odd numbers and then all the even numbers. Each of these numbers is “greater” than any of its predecessors. each of these arrangements of integers is to be regarded as resulting from some deﬁnite relation. we see that the ﬁrst is similar to the part of the second which omits the last term. ω + . . in the following sense:— One serial number is said to be “greater” than another if any series having the ﬁrst number contains a part having the second number. we still have a progression. but the second is not similar to any part of the ﬁrst. . the number . But if we add a term at the beginning of a progression instead of the end. . n. namely. as a rule. the one which merely removes to the end will be deﬁned by the following relation: “x and y are ﬁnite integers. . and its serial | number is ω. . the general rule is that µ + ν is not equal to ν + µ. . x . which is itself a progression. E. .Chap. If we compare the two series . but a couple of progressions gives a series which is twice as long as a progression. is quite exceptional. . Thus + ω = ω. The series we ﬁnally reached just now consisted of ﬁrst all the odd numbers and then all the even numbers. y . xn . is sometimes called ω . . Thus + ω is not equal to ω + . the number of a series consisting of two progressions. . in which there is equality. . . . . y . depends upon the order of the factors: a progression of couples gives a series such as x . . and either y is and x is not . yn . IX. . The serial numbers of these various series are ω + . where n is ﬁnite.) Thus the second series has a greater serial number than the ﬁrst. x . The case of ﬁnite ordinals. but is easily demonstrated.” We shall not trouble. . . . ω. . or neither is and x is less than y. namely. . (This is obvious. .e. .

or in which there are subordinate parts having no beginning. is called ω . all that can be obtained by thinning out a progression. . The ordinal number of the series of all ordinals that can be made out of an ℵ .. then their doubles. of which we shall treat later. having no beginning. (This is not diﬃcult to prove. .. This question is connected with the multiplicative axiom. . . .” A well-ordered series is one which has a beginning. . . For example. ω . we can place ﬁrst the odd numbers. . We can proceed from ω and ℵ to ω and ℵ by a process exactly analogous to that by which we advanced from ω and ℵ to ω and ℵ . but . we can always go further. however far we have gone. . for a progression of couples. We can proceed to ω . . . . Inﬁnite Series and Ordinals to distinguish between ω and ω . ωω . . Usage is variable. and so on. and has a term next after any selection of its terms. . . The series of all the ordinals that can be obtained in this way. IX. on the one hand..) The cardinal number of the class of such ordinals can be shown to be greater than ℵ . β” will have to stand for a suitably constructed sum of α relations each having β terms. . . . . then the doubles of these. and this decision of course governs our general interpretation of “α.. . . of which the number is ω . This excludes. and so on. it is the number which Cantor calls ℵ . . . . and has consecutive terms. in which there are terms between any two. provided there are any terms after the selection. for aught we know. We can proceed indeﬁnitely with the process of thinning out the inductive numbers. It is not even known whether it is comparable with them in magnitude. . i. Thus a series whose ordinal number is ω has a ﬁeld whose cardinal number is ℵ . The series of negative integers in order of magnitude. . it may be neither equal to nor greater nor less than any one of the Alephs. . but ending with −.Chap. . Any one of the progressions in this new series can of course be | thinned out as we thinned out our original progression. . taken in order of magnitude. compact series. is not well-ordered. we shall use ω for a couple of progressions and ω . It is not known whether ℵ is equal to any of the cardinals in the series of Alephs. is itself longer than any series that can be obtained by re-arranging the terms of a progression. We thus obtain the series . All the series we have been considering so far in this chapter have been what is called “well-ordered. .e. and on the other hand series which have no beginning. since it is a progression of progressions. .β” when α and β are relation-numbers: “α . And there is nothing to prevent us from advancing indeﬁnitely in this way to new cardinals and new ordinals.

Inﬁnite Series and Ordinals taken in the reverse order. and so on. Of the usual formal laws of addition. but when we remove the dot. the null-class) has a ﬁrst term. IX. and exponentiation. multiplication. but diﬀer at the second digit. instead of starting with ·. It follows from this example that it is possible to construct compact series having ℵ terms. A property may be said to be “transﬁnitely hereditary” if. This makes it possible to prove many propositions concerning well-ordered series which are not true of all series. and ratios in order of magnitude form a compact series. Omitting these. An “ordinal” number means the relation-number of a well-ordered series. there is no diﬃculty. but only some are obeyed by transﬁnite ordinals. but ﬁrst of all the one with no second digit. but not as regards the (n + )th . It is easy to arrange the inductive numbers in series which are not well-ordered. and we have a compact series consisting of all ﬁnite integers except such as divide by . Generally. a generalised form of mathematical induction applies. we will transfer to the right any ’s that occur at the beginning of our decimal. We shall resume this topic in the next chapter. Of two that do begin with the same digit. | as the reader can easily convince himself. and returning to the ones that have no ’s at the beginning. This rule of arrangement. and those that are obeyed by them are . when it belongs to a certain selection of the terms in a series. beginning with −. between any two there are always an inﬁnite number of others. as we saw. Now omit the dot at the beginning of each. the one with the smaller second digit comes ﬁrst. it belongs to their immediate successor provided they have one. we can state the rule for the arrangement of our integers as follows: Of two integers that do not begin with the same digit. there is no diﬃculty about including those that are divisible by . a transﬁnitely hereditary property belonging to the ﬁrst term of the series belongs to the whole series. These form a compact series. that one comes ﬁrst which has either no (n + )th digit or a smaller one than the other. thus we have here another example. the one that begins with the smaller digit comes ﬁrst. all are obeyed by transﬁnite cardinals. The deﬁnition is: | A “well-ordered” series is one in which every sub-class (except.Chap. and even to arrange them in compact series. being in fact a progression. If we wish to include those that divide by . if two integers agree as regards the ﬁrst n digits. arranged in order of magnitude. we can adopt the following plan: consider the decimals from · (inclusive) to (exclusive). In fact. gives rise to a compact series containing all the integers not divisible by . and. it is well-ordered. It is thus a species of serial number. Among well-ordered series. of course. we will include all decimals less than . we have already seen that there are ℵ ratios. For example. In a well-ordered series.

the distributive law (adopting the convention | we have adopted above as regards the order of the factors in a product) holds in the form (β + γ)α = βα + γα. whether ﬁnite or inﬁnite. ∗ –. The distributive law: α(β + γ) = αβ + αγ. When the commutative law does not hold. the exponential laws α β . The commutative law does not hold. The associative law: (α + β) + γ = α + (β + γ) and (α × β) × γ = α × (β × γ). III. II. which is obviously connected with the commutative law for multiplication.Chap. some hold and some do not. The laws of exponentiation: α β . and for ﬁnite ordinals. one form may be true and the other false. The commutative law: α + β = β + α and α × β = β × α. The reader who wishes to know what they are and how the above laws are proved must consult the second volume of Principia Mathematica. the associative law does hold. β γ = (αβ)γ . α γ = α β+γ . The deﬁnitions of multiplication and exponentiation that are assumed in the above propositions are somewhat complicated. but not in the form α(β + γ) = αβ + αγ. But when we come to inﬁnite ordinals. (α β )γ = α βγ . . or indeed to relation-numbers in general. but not the law α γ . IV. α γ . the above form of the distributive law must be distinguished from (β + γ)α = βα + γα. β γ = (αβ)γ . Inﬁnite Series and Ordinals obeyed by all relation-numbers. As we shall see immediately. All these laws hold for cardinals. IX. By the “usual formal laws” we mean the following:— I. α γ = α β+γ and (α β )γ = α βγ still hold.

though by no means small. Inﬁnite Series and Ordinals Ordinal transﬁnite arithmetic was developed by Cantor at an earlier stage than cardinal transﬁnite arithmetic. Cardinals are essentially simpler than ordinals. IX. and it is a curious historical accident that they ﬁrst appeared as an abstraction from the latter. have more diﬃculty in understanding and using notions which are comparatively “simple” in the logical sense than in manipulating more complex notions which are | more akin to their ordinary prac.Chap. and is very largely merged in that of the more general conception of relation-numbers. And mathematicians. This does not apply to Frege’s work. For these reasons. probably in the main on account of the diﬃculty of his symbolism. But from the point of view of the philosophy of mathematics it is less important and less fundamental than the theory of transﬁnite cardinals. in which cardinals. were treated in complete independence of ordinals. is distinctly less than that of cardinals. and only gradually came to be studied on their own account. tice. like other people. . it was only gradually that the true importance of cardinals in mathematical philosophy was recognised. ﬁnite and transﬁnite. but it was Cantor’s work that made the world aware of the subject. The importance of ordinals. because it has various technical mathematical uses which led him to it. while Frege’s remained almost unknown.

. . namely. . A given point on a line may be the limit of a set of points on the line. so that among those others there would be some diﬀering by less than any assigned quantity. without its being necessary to bring in co-ordinates or measurement or anything quantitative. not a quantitative fact. but we will here repeat the deﬁnitions which lead to it. ematics has been found continually greater than had been thought. n. what really occurs is a set of ﬁnite quantities having zero for their lower limit. has been already deﬁned. in a general form in which they do not demand that the relation concerned shall be serial. The cardinal number ℵ is the limit (in the order of magnitude) of the cardinal numbers . but Weierstrass showed that this is an error: wherever inﬁnitesimals were thought to occur. . it was supposed that inﬁnitesimals were involved in the foundations of these subjects. .” of increasing complexity.CHAPTER X LIMITS AND CONTINUITY The conception of a “limit” is one of which the importance in math. It used to be thought that “limit” was an essentially quantitative notion. which is an ordinal fact. the notion of a quantity to which others approached nearer and nearer. . from which the rest are derived. in the series. | There are various forms of the notion of “limit. depends upon limits. . . Formerly. ﬁnite numbers get no nearer to ℵ as they grow larger. . What makes ℵ the limit of the ﬁnite numbers is the fact that. not involving quantity at all (except by accident when the series concerned happens to be quantitative). The simplest and most fundamental form. The “maxima” with respect to P are the minima with respect to the converse of P. But in fact the notion of “limit” is a purely ordinal notion. it comes immediately after them. although the numerical diﬀerence between ℵ and a ﬁnite cardinal is constant and inﬁnite: from a quantitative point of view. The deﬁnitions are as follows:— The “minima” of a class α with respect to a relation P are those members of α and the ﬁeld of P (if any) to which no member of α has the relation P. . The whole of the diﬀerential and integral calculus. indeed practically everything in higher mathematics.

one sequent.e. . i. of all fractions of the form − /n for diﬀerent ﬁnite values of n. Or. The “lower limits” with respect to P are the upper limits with respect to the converse of P. . We can.Chap. . If a class has no boundary. we can speak of “the limit” (if any).” . Limits and Continuity The “sequents” of a class α with respect to a relation P are the minima of the “successors” of α. Take. consider the prime numbers. This will be a segment in the sense deﬁned | in Chapter VII. but if α has a maximum. for example.. it has no upper limits. An “upper limit” of α is a boundary which is not a member of α. and the whole of α is therefore included in the segment deﬁned by α. and it is clear that the segment which it deﬁnes (in the whole series of fractions in order of magnitude) is the class of all proper fractions. the segment deﬁned by α consists of all the terms that precede some term or other of α. Whenever P has connexity. the class consisting of the fractions . If P is serial. we can greatly simplify the above deﬁnition of a limit. . deﬁne ﬁrst the “boundary” of a class α. in that case. If α has a maximum. considered as a selection from the cardinals (ﬁnite and inﬁnite) in order of magnitude. This series of fractions has no maximum. and the “successors” of α are those members of the ﬁeld of P to which every member of the common part of α and the ﬁeld of P has the relation P. or of what is called a “gap. . in the cases we are concerned with in practice. again.. Assuming that P is serial. Thus.” We will speak of the “segment of P deﬁned by a class α” as all those terms that have the relation P to some one or more of the members of α. every segment in the sense there deﬁned is the segment deﬁned by some class α. For this purpose it is best to use the notion of “segment. it has neither maximum nor limit. the “boundary” of a class α will be the term x (if it exists) whose predecessors are the segment deﬁned by α. its limit or maximum. This is the case of an “irrational” Dedekind cut. every member of α precedes some other member of α.e. etc. In this case the segment deﬁned consists of all ﬁnite integers. and then proceed to distinguish the case where the boundary is the limit from the case where it is a maximum. The “upper limits” of α with respect to P are the sequents provided α has no maximum. When P is a serial relation. a class can have at most one maximum. the segment will be all the predecessors of the maximum. one minimum. The “precedents” with respect to P are the sequents with respect to the converse of P. X. But if α has no maximum. A “maximum” of α is a boundary which is a member of α. i. indeed. .

. The upper limiting-points of the ﬁeld of this new series will be . ω . . The word “continuity” had been used for a long time. every member of this series is both an upper and a lower limiting-point for suitably chosen sets. ω. . . . . We saw in Chapter VII. ω. If we consider. ω + ω. The ﬁrst deﬁnition that would naturally occur to a man seeking a precise meaning for the continuity of series would be to deﬁne it as consisting in what we have called “compactness. . ω + ω . Limits and Continuity Thus the “upper limit” of a set of terms α with respect to a series P is that term x (if it exists) which comes after all the α’s. but Cantor’s deﬁnition is narrower than Dedekind’s: a series which has Cantorian continuity must have Dedekindian continuity. . But if we consider such a series as the series of ratios. . ω . . . ω. ω . Such a state of aﬀairs seems contrary to the vague feeling we have as to what . . because there are no terms except the last that have no immediate successors. . . We shall. ω + . ω . . . . for example. . this set (the rationals) will have all the real numbers as upper and lower limiting-points. . .Chap. ω. With regard to limits. the series of ordinals—and indeed every wellordered series—has no lower limiting-points. have to distinguish upper limiting-points from lower limiting-points. We may deﬁne all the “upper limiting-points” of a set of terms β as all those that are the upper limits of sets of terms chosen out of β. . ω . | the upper limiting-points of the ﬁeld of this series are those that have no immediate predecessors. . . .e. . .” and the limiting-points of the ﬁrst derivative are called the second derivative.” i. . of course. the series of ordinal numbers: . of which one wholly precedes the other. But this would be an inadequate deﬁnition. but the converse does not hold. . we may distinguish various grades of what may be called “continuity” in a series. The limiting-points of a set are called its “ﬁrst derivative. . and of which the ﬁrst has no last term. in the fact that between any two terms of the series there are others. ω . and so on. and select out of it the rational real numbers. . that there are innumerable ways in which the series of ratios can be divided into two parts. . . If we consider the series of real numbers. ω . . . because of the existence of “gaps” in series such as the series of ratios. i. ω . X. . . . but had remained without any precise deﬁnition until the time of Dedekind and Cantor. ω. On the other hand. ω. ω + . | while the second has no ﬁrst term. Each of these two men gave a precise signiﬁcance to the term. .e. but is such that every earlier term comes before some of the α’s.

it shows that the series of ratios is not the sort of series that is needed for many mathematical purposes. a series is Dedekindian when there are no gaps. If one of these is assumed. ﬁrst travel through some of Cantor’s conceptions in this subject before giving his deﬁnition of continuity. for example: we wish to be able to say that when two straight lines cross each other they have a point in common. Take geometry. The absence of gaps may arise either through terms having successors. the other can be deduced. We are thus led to a closer investigation of series with respect to limits. Limits and Continuity should characterise “continuity. therefore. and this is a further property which our deﬁnition does not enable us to deduce. the two lines might cross in a “gap” and have no point in common. this is a property belonging to compact series. that led to the deﬁnition of “Dedekindian” continuity. This is a crude example. (It is suﬃcient to assume that there is always an upper boundary. It will be remembered that we deﬁned a series as Dedekindian when every sub-class of the ﬁeld has a boundary. be ﬁttingly called continuity. It was the needs of geometry. that we desire to be able to assign such properties to geometrical space as shall make it certain that every point can be speciﬁed by means of co-ordinates which are real numbers: this is not insured by Dedekindian continuity alone. and so is the series of real numbers. this deﬁnition somewhat conceals the considerations which have given rise to it. but if the series of points on a line were similar to the series of ratios. There is no correction required so far as concerns the property that all its points are to be limiting-points. We want to be sure that every point which cannot be speciﬁed by rational co-ordinates can be speciﬁed as the limit of a progression of points | whose co-ordinates are rational. But this deﬁnition is still too wide for many purposes. Thus we are led to the deﬁnition: A series has “Dedekindian continuity” when it is Dedekindian and compact. in that case our series must have a property which may. in its simplest form. for example. and to .” and. But this deﬁnition does not express quite accurately what he means. We shall.) That is to say. as much as anything. although. for many purposes. what is more.Chap. X. or that there is always a lower boundary. Cantor deﬁnes a series as “perfect” when all its points are limitingpoints and all its limiting-points belong to it. The former sort of Dedekindian series is excluded by assuming that our series is compact. or through the existence of limits in the absence of maxima. Thus a ﬁnite series or a well-ordered series is Dedekindian. Suppose. This investigation was made by Cantor and formed the basis of his deﬁnition of continuity. but many others might be given to show that compactness is inadequate as a mathematical deﬁnition of continuity.

or by regressions (which are the converses of progressions). | We come now to the second property by which perfection was to be deﬁned.) Thus. i. and of which the ﬁrst has no immediate predecessor. Cantor calls our series “condensed in itself” (insichdicht). Cantor assumes that they are to be deﬁned by progressions.e. the series in which every point is a limiting-point are compact series. without specifying which. . Limits and Continuity no others if all points are to be upper limiting. as we saw. This. X. the series of decimals in which a decimal ending in a recurring is distinguished from the corresponding terminating decimal and placed immediately before it. and every progression or regression contained in the series has a limit in the series. but has exceptional terms which are consecutive. Apart from such series. and limiting-points are taken in relation to the larger series. was ﬁrst deﬁned as consisting in the fact that all the limiting-points of a series belong to it. i.. (It is not clear how far he recognises that this is a limitation. Otherwise. it cannot fail to contain its limiting-points. and this holds without qualiﬁcation if it is speciﬁed that every point is to be an upper limiting-point (or that every point is to be a lower limiting-point). every subordinate series which has no maximum has a limit. e. we ﬁnd that the deﬁnition we want is the following:— A series is said to be “closed” (abgeschlossen) when every progression or regression contained in the series has a limit in the series. i. Such a series is very nearly compact. But this only has any eﬀective signiﬁcance if our series is given as contained in some other larger series (as is the case. But if it is only assumed that they are limiting-points one way. when every term is the limit of a progression or regression.e. What Cantor means is not exactly what he says.Chap. every subordinate series has a boundary. We then have the further definition:— A series is “perfect” when it is condensed in itself and closed. on other occasions he says something rather diﬀerent. we must distinguish diﬀerent kinds of limiting-points according to the nature of the smallest sub-series by which they can be deﬁned. there will be other series that will have the property in question—for example. but only for progressions and regressions. the property which Cantor calls that of being “closed” (abgeschlossen). if a series is considered simply on its own account. with a selection of real numbers). which is what he means. while the second has no immediate successor. indeed.e. What he really means is that every subordinate series which is of the sort that might be expected to have a limit does have a limit within the given series. namely.g. When every member of our series is the limit of a progression or regression. But Cantor does not state this for every subordinate series. ﬁnally. Although Cantor does not explicitly consider the matter.or all lower limitingpoints.

as well as for other reasons. added to perfection. or (to speak more cautiously) of a kind of continuity which can. but derivative from the above and more complicated. The notions of limit and continuity which we have been deﬁning must not be confounded with the notions of the limit of a function for approaches to a given argument. () it contains a median class having ℵ terms. This property. the continuity of space and time (if they are continuous) is an instance of the continuity of series. These are diﬀerent notions. The deﬁnitions of continuity which we have been considering. For this purpose we have to add a further property. those of Dedekind and Cantor. however | little the two may diﬀer. We may slightly simplify his deﬁnition. Limits and Continuity In seeking a deﬁnition of continuity. All series having Cantorian continuity are similar. Among the real numbers some are rational. X. This gives a further property which suﬃces to characterise continuity completely. The continuity of motion (if motion is continuous) is an instance of the continuity of a function. we say: A “median class” of a series is a sub-class of the ﬁeld such that members of it are to be found between any two terms of the series. be reduced to the continuity of series. it will be well to deal brieﬂy with the notions of limits and continuity as applied to functions. however near together. although the number of irrationals is greater than the number of rationals. do not correspond very closely . on the other hand. This class Cantor deﬁnes as that of continuous series. To avoid confusion. by suﬃcient mathematical | manipulation. namely. The number of rationals.” It will be seen that it implies Dedekindian continuity. is ℵ . In view of the fundamental importance of motion in applied mathematics. suﬃces to deﬁne a class of series which are all similar and are in fact a serial number. We then ﬁnd that Cantor’s deﬁnition is equivalent to the following:— A series is “continuous” when () it is Dedekindian. but to no others. or the continuity of a function in the neighbourhood of a given argument. To begin with. but this subject will be best reserved for a separate chapter. namely. as we saw. yet there are rationals between any two real numbers. what Cantor has in mind is the search for a deﬁnition which shall apply to the series of real numbers and to any series similar to that. we shall speak of this kind as “Cantorian continuity. but not all series having Dedekindian continuity. Thus the rationals are a median class in the series of real numbers.Chap. some are irrational. the property of containing a class of ℵ members in such a way that some of this class occur between any two terms of our series. but the converse is not the case. It is obvious that there cannot be median classes except in compact series. very important.

the sort of general obliteration of distinctions which characterises a thick fog. quite deﬁnitely and uncompromisingly. It would be quite impossible | to justify this thesis within the limits of the present volume. though it can be made less than any given ﬁnite amount assigned in advance. in order to return to topics more closely connected with mathematics.” is one which is certainly quite diﬀerent from that which we have been deﬁning. A fog gives an impression of vastness without deﬁnite multiplicity or division. e. Take. and its distance from every other unit is ﬁnite. by what we see at a given time. but it may.Chap. separate unit. as well as in parts of Our Knowledge of the External World. be very well maintained that the mathematical conception which we have been considering in this chapter gives the abstract logical scheme to which it must be possible to bring empirical material by suitable manipulation. the series of real numbers. The reader who is interested may read an attempt to justify it as regards time in particular by the present author in the Monist for −. I think. if that material is to be called “continuous” in any precisely deﬁnable sense. interesting as it is. It is not to be maintained that the two kinds are simply identical. for example. They conceive continuity rather as absence of separateness. to be characteristic of his mental life and of that of children and animals. It is this sort of thing that a metaphysician means by “continuity. X. With these indications. . or by the word “ﬂux. Limits and Continuity to the vague idea which is associated with the word in the mind of the man in the street or the philosopher. is a diﬃcult and intricate one. Each is what it is. The question of the relation between the kind of continuity existing among the real numbers and the kind exhibited. we must leave this problem.g. it does not pass over by imperceptible degrees into another.” declaring it. The general idea vaguely indicated by the word “continuity” when so employed. very truly. it is a hard.

∗ –. Let us consider any ordinary mathematical function f x. there is only one value that f x can have. but errors incorporated. wrong views upon our present topics have become so ﬁrmly embedded in the minds of professional philosophers that a prolonged and considerable eﬀort is required for their uprooting. It has been thought ever since the time of Leibniz that the diﬀerential and integral calculus required inﬁnitesimal quantities. and philosophers have tended to ignore the work of such men as Weierstrass. as Dr Whitehead has shown. e. Mathematicians (especially Weierstrass) proved that this is an error.” When a function is what we call “continuous. we can make the diﬀerences in f x fall below any assigned amount. Limits and continuity of functions. We call x the “argument. vol. and also with the deﬁnition of what is meant by a “continuous function. begin with the deﬁnitions in the text-books. in works on ordinary mathematics. and if we make the diﬀerences in x small enough. when x is given. where | x and f x are both real numbers. if a function is to be continuous. die hard.e.CHAPTER XI LIMITS AND CONTINUITY OF FUNCTIONS In this chapter we shall be concerned with the deﬁnition of the limit of a function (if any) as the argument approaches a given value. and f x is one-valued—i. especially through the so-called inﬁnitesimal calculus. . ii.g.” Both of these ideas are somewhat technical. We do not want. that See Principia Mathematica. This is not essential. and would hardly demand treatment in a mere introduction to mathematical philosophy but for the fact that.” and f x the “value for the argument x. and not only to such as are numerical or numerically measurable. however. are deﬁned in terms involving number. and proceed afterwards to show how these deﬁnitions can be generalised so as to apply to series in general. We will.” the rough idea for which we are seeking a precise deﬁnition is that small diﬀerences in x shall correspond to small diﬀerences in f x. in what Hegel has to say about mathematics.

so that it passes more and more quickly through the cycle of values from one multiple of π/ to another as x becomes smaller and smaller.” where x is a real number. x . any change. for some value of x. or generally from (n − )π/ to (n + )π/. . as a non-mathematical example.Chap. they are the exceptions: there are inﬁnitely more discontinuous functions than continuous ones. will be very small. Consequently sin /x passes more and more quickly from − | to and back again. where is some number which. . It is clear that continuity at a given point has to do with what happens in any neighbourhood of that point. and then the value changes suddenly from one birthplace to the other. The actual fact is that. or in ℵ places. will make a change in f x which exceeds some assigned ﬁnite amount. and we cannot diminish the oscillations by making the interval smaller. all values for arguments throughout this neighbourhood . Limits and Continuity of Functions there shall be sudden jumps. Many functions are discontinuous for one or several values of the variable. or from π/ to π/. log x. Take as an example sin /x. and so on. and then gives a sudden jump. What we desire is this: If a is the argument for which we wish our function to be continuous. Now if we consider /x when x is very small. The function sin θ passes through all values from − to every time that θ passes from −π/ to π/. XI. in important cases. “the place of birth of the youngest person living at time t. let us ﬁrst deﬁne a neighbourhood (α say) containing the value f a which the function has for the argument a. In fact.” This is a function of t. This function remains constant from one integer to the next. sin x. to x . It is easy to manufacture functions which are discontinuous in several places. when argument and value are both real numbers. say the interval from − to + where is some very small number. or everywhere. Take. Examples will be found in any book on the theory of functions of a real variable. if we take a suﬃciently small neighbourhood containing a. however small. Proceeding now to seek a precise deﬁnition of what is meant by saying that a function is continuous for a given argument. . its value is constant from the time of one person’s birth to the time of the next birth. if we take any interval containing . but continuous for all other values. so that. though continuous functions are more familiar. for example. But it is not at all diﬃcult to deﬁne discontinuous functions. sin /x will go through an inﬁnite number of oscillations in this interval. let us ﬁrst deﬁne a “neighbourhood” of a number x as all the numbers from x − to x + . where n is any integer. we desire that. we see that as x diminishes /x grows faster and faster. Thus round about the argument the function is discontinuous. An analogous mathematical example would be “the integer next below x. as x grows smaller. however small. The ordinary simple functions of mathematics have this property: it belongs.

but as small as we please. Given any number in this section. The general rule is that a function oscillates. and that. But it is only the exceptionally “tame” function that has a deﬁnite limit as the argument approaches a given point. however small. If this can be done. the value of the function lies within the neighbourhood from f (a) − σ to f (a) + σ . namely. there are values at least as great as this number for arguments between a − and a. we could have deﬁned the continuity of a function diﬀerently: a function is continuous at a point where its value is the same as the limit of its values for approaches either from above or from below. given any neighbourhood of a given argument. That is to say. no matter how small we may have made α. in important cases. for every positive number σ . σ ﬁrst deﬁnes a neighbourhood of f (a). Let us consider what may happen as the argument approaches some value a from below. In this deﬁnition. will be very small. such that. So far we have not deﬁned the “limit” of a function for a given argument. for all arguments within this neighbourhood. that from a − to a + . where is some number which. the function is “continuous” for the argument a. having a in the middle of it. That is to say. As this is the general rule. a whole stretch of values will occur for arguments within this neighbourhood. diﬀerent from . the section consisting of those numbers that are not greater than all the values for arguments from a − to a. And this is to remain true whatever tiny amount we may select. we can always ﬁnd a stretch of real numbers. however σ may be chosen. XI. Limits and Continuity of Functions shall be contained in the neighbourhood α. there exists a positive number . The common part of all these sections we will call the “ultimate section” A number is said to be “numerically less” than when it lies between − and + . the diﬀerence f (a + δ) − f (a) is numerically less than σ . for all values of δ which are numerically | less than . i. diﬀerent from . such that. such that throughout this stretch f x will not diﬀer from f a by more than the prescribed tiny amount. namely. The values of the function for arguments from a − to a (a excluded) will be a set of real numbers which will deﬁne a certain section of the set of real numbers. namely. the neighbourhood from f (a) − σ to f (a) + σ .Chap.e. let us consider it ﬁrst. . for arguments that fall very little short | of a (if is very small). if we decree that our function is not to diﬀer from f a by more than some very tiny amount. If we had done so. we wish to consider what happens for arguments contained in the interval from a− to a. Let us take all possible ’s and all possible corresponding sections. Hence we are led to the following deﬁnition:— The function f (x) is said to be “continuous” for the argument a if. The deﬁnition then proceeds to say that we can (by means of ) deﬁne a neighbourhood.

If a term z belongs both to the ultimate section and to the ultimate upper section. Let us begin with the “ultimate section. this deﬁnes an upper section which will vary as varies. if the ultimate oscillation is null. i. to sections that go from some point up to the top. there are arguments between a − and a for which the value of the function is not less than z. we shall still ﬁnd values as great as x and values as small as x. together with the positive numbers up to and including . we obtain the “ultimate upper section. or one term. however small we may make . But if the ultimate oscillation has many terms. whatever may be. in order to ﬁt in with the above deﬁnitions.e. We shall assume.” We may illustrate the matter by considering once more the function sin /x as x approaches the value . positive and negative. the boundary of the ultimate section is the same as that of the ultimate upper section. that this value is approached from below.” Between − and . In the ﬁrst two cases the function has a deﬁnite limit for approaches from below. The ultimate oscillation may contain no terms. it consists of all negative numbers together with . Similarly the “ultimate upper section” consists of all positive numbers together with . both included. however near we come to a. the function will assume the value for certain arguments. for it is not diﬃcult to prove that. but will never assume any greater value. up to and including . It is equally true if it has none. Limits and Continuity of Functions as the argument approaches a. we shall say that it belongs to the “ultimate oscillation. Here we take those numbers that are not less than all the values for arguments from a − to a.” To say that a number z belongs to the ultimate upper section is to say that.e. To say that a number z belongs to the ultimate section is to say that. Taking the common part of all such sections for all possible ’s. the lower boundary of the ultimate upper section and the upper boundary of the ultimate section) as the lower and upper limits of . however small we make . this is fairly obvious. Hence the ultimate section consists of all real numbers.e. i. there are arguments between a − and a for which the value of the function is not greater than z. If the ultimate oscillation has one term. and may be deﬁned as the limit of the function for approaches from below. Thus the “ultimate oscillation” consists of all real numbers from − to . XI. In this case we can take the lower and upper boundaries of the ultimate oscillation (i. or many terms. there is no deﬁnite limit to the function for approaches from below. together with the negative numbers down to and including −. | We may say generally that the “ultimate oscillation” of a function as the argument approaches a from below consists of all those numbers x which are such that. instead of from the bottom up to some point.Chap. We may apply exactly the same process to upper sections.

namely:— () Given any real number less than f a. This may be taken as deﬁning continuity: it is equivalent to our former deﬁnition. for which it has the value f a. for this argument and all arguments greater than this. just as the earlier deﬁnition of continuity proceeded. Similarly we shall say that a function “converges into α as the argument approaches x from below” if there is some argument y less than x such that throughout the interval from y (included) to x (excluded) the function has values which are members of α.e. Limits and Continuity of Functions its “ultimate” values for approaches from below. and our arguments both lie between a − and a (a excluded). even when both exist. the function is continuous for this argument. () and () Similar conditions for approaches to a from above. in the general case. () Given any real number greater than f a. if it satisﬁes four conditions. These two limits. in that case the function has a limit for approaches from below. i. This is to hold for any σ . it is necessary and suﬃcient that. Let us deﬁne the limit for approaches from below. If there is to be a deﬁnite limit for approaches to a from below. We can deﬁne the limit of a function for a given argument (if it exists) without passing through the ultimate oscillation and the four limits of the general case.Chap. then the diﬀerence between the values for these arguments will be less than σ . the function converges into the successors of this number as the argument approaches a from below. the value of the function is a member of the class α. We may now say that a function is continuous for the argument a. If it is also the value for the argument a. however small. given any small number σ . in that case. A function is called “continuous” (without qualiﬁcation) when it is continuous for every argument. and is then their common value. they still need not be identical with the value for the argument a. Thus we have. Similarly we deﬁne the case when there is a limit for approaches from above. It is only in this last case that we call the function continuous for the argument a. and if they are identical. . The limit for a given argument a only exists when all these four are equal. need not be identical. Another slightly diﬀerent method of reaching the deﬁnition of continuity is the following:— Let us say that a function “ultimately converges into a class α” if there is some real number such that. Similarly we obtain lower and upper limits of the “ultimate” values for approaches from above. XI. the function converges into the predecessors of this number as the argument approaches a from below. four limits to a function for approaches to a given argument. The deﬁnition proceeds. two values for arguments suﬃciently near to a (but both less than a) will diﬀer | by less than σ . if is suﬃciently small.

though without adequate evidence. while its converse domain is contained in the ﬁeld of Q. | We may now generalise our deﬁnitions so as to apply to series which are not numerical or known to be numerically measurable. If f (t) was the policeman’s position at time t. whereas the value for the argument t was —. from the case of motion. Let R be a one-many relation whose domain is contained in the ﬁeld of P. was being attacked by a policeman.) . and it is assumed. with x or y themselves. Suppose. The deﬁnitions given for the case of functions where argument and value are real numbers can readily be adapted for more general use. there is an interval on the Q-series containing a not as an end-point and such that. There is a story by H. we shall mean by the “P-interval x to y” all terms z such that x has the relation P to z and z has the relation P to y—together. R the relation of the position of our particle on the line at time a to the time a. Let P and Q be two relations. Limits and Continuity of Functions The advantage of this form of deﬁnition is that it analyses the conditions of continuity into four. the power of realising his wishes. given any interval α on the P-series containing the value of the function for the argument a. derived from considering arguments and values respectively greater or less than the argument and value for which continuity is to be deﬁned. It is the meaning of “continuity” involved in such statements which we now wish to deﬁne as simply as possible.Chap. so that “the R of a” is its position at time a.e. without his knowledge. Wells which will illustrate. This illustration may be borne in mind throughout our deﬁnitions. while its values belong to the ﬁeld of P. given any body. But such occurrences are supposed to be rare in the real world. for example. G. that all motions are continuous. if x and y are two members of the ﬁeld of P. if f (t) is its position at time t. P the series of points on our line from left to right. and x has the relation P to y. f (t) is a continuous function of t. whose arguments belong to the ﬁeld of Q. that.e. the function has values which are members of α. i. who possessed. the limit of the policeman’s positions as t approached to t from below would be in contact with the hero. which it is well to imagine serial. Then R is (in a generalised sense) a function. i. the diﬀerence between the limit of a function for a given argument and its value for the same argument. that we are dealing with a particle moving on a line: let Q be the time-series. (We mean by an “interval” all the terms between any two. The hero of the story. throughout this interval. but on ejaculating “Go to——” he found that the policeman disappeared. XI. though it is not necessary to our deﬁnitions that they should be so. when so stated. and t the moment of the ejaculation. We shall say that the function R is continuous for the argument | a if. The case of motion is a convenient one to bear in mind.

take any argument y which precedes a (i. then halved again. the length of our bit is /n of an inch. putting b for the value for the argument a: | Given any term having the relation P to b. and form the section of P deﬁned by these values. and many propositions about them can be proved for any two series (one being the argument-series and the other the value-series). and take their common part. and so on indeﬁnitely. the ﬁrst is.e. growing smaller without any limit short of zero. this will be the ultimate section. those members of the P-series which are earlier than or identical with some of these values. This is analogous to the fact that if a line an inch long be halved. The adaptation of the deﬁnition of convergence and the resulting alternative deﬁnition of continuity oﬀers no diﬃculty of any kind. We say that a function R is “ultimately Q-convergent into α” if there is a member y of the converse domain of R and the ﬁeld of Q such that the value of the function for the argument y and for any argument to which y has the relation Q is a member of α. The second condition is obtained by replacing P by its converse. in the notions of the limit of a function or the continuity of a function. Form all such sections for all y’s that precede a. R Q-converges into the successors of b (with respect to P) as the argument approaches a from below. but they do not involve any intervals that are not ﬁnite. The process of successive bisection does not lead to divisions whose ordinal number is inﬁnite. Thus inﬁnitesimals are not to be reached in this way. Of the four conditions that a function must fulﬁl in order to be continuous for the argument a. and this is ﬁnite whatever ﬁnite number n may be. Both can be deﬁned generally. It will be seen that the deﬁnitions do not involve inﬁnitesimals. We say that R “Q-converges into α as the argument approaches a given argument a” if there is a term y having the relation Q to a and belonging to the converse domain of R and such that the value of the function for any argument in the Q-interval from y (inclusive) to a (exclusive) belongs to α. The ultimate upper section and the ultimate oscillation are then deﬁned exactly as in the previous case.Chap. we never reach inﬁnitesimals in this way: after n bisections.e. There is thus nothing. has the relation Q to a). Confusions on such topics have had much to . XI. take the values of the function for all arguments up to and including y. the third and fourth are obtained from the ﬁrst and second by replacing Q by its converse. that essentially involves number. i. Limits and Continuity of Functions We can easily deﬁne the “ultimate section” and the “ultimate oscillation. since it is essentially a one-by-one process. They involve inﬁnite classes of intervals.” To deﬁne the “ultimate section” for approaches to the argument a from below.

Chap. . Limits and Continuity of Functions do with the diﬃculties which have been found in the discussion of inﬁnity and continuity. XI.

This sometimes demands a certain amount of ingenuity. in the case of relationnumbers) having the required number of terms. and which is convenient. in the sense that many interesting propositions. Exactly analogously we can deﬁne µ + ν. Before enunciating the multiplicative axiom. given that µ is the number of some class α and ν is the number of some class β. How shall we deﬁne µ + µ? For this purpose we must have two classes having µ terms. But in the case of multiplication. though not indispensable. cannot be proved without its help. class and whose second term is a class consisting of a single member of α. and the deﬁnition of multiplication when the number of factors may be inﬁnite. as the simplest example. because even without those propositions the subjects in which they occur still exist. as a rule. . are merely a question of a suitable technical device. we must ﬁrst explain the theory of selections. where the number of factors may be inﬁnite. important problems arise out of the deﬁnition. though in a somewhat mutilated form.CHAPTER XII SELECTIONS AND THE MULTIPLICATIVE AXIOM In this chapter we have to consider an axiom which can be enunciated. It is convenient. in terms of logic. and a class α which has µ terms. of which the following is perhaps the simplest: Form ﬁrst all the ordered couples whose ﬁrst term is a class consisting of a single member of α. the case of addition. We can construct such classes from α in various ways. the only correct procedure is to construct an actual class (or relation. and they must not overlap. form all the ordered couples whose ﬁrst term is | the null. Take. which it seems natural to suppose true. In deﬁning the arithmetical operations. and the logical sum of the two classes will have µ + µ terms. in certain portions of mathematics. and whose second term is the null-class. These two classes of couples have no member in common. Such deﬁnitions. secondly. but it is not indispensable. but not proved. Suppose we are given a cardinal number µ. then. but it is essential in order to prove the existence of the number deﬁned.

For example. ii. The following method of deﬁning multiplication generally is due to Dr Whitehead. (x . not with the case when its members are. x ). Then the class which is used to deﬁne the product µ × µ is the class of couples: (x . vol. (x . suppose the number of terms in each of these classes is given. It is the case when κ is inﬁnite. It is explained and treated at length in Principia Mathematica. It is to be observed that the problem is to be able to deal with the case when κ is inﬁnite. that we have to ﬁnd a way of dealing with. (x . it even remains adequate when α and β are identical. x ). Let us now set to work to choose one term out of each class to be its representative. How shall we deﬁne the product of all these numbers? If we can frame our deﬁnition generally. Let us suppose to begin with that κ is a class of classes no two of which overlap—say the constituencies in a country where there is no plural voting. x ). x . assuming that by law each constituency has to elect a man who is a voter in that constituency. x ). and therefore if there are µ voters in a constituency. the number of possible Parliaments is obtained by multiplying together the numbers of voters in the . It will be seen that this deﬁnition does not require that α and β should not overlap. We thus arrive at a class of representatives. (x . x ). and vol.. The choices of the diﬀerent constituencies are independent. No diﬃculty arises as regards this deﬁnition. XII. even though its members may be ﬁnite. (x . the method deﬁned above is just as applicable when its members are inﬁnite as when they are ﬁnite. The problem of multiplication when the number of factors may be inﬁnite arises in this way: Suppose we have a class κ consisting of classes. it can make µ choices. x ). Given two classes α and β. x ). (x . except that it cannot be extended to an inﬁnite number of factors. ∗ . Selections and the Multiplicative Axiom Multiplication when the number of factors is ﬁnite oﬀers no diﬃculty. This deﬁnition remains applicable when µ or ν or both are inﬁnite. when the total number of constituencies is ﬁnite. x ). each constituency being considered as a class of voters. i. one being selected out of each constituency. let α be the class whose members are x . (x . How many diﬀerent possible ways of choosing a Parliament are there? Each constituency can select any one of its voters.Chap. (x . thus it is obvious that. as constituencies do when they elect members of Parliament. ∗ ﬀ. If | κ is not inﬁnite. we can deﬁne µ × ν as the number of ordered couples that can be formed by choosing the ﬁrst term out of α and the second out of β. of which the ﬁrst has µ terms and the second ν terms. and it can be extended step by step to three or four or any ﬁnite number of factors. x . who make up our Parliament. x ). it will be applicable whether κ is ﬁnite or inﬁnite.

and x is the term which has the relation R to α. The formal deﬁnition is: A “selector” from a class of classes κ is a one-many relation.e. For purposes of deﬁning multiplication. will be the class of selections. We must now drop our illustration. is deﬁned as the product of the numbers of the members of κ. and the multiplicative class. giving rise to diﬀerent selectors in the two cases. and if α be any member of κ. The class of all “selections” from κ we shall call the “multiplicative class” of κ. µ and α have exactly one term in common. and this is to be all that R does.e. Selections and the Multiplicative Axiom various constituencies. and such that. We might.Chap. But when the members of κ overlap. This deﬁnition is equally applicable whether κ is ﬁnite or inﬁnite. there may be more selectors than selections. and let us assume to begin with that no two members of κ overlap. Before we can be wholly satisﬁed with these deﬁnitions. If R is a selector from κ. This is the method by which inﬁnite products are deﬁned. i. i. the number of possible selections from κ. since a term x which belongs to two classes α and β may be selected once to represent α and once to represent β. and proceed to exact statements. given any member α of κ. if x has the relation to α. having κ for its converse domain. We can deﬁne exponentiation by an adaptation of the above | plan. instead of deﬁning ﬁrst a class called a “selection. it picks out one term as the representative of that member. as before. When we do not know whether the number of constituencies is ﬁnite or inﬁnite. i. For this purpose. The number of terms in the multiplicative class of κ. i. deﬁne µν as the number of selectors from . but to the same selection.” we will deﬁne ﬁrst a relation which we will call a “selector. A “selection” from κ will now be deﬁned as the domain of a selector. then no member of the one is a member of the other. µ is a “selection” from κ if every member of µ belongs to some member | of κ. if. that if α and β are two diﬀerent members of κ. Thus we deﬁne: “The product of the numbers of the members of a class of classes κ” is the number of selectors from κ. of course. Let κ be a class of classes. We shall call a class a “selection” from κ when it consists of just one term from each member of κ.e.” A relation R will be called a “selector” from κ if. from every member of κ. then x is a member of α. we may take the number of possible Parliaments as deﬁning the product of the numbers of the separate constituencies. there is just one term x which is a member of α and has the relation R to α.e. it is the selectors we require rather than the selections. we must remove the restriction that no two members of κ are to overlap. XII. and α is a member of κ. we call x the “representative” of α in respect of the relation R.

but the set is one having a certain structure and a more manageable composition than is the case in general. all the couples that can be formed of a variable member of α and a ﬁxed member of β. This deﬁnition. without assuming that we can know whether it is true or false. But there is one thing we cannot prove: we cannot prove that a product is only zero when one of its factors is zero. The relevance of this to the multiplicative axiom will appear shortly. plication and exponentiation. there must be selectors from them. and α has µ members. Thus µν is deﬁned by the selectors from a certain set of ν classes each having µ terms. A selection from our class of classes will be a set of ordered couples. But there are objections to this deﬁnition. but we prefer to deﬁne it as the number of ordered couples to be formed consisting of a member of α followed by a member of β. What applies to exponentiation applies also to the product of two cardinals. These ν classes are each of them a class of couples. as we register the axiom of parallels. since our classes of couples are mutually exclusive. We might deﬁne “µ × ν” as the sum of the numbers of ν classes each having µ terms. namely. until at last we become content to register the assumption and its consequences. We can prove this when the number of factors is ﬁnite. at ﬁrst sight. In other words. they seem obviously true. there must be at least one class consisting of one term out of each of the given classes. each of which has µ terms. is designed to evade the necessity of assuming the multiplicative axiom. XII. for. or that. These things cannot be proved.Chap. of which there will be exactly one having any given member of β for its second term. the number of selectors is the same as the number of selections. but not when it is inﬁnite. We deﬁne µν as the number of selectors from the class consisting of these ν classes. derived from the fact that the multiplicative axiom (of which we shall speak shortly) is unnecessarily involved if it is adopted. If we now form all the classes of this sort that result from varying y. and although. yet reﬂection brings gradually increasing doubt. we cannot prove that. also. since any member of α may be chosen for the ﬁrst term. Let y be a member of β. and the ﬁrst term may be any member of α. given a class of classes none of which is null. and β has ν members. The assumption. . given a class of mutually exclusive classes. | With our deﬁnitions. There will be µ such couples for a given y. and β a class having ν terms. Selections and the Multiplicative Axiom ν classes. and form the class of all ordered couples that have y for their second term and a member of α for their ﬁrst term. Or we may equally well deﬁne µν as the number of selections. since y may be any member of β. we can prove the usual formal laws of multi. we obtain altogether ν classes. where α has µ terms and β has ν terms. We adopt instead the following construction:— Let α be a class having µ terms.

What Zermelo does. mathematicians had used it without a qualm. that. Such a relation picks out a “representative” from each sub-class.” This proposition we will call the “multiplicative axiom.” We will ﬁrst give various equivalent forms of the proposition. and κ all the sub-classes of α with the exception | of the null-class. i. Also vol. i. Selections and the Multiplicative Axiom loosely worded. The multiplicative axiom is equivalent to the proposition that a product is only zero when at least one of its factors is zero. The multiplicative axiom is equivalent to the assumption that if α be any class. the null-class).e. i.” i. dass jede Menge wohlgeordnet werden kann. –. but it would seem that they had done so unconsciously.e. The multiplicative axiom is equivalent to the proposition that. then there is at least one selector from κ. We may begin with the following:— “Given any class of mutually exclusive classes. Mathematische See . given any class α. in his “Beweis. then there is at least one one-many relation implying R and having κ for its converse domain. there is at least one one-many relation R whose converse domain consists of all existent sub-classes of α and which is such that. There are many equivalent ways of stating it precisely. and κ any class contained in the converse domain of R. the result cannot be unless one of the numbers concerned is . lix. there is at least one class which has exactly one term in common with each of the given classes. The full proof of this proposition is diﬃcult. can be arranged in a series in which every sub-class has a ﬁrst term (except. This is the form in which the axiom was ﬁrst brought to the notice of the learned world by Zermelo. of which none is null. In this form we shall speak of it as Zermelo’s axiom.Chap. if any number of cardinal numbers be multiplied together. vol. XII. but it is not diﬃcult to see the general principle upon which it proceeds. in the above-mentioned proof. to be equivalent to the proposition that every class can be well-ordered. is to count oﬀ the members of α. And the credit due to Zermelo for having made it explicit is entirely independent of the question whether it is true or false. It uses the form which we call “Zermelo’s axiom. The multiplicative axiom has been shown by Zermelo. is that selectors and selections exist when we should expect them. Annalen. it will often happen that two sub-classes have the same representative. pp. until he made it explicit. iii. of course. then x is a member of ξ.” Zermelo regards the axiom as an unquestionable truth. it assumes that. of course. It must be confessed that. by means Principia Mathematica. one by one. if x has the relation R to ξ. ∗ .e. ∗ –. if R be any relation. in eﬀect. and then consider certain ways in which its truth or falsehood is of interest to mathematics. vol.

nor greater than ν. one-one correlations of α and β. We naturally think that the sum of ν mutually exclusive classes. It must be diﬀerent from x . Now. there will be cardinals µ and ν such that µ is neither less than. Then take the representative of the class consisting of all of α except x . nothing is known at present. . Selections and the Multiplicative Axiom of R and transﬁnite induction. xn . and are therefore similar. the existence of certain selectors can be proved. call it x . We have seen that ℵ and ℵ possibly form an instance of such a pair. If the axiom is false. because every representative is a member of its class. and we wish to prove that the sum of one set has as many terms as the sum of the other. . The propositions that depend upon the axiom. . let xω be the representative of what is left of α. and similarly for every other pair. each having µ terms.) This proposition is called “Zermelo’s theorem. The way the multiplicative axiom enters in is as follows: Suppose we have two sets of ν mutually exclusive classes. only a hint of the general lines of the proof. . we cannot in . this can be proved. Now α and β each have µ terms. accordingly. but the above are the most important of the forms known at present. In this way we ﬁrst obtain a progression x . we must establish a one-one relation.Chap.. x . In this way we can go on until nothing is left. of course. But when ν is inﬁnite. without being known to be equivalent to it. equal to. owing to some special circumstance. Then if κ and λ are the two sets of classes. In order to prove this. and α is some member of κ. Proceed similarly to take away x .” The multiplicative axiom is also equivalent to the assumption that of any two cardinals which are not equal. The trouble is that there are so many. If κ and λ are inﬁnite. there will be a member β of λ which will be the correlate of α with respect to S. assuming that α is not ﬁnite. (The above is. We then take away the whole progression. There are. one class of the set being all the one-one correlators of α with β. there is some one-one relation between the two sets of classes. and x is shut out from this class. Take ﬁrst the connection of addition and multiplication. We put ﬁrst the representative of α. This requires a selection from a set of classes | of correlators. When ν is ﬁnite. In order to obtain a one-one correlation of the sum of κ with the sum of λ. each having µ terms. we have to pick out one correlator of α with β. and let x be the representative of what is left. since there are in each case ν classes. Let us consider some one-one relation S between the classes. Many other forms of the axiom might be given. it cannot be proved without the multiplicative axiom. one must be the greater. . but what we want is a one-one relation between their terms. must have µ × ν terms. As to the truth or falsehood of the axiom in any of its forms. call it x . XII. . are numerous and important. The successive representatives will form a | well-ordered series containing all the members of α. except where.

but this inference is fallacious. though we can prove that it is sometimes the case. It is also easy to see that. since that number is not increased by doubling. Hence we might suppose that the sum of ℵ pairs must have ℵ terms. In our case it can be done with the boots. is involved in most of the recognised theory of ordinals of the second class. But this is an instance of the diﬃculty. In view of the way in which the multiplicative axiom is involved. The problem is: How many boots had he. the proposition and its corollary cannot be regarded as proved. It is commonly inferred from this that the sum of ℵ classes each having ℵ members must itself have ℵ members. We know that × ℵ = ℵ . since we do not know that the number of terms in such a sum is ℵ × ℵ . such that a series having this ordinal number will have ℵ terms in its ﬁeld. with the corollary that ω (the smallest ordinal of the third class) is not the limit of any progression. except by some very artiﬁcial device.Chap. the predecessors of their limit form at most the sum of ℵ classes each having ℵ terms. This is illustrated by the millionaire who bought a pair of socks whenever he bought a pair of boots. we know that ℵ = ℵ × ℵ = ℵ . and therefore that the limit is a number of the “second class. All that can be said at present is that we do not know.” i. Another illustration may help to make the point clearer. Hence we cannot establish the usual kind of connection between addition and multiplication. cannot be proved to happen always | unless we assume the multiplicative axiom. To begin with. unless we can know that the multiplicative axiom is true.” That is to say. nor consequently that it is ℵ . already noted. XII. and who had such a passion for buying both that at last he had ℵ pairs of boots and ℵ pairs of socks. unless the multiplicative axiom is true—that the predecessors of the limit are ℵ in number. or they may not. This has a bearing upon the theory of transﬁnite ordinals. This fact has various curious consequences. of connecting the sum of ν classes each having µ terms with µ × ν.e. if we take any progression of ordinals of the second class. and how many socks? One would naturally suppose that he had twice as many boots and twice as many socks as he had pairs of each. The reason for the diﬀerence is this: Among boots we can distinguish right and left. it is supposed to be proved that any progression of ordinals of the second class has a limit which is again an ordinal of the second class. but not with the socks. It is easy to prove that an ordinal which has ℵ predecessors must be one of what Cantor calls the “second class. They may be true. This proposition. But this. Thus the greater part of the theory of ordinals of the second class must be regarded as unproved. Selections and the Multiplicative Axiom general know that such a selection exists. and that therefore he had ℵ of each. It is inferred thence—fallaciously. and therefore we can make a selection of one out of each pair. sometimes it cannot. Sometimes this can be done. . and never at any other time.

(The class may. in this way we obtain a progression of all the boots. To prove that a class has ℵ terms. with each pair. Of course. thus we can choose. may serve to show how a selection might be impossible. Now there is no diﬃculty in showing that a non-inductive class must contain more terms than any inductive class. but we cannot be sure of this in the case of the socks unless we assume the multiplicative axiom or fall back upon some fortuitous geometrical method of selection such as the above. of course. the centres of mass of the two socks are not both at exactly the same distance from p. like socks. that there is any class consisting of one sock out of each pair. we do not know that a selection is even theoretically possible. what comes to the . For example. we can choose all the right boots or all the left boots. XII. if µ is an inﬁnite number. or. and we cannot be sure. Within each pair. Hence the problem. a relation which is a selector.e. But there is no theoretical reason why a method of selection such as this should always be possible. which to put ﬁrst. from each pair. with a little goodwill on the part of the reader. with any | pair. It will be remembered that in Chapter VIII. one set of µ pairs may not contain the same number of terms as another set of µ pairs.) Thus we have to prove. if it were impossible to select one out of each pair of socks. take the left boot ﬁrst and the right second. there are certainly ℵ boots. itself have ℵ terms.Chap. that. There is no diﬃculty in doing this with the boots. that sock which has its centre of mass nearer to p. take the centres of mass of the socks: there will be points p in space such that. in the case of objects in space. and therefore that there were not ℵ of them. we pointed out that a reﬂexive number must be non-inductive. but that the converse (so far as is known at present) can only be proved if we assume the multiplicative axiom. The way in which this comes about is as follows:— It is easy to prove that a reﬂexive class is one which contains sub-classes having ℵ terms. it is necessary and suﬃcient to ﬁnd some way of arranging its terms in a progression. This case illustrates that. unless we assume the multiplicative axiom. it is possible to choose a progression out of its terms. It is to be observed that. We may put the matter in another way. The pairs are given as forming an ℵ . keeping the order of the pair unchanged. but with socks no such principle of selection suggests itself. we always can ﬁnd some principle of selection. i. given ℵ pairs of boots. given any noninductive class. it would follow that the socks could not be arranged in a progression. and therefore as the ﬁeld of a progression. Another important problem involving the multiplicative axiom is the relation of reﬂexiveness to non-inductiveness. for. if we can. and an inﬁnite number of arbitrary choices is an impossibility. Selections and the Multiplicative Axiom namely. But with the socks we shall have to choose arbitrarily. and the case of the socks. Unless we can ﬁnd a rule for selecting.

if it does. say x . α . . The class α . α . We may . then classes having term (as many as there are members of α). . let the one term be x . provided µ is not µ inductive. assuming the multiplicative axiom. . . . but in that case α must introduce a new term. it introduces one new term. x . We thus get a progression of sets of sub-classes. one containing three.e. at the very most. so that µ is the number of sub-classes of α and is the number of collections of sub-classes. α .e. . α . it must introduce two new terms. We now know that. . But by that time the old terms would no longer be suﬃciently numerous to form a next class with the right number of members. and so introduces no new terms. ν(ν + )/ terms. each set consisting of all those that have a certain given ﬁnite number of terms. The ﬁrst ν classes α . instead of a progression of collections of sub-classes. we select one sub-class containing one term. α . . thus it would be possible. Selections and the Multiplicative Axiom same thing. XII. of sub-classes of α. µ must be a reﬂexive number. But this is a long way from what we set out to prove. From each set of sub-classes let us choose out one. (We can do this if the multiplicative axiom is assumed. . . i. we can be sure that new members keep on coming in from time to time. but we have only proved that the number of collections of sub-classes of α is a reﬂexive number. consisting of two terms.e. . therefore new terms must come in at this point if not sooner. . α . if we omit from our progression α . say. then. . α . if µ is a non-inductive number. we shall still have a progression. which consists of one term. i. α . thus we are one step nearer to our goal. although we cannot be sure that new members of α come in at any one speciﬁed stage in the progression α . . if µ is the number of µ members of α. αν contain. to go on with only repetitions from the (ν + )th | class to the ν(ν + )/th class. α . must be reﬂexive. β . we do not know whether we can always do it or not. ν(ν + )/ + . + + + . say.) We have now a progression α . . Let us illustrate. In this case it is possible that α consists of x . that. there are sub-classes of α that have ν terms. Thus we can form sets of ﬁnite sub-classes of α: First one class having no terms. we must employ the multiplicative axiom. say. and if it does not. β . x . The class α . that if α is a non-inductive class and ν is any inductive number. i. (We shall have α = β and α = β . and so on. + ν terms. is a new beginning. omitting the sub-class consisting of the null-class alone. α . So far we have not used the multiplicative axiom. one containing two terms. α . .Chap. if there were no repetitions in the ﬁrst ν classes. Let our new progression be called β . The next step is to notice that. because α and α must introduce new terms. then classes having | terms. In order to advance beyond this point. It follows that. That is to say. otherwise. may or may not contain x . x . and so on. all those classes that are composed entirely of members that have occurred in previous classes.

for a well-ordered series must have either a ﬁnite or a reﬂexive number of terms in its ﬁeld. x is a member of γ . γ . every class can be well-ordered. γ will consist of the other member of α . Assuming the multiplicative axiom. . Let γµ be the part of βµ which consists of new members. . From this point of view. γ . The multiplicative axiom is equivalent to the assumption that all cardinal numbers are multipliable. There is one advantage in the above direct argument. Selections and the Multiplicative Axiom or may not have α = β . which would prove the multiplicative . . and is a sub-class of α. but if α does contain this one member. we shall have γ = β = α . . such a selection can be made. βµ will be αν . when | it is possible. i. Thus by twice using this axiom we can prove that. It is not improbable that there is much to be discovered in regard to the topics discussed in the present chapter. i. the β’s are some of the α’s. and so on. Cases may be found where propositions which seem to involve the multiplicative axiom can be proved without it. We can prove that a ﬁnite number is always multipliable. we only need the very much smaller assumption that ℵ is multipliable. x . It is conceivable that the multiplicative axiom in its general form may be shown to be false. For this reason it is better. Thus we get a new progression γ . or to deal with the problem of the boots and socks.) Now these β’s are such that any one of them. We may state this assumption in the form: “ℵ is a multipliable number. to content ourselves with the more restricted assumption. if the axiom is true. but only its truth as applied to a set of ℵ classes. say βµ . or to show that any progression of numbers of the second class is of the second class.e. Hence a selection from them will be a progression. . but. if x is the member of γ . if the axiom is true. The assumption made in the above direct argument is that a product of ℵ factors is never zero unless one of the factors is zero. is a progression. (Again γ will be identical with β and with α . that. every non-inductive cardinal must be reﬂexive. Zermelo’s theorem oﬀers the best hope: the continuum or some still more dense series might be proved to be incapable of having its terms well-ordered. but we cannot prove that any inﬁnite number is so. . where ν is some number greater than µ. as against deduction from Zermelo’s theorem.) This new progression of γ’s consists of mutually exclusive classes. XII. speaking generally. x is a member of γ . It may happen that the axiom holds for ℵ classes.Chap. contains members which have not occurred in any of the previous β’s.” where a number ν is deﬁned as “multipliable” when a product of ν factors is never zero unless one of the factors is zero. x . that the above argument does not demand the universal truth of the multiplicative axiom. This could also be deduced from Zermelo’s theorem. But in order to identify the reﬂexive with the non-inductive. then x . though not for larger numbers of classes. if α does not contain the one member of α .e.

Chap. XII. Selections and the Multiplicative Axiom

axiom false, in virtue of Zermelo’s theorem. But so far, no method of obtaining such results has been discovered, and the subject remains wrapped in obscurity.

CHAPTER XIII

**THE AXIOM OF INFINITY AND LOGICAL TYPES
**

The axiom of inﬁnity is an assumption which may be enunciated as follows:— “If n be any inductive cardinal number, there is at least one class of individuals having n terms.” If this is true, it follows, of course, that there are many classes of individuals having n terms, and that the total number of individuals in the world is not an inductive number. For, by the axiom, there is at least one class having n + terms, from which it follows that there are many classes of n terms and that n is not the number of individuals in the world. Since n is any inductive number, it follows that the number of individuals in the world must (if our axiom be true) exceed any inductive number. In view of what we found in the preceding chapter, about the possibility of cardinals which are neither inductive nor reﬂexive, we cannot infer from our axiom that there are at least ℵ individuals, unless we assume the multiplicative axiom. But we do know that there are at least ℵ classes of classes, since the inductive cardinals are classes of classes, and form a progression if our axiom is true. The way in which the need for this axiom arises may be explained as follows. One of Peano’s assumptions is that no two inductive cardinals have the same successor, i.e. that we shall not have m + = n + unless m = n, if m and n are inductive cardinals. In Chapter VIII. we had occasion to use what is virtually the same as the above assumption of Peano’s, namely, that, if n is an inductive cardinal, | n is not equal to n + . It might be thought that this could be proved. We can prove that, if α is an inductive class, and n is the number of members of α, then n is not equal to n + . This proposition is easily proved by induction, and might be thought to imply the other. But in fact it does not, since there might be no such class as α. What it does imply is this: If n is an inductive cardinal such that there is at least one class having n members, then n is not equal to n + . The axiom of inﬁnity assures us (whether truly or falsely) that there are classes

Chap. XIII. The Axiom of Inﬁnity and Logical Types

having n members, and thus enables us to assert that n is not equal to n + . But without this axiom we should be left with the possibility that n and n + might both be the null-class. Let us illustrate this possibility by an example: Suppose there were exactly nine individuals in the world. (As to what is meant by the word “individual,” I must ask the reader to be patient.) Then the inductive cardinals from up to would be such as we expect, but (deﬁned as + ) would be the null-class. It will be remembered that n + may be deﬁned as follows: n + is the collection of all those classes which have a term x such that, when x is taken away, there remains a class of n terms. Now applying this deﬁnition, we see that, in the case supposed, + is a class consisting of no classes, i.e. it is the null-class. The same will be true of + , or generally of + n, unless n is zero. Thus and all subsequent inductive cardinals will all be identical, since they will all be the null-class. In such a case the inductive cardinals will not form a progression, nor will it be true that no two have the same successor, for and will both be succeeded by the null-class ( being itself the null-class). It is in order to prevent such arithmetical catastrophes that we require the axiom of inﬁnity. As a matter of fact, so long as we are content with the arithmetic of ﬁnite integers, and do not introduce either inﬁnite integers or inﬁnite classes or series of ﬁnite integers or ratios, it is possible to obtain all desired results without the axiom of inﬁnity. That is to say, we can deal with the addition, | multiplication, and exponentiation of ﬁnite integers and of ratios, but we cannot deal with inﬁnite integers or with irrationals. Thus the theory of the transﬁnite and the theory of real numbers fails us. How these various results come about must now be explained. Assuming that the number of individuals in the world is n, the number of classes of individuals will be n . This is in virtue of the general proposition mentioned in Chapter VIII. that the number of classes contained in a class which has n members is n . Now n is always greater than n. Hence the number of classes in the world is greater than the number of individuals. If, now, we suppose the number of individuals to be , as we did just now, the number of classes will be , i.e. . Thus if we take our numbers as being applied to the counting of classes instead of to the counting of individuals, our arithmetic will be normal until we reach : the ﬁrst number to be null will be . And if we advance to classes of classes we shall do still better: the number of them will be , a number which is so large as to stagger imagination, since it has about digits. And if we advance to classes of classes of classes, we shall obtain a number represented by raised to a power which has about digits; the number of digits in this number will be about three times . In a

It would be natural to suppose—as I supposed myself in former days—that. . Even if there were no individuals at all. at the stage concerned. classes. we shall certainly this subject see Principia Mathematica. whatever inductive number µ may be. etc. is ﬁnite. at the next stage. for there would then be one class.. the number of terms in our whole set will be n + n + . all taken together. classes of classes of classes. the null-class. vol. classes of classes (namely. The Axiom of Inﬁnity and Logical Types time of paper shortage it is undesirable to write out this number. ∗ ﬀ. we have a very similar state of aﬀairs. and if we want larger ones we can obtain them by travelling further along the logical hierarchy. if we still do not succeed. As regards ratios. it is necessary that we should be able to make a single class or series in which no inductive cardinal is null. Ultimately. If we cannot succeed by counting individuals.Chap. iii. see ibid. If a ratio µ/ν is to have the expected properties. We need the whole series of ratios in order of magnitude in order to deﬁne real numbers as segments: this deﬁnition will not give the desired result unless the series of ratios is compact. then if we form the complete set of individuals. the null-class of classes and the class whose only member is the null-class of individuals). We need the whole class of inductive cardinals in order to establish the existence of ℵ . Thus no such assumption as the axiom of inﬁnity is required in order to reach any given ratio or any given inductive cardinal. at the next stage. and the whole series in order to establish the existence of progressions: for these results. we shall reach a stage where there are many more than µ objects. . merely by travelling along the hierarchy for a suﬃcient distance. n which is ℵ . however few individuals there may be in the world. classes of classes. On . XIII. Thus taking all kinds of objects together... ii. vol. It may be said: Let us assume that the number of individuals is n. namely. ad inf. there must be enough objects of whatever sort is being counted to insure that the null-class does not suddenly obtrude itself. for any given ratio µ/ν. without the axiom of | inﬁnity. which it cannot be if the total number of ratios. and so on. But this can be insured. On the corresponding problems as regards ratio. It is when we wish to deal with the whole class or series of inductive cardinals or of ratios that the axiom is required. the axiom of inﬁnity could be proved. this would still be true. ∗ ﬀ. where n may be without spoiling our argument. we can try counting classes of individuals. and so on. by means of constructions such as we have been considering. and not | conﬁning ourselves to objects of any one type. In this way any assigned inductive cardinal can be made to ﬁnd its place among numbers which are not null. we can try classes of classes. by merely travelling up the hierarchy a suﬃcient distance. .

classes of individuals. Since all its sub-classes will be members of it.” To explain the subject of “types” fully would require a whole volume. the following:— . and in connection with the axiom of inﬁnity it is particularly easy to see the necessity of some such doctrine. I attempted to discover some ﬂaw in Cantor’s proof that there is no greatest cardinal. But the need of some doctrine of types is less doubtful than the precise form the doctrine should take. before going into this argument. It would be a mistake to lay too much stress on such feelings of hocus-pocus. So it might be said. though he may be unable to say where the ﬂaw is in the above construction. will feel convinced that it is impossible to manufacture an inﬁnite collection out of a ﬁnite collection of individuals. The Axiom of Inﬁnity and Logical Types obtain an inﬁnite class. The fallacy involved is the fallacy which may be called “confusion of types. confused. if he has a robust sense of reality. the ﬁrst thing to observe is that there is an air of hocus-pocus about it: something reminds one of the conjurer who brings things out of the hat. like other emotions. add together into one class the individuals. etc. from the “contradiction of the greatest cardinal.Chap. and we inferred that there is no greatest cardinal number. When I ﬁrst came upon this contradiction. XIII. The class consisting of all objects that can be counted. And when the above argument is scrutinised it will. that the number of classes contained in a given class is always greater than the | number of members of the class. But they aﬀord a prima facie ground for scrutinising very closely any argument which arouses them. it is the purpose of this book to avoid those parts of the subjects which are still obscure and controversial. we should obtain a class of which its own sub-classes would be members. be found to be fallacious. for example. there cannot be more of them than there are members. those parts which can be accepted as embodying mathematically ascertained truths. in the year . which we gave in Chapter VIII. moreover. though the fallacy is a subtle one and by no means easy to avoid consistently.. have a cardinal number which is the greatest possible. they may easily lead us astray. isolating. Hence we arrive at a contradiction. and obscure. This necessity results. But if we could. as we suggested a moment ago. Now. classes of classes of individuals. for the convenience of beginners. Applying this proof to the supposed class of all imaginable objects. I was led to a new and simpler contradiction. The man who has lent his hat is quite sure there wasn’t a live rabbit in it before. but he is at a loss to say how the rabbit got there. and shall therefore not need the axiom of inﬁnity. must. if there be such a class. So the reader. in my opinion. of whatever sort.” We saw in Chapter VIII. Now the theory of types emphatically does not belong to the ﬁnished and certain part of our subject: much of this theory is still inchoate. namely.

Form now the assemblage of all classes which are not members of themselves. though not really. The Axiom of Inﬁnity and Logical Types The comprehensive class we are considering. and n classes of individuals. names for classes can occur signiﬁcantly: a sentence or set of symbols in which such pseudo-names occur in wrong ways is not false. it is a member of itself. more brieﬂy.e. chap.. i. we cannot form a new class. consisting of both individuals and classes and having n + n members. but strictly devoid of meaning. pp. classes which are not pure as to “type. For the present an outline of e the solution must suﬃce. In this way the attempt to escape from the need for the axiom of inﬁnity breaks down. for example. of another class of individuals will be to suppose nonsense. Mankind. in rough outline. or that it is not. ∗ and ∗ . This places a limitation upon the ways in which what are nominally. is not a man.e. and is a member of the class “everything. i. XIII. And more generally. and also. if there is such a thing as “everything. or is not a member. and to construct symbolically any class whose members are not all of the same grade in the logical hierarchy is to use symbols in a way which makes them no longer symbolise anything. This is a class: is it a member of itself or not? If it is.. which is to embrace everything. and that it is not..” then “everything” is something. –. –. it is one of those classes that are not members of themselves. I have aimed only at saying just so much as was required in order to show that we cannot prove the existence of inﬁnite numbers and classes by i. it is not a member of itself. “Mathematical Vol. .” vol. “Les paradoxes de la logique. a member of itself is meaningless in just this way. to suppose that one class of individuals is a member. it is not one of those classes that are not members of themselves. This is a contradiction. or done more than indicate. Introduction. I do not pretend to have explained the doctrine of types. ii. classes are logical ﬁctions. pp. Logic as based on the Theory of Types. ii. Prefatory Statement.e. There is no diﬃculty in manufacturing similar contradictions ad lib. i.Chap. The supposition that a class is. in articles by the present author in the American Journal | of Mathematics and in the Revue de M´taphysique et de Morale. In other words. The solution of such contradictions by the theory of types is set forth fully in Principia Mathematica. The fallacy consists in the formation of what we may call “impure” classes..” But normally a class is not a member of itself. Thus of the two hypotheses—that it is. . Thus if there are n individuals in the world. xxx.” As we shall see in a later chapter. must embrace itself as one of its members. If it is not. a member of itself—each implies its contradictory. and a statement which appears to be about a class will only be signiﬁcant if it is capable of translation into a form in which no mention is made of the class. vol.” . why there is need of such a doctrine.

not only on its own account. Hence the relation “idea of” constitutes a reﬂexion of the whole class of objects into a part of itself. and still more because. suggested by a passage in Plato’s Parmenides. however. and together with and being gives a class of three terms. it would be found that numbers do not have being—they are. for if n is inductive this will not be the case. There is an argument employed by both Bolzano and Dedekind to prove the existence of reﬂexive classes. but there is (at least in the realm of being) an idea of any object. given a class of n terms. the class of objects and the class of ideas are both inﬁnite. Accordingly. in fact. is quite diﬀerent from the same equation as applied to a reﬂexive number. ). Was sind und was sollen die Zahlen? No. This argument is interesting. into that part which consists of ideas. Bolzano.” as we shall see when we come to consider the deﬁnition of classes. if n is an inductive cardinal. This argument is fallacious. n is not equal to n + . The Axiom of Inﬁnity and Logical Types such conjurer’s methods as we have been examining. to the eﬀect that. There remain. it does not mean that. then has being. and therefore there is such a number as . in brief. As applied to a reﬂexive number. . . what are called “logical ﬁctions. it merely means that there is no class of n individuals. | In so far as these arguments assume that. as we have seen. The argument. § (p. There is an argument. The relation of an object to the idea of it is one-one. this class is “similar” to that obtained by adding another term. partly because “being” is not a term having any deﬁnite meaning. if a deﬁnite meaning were invented for it. and no class of n + individuals. namely. But as applied to a number which is too great for the actual world. The argument that the number of numbers from to n (both inclusive) is n + depends upon the assumption that up to and including n no number is equal to its successor. they have been already dealt with. It must be understood that the equation n = n + . Various arguments professing to prove the existence of inﬁnite classes are given in the Principles of Mathematics. which. we shall then ﬁnd this class “similar” to one of n + terms. but because the mistakes in it (or what I judge to be Dedekind. and so on. and ideas are only some among objects. which might be true for a ﬁnite n if n exceeded the total number of individuals in the world. certain other possible methods which must be considered. if there is such a number as . it means that. if we mount the hierarchy of types suﬃciently far to secure the existence of a class of n terms. will not be always true if the axiom of inﬁnity is false.Chap. quite independently of the truth or falsehood of the axiom of inﬁnity. Paradoxien des Unendlichen. and therefore and being are two. . is this: An object is not identical with the idea of the | object. but is not identical with being. XIII.

Beyond the third or fourth stage they become mythical. as we have already noted. whatever may be said of logical arguments. to “meine Gedankenwelt”—the world of my thoughts.” or as “the husband of Xantippe. which are not relevant to our main purpose. We are then to suppose that.Chap. exceedingly diﬃcult to decide what is meant by an “idea”. and so on ad inf. Socrates (e. We certainly cannot obtain this result empirically. but let us assume that we know. as Dedekind does. the empirical arguments derivable from space and time. The Axiom of Inﬁnity and Logical Types mistakes) are of a kind which it is instructive to note. each of which could be called an idea of the object in the sense in which we might say “my idea of Socrates is quite diﬀerent from yours. and then the idea of the idea of Socrates. the “ideas” intended must be Platonic ideas laid up in heaven. it may be identical with the object. starting (say) with Socrates. the diversity of colours. for certainly they are not on earth. In the former case the argument fails. there is the idea of Socrates. In the second case the argument also fails. it must be on the basis of some logical theory. or apply it. etc. Nor. of course. Now it is plain that this is not the case in the sense that all these ideas have actual empirical existence in people’s minds. proving that it is necessary to a thing that there should be an idea of it. scriptions of any given object. It is.” If—to take up the remaining hypothesis—“idea” is to be interpreted psychologically. it must be maintained that there is not any one deﬁnite psychological entity which could be called the idea of the object: there are innumerable beliefs and attitudes. or it may stand for a description (in the sense to be explained in a subsequent chapter). If the argument is to be upheld.g. because the relation of object and description is not | one-one: there are innumerable correct de. But a few further points should be noted. But then it at once becomes doubtful whether there are such ideas. we should have to enter upon a number of psychological and logical inquiries.) may be described as “the master of Plato. are quite suﬃcient to prove the actual exis- .” but there is not any central entity (except Socrates himself) to bind together various “ideas of Socrates. If “idea” is to be understood logically.” and thus there is not any such one-one relation of idea and object as the argument supposes. The main error consists in assuming that there is an idea of every object.” or as “the philosopher who drank the hemlock. is it true psychologically that there are ideas (in however extended a sense) of more than a tiny proportion of the things in the world. XIII. of course. the above argument in favour of the logical existence of reﬂexive classes must be rejected. If we were concerned to examine fully the relation of idea and object.. For all these reasons. because it was essential to the proof of reﬂexiveness that object and idea should be distinct. If we are to know that there are. It might be thought that.

not mathematical ﬁctions. We naturally regard space and time as continuous. illustrates the fact that physics can never aﬀord proof of continuity. whether true or false. whether it is true or false in this world. The conclusion is. therefore. some inﬁnite. at least. as compact. but a few words before we leave this topic may do something to diminish the obscurity which would otherwise envelop the meaning of these words. and we have no means of knowing to which of these two kinds our actual world belongs. “Brutus killed Cæsar” expresses a relation between Brutus and Cæsar. for the present I am merely suggesting them for the reader’s consideration.” where Cæsar . that some of the possible worlds are ﬁnite. from the substantives which express the subject of the attribute or the terms of the relation. as anyone may discover in a cinema. It would be impossible to explain them adequately without a longer disquisition on the theory of types than would be appropriate to the present work. The Axiom of Inﬁnity and Logical Types tence of an inﬁnite number of particulars. or. to adopt a Leibnizian phraseology. The senses are not suﬃciently exact to distinguish between continuous motion and rapid discrete succession. at any rate in the sense in which space and time are physical facts. A world in which all motion consisted of a series of small ﬁnite jerks would be empirically indistinguishable from one in which motion was continuous. we must conclude that nothing can be known a priori as to whether the number of things in the world is ﬁnite or inﬁnite. expressing an attribute or relation. since the same occurrence may be expressed in the words “Cæsar was killed by Brutus. we may call both Brutus and Cæsar subjects of this proposition: the fact that Brutus is grammatically subject and Cæsar object is logically irrelevant. though it is theoretically conceivable that some day there might be evidence pointing. We have no reason except prejudice for believing in the inﬁnite extent of space and time. in that direction. but is also not demonstrable logically. XIII. I do not believe this. The theory of “quanta” in physics. though it might quite possibly aﬀord disproof. it follows that there is no empirical reason for believing the number of particulars in the world to be inﬁnite. though not conclusively. but this again is mainly prejudice. From the fact that the inﬁnite is not self-contradictory. also that there is at present no empirical reason to believe the number to be ﬁnite. In an ordinary statement we can distinguish a verb. If they are valid. we cannot tell. Using the word “subject” in a generalised sense. “Cæsar lived” ascribes an attribute to Cæsar.Chap. The axiom of inﬁnity will be true in some possible worlds and false in others. and that there never can be. It would take up too much space to | defend these theses adequately. Throughout this chapter the synonyms “individual” and “particular” have been used without explanation.

“A gives B to C” is a relation of three terms. It also happens that the verb may grammatically be made subject: e. it is false throughout this hierarchy. “Killing is a relation which holds between Brutus and Cæsar.” It is to the number of these that the axiom of inﬁnity is assumed to apply. and it is these that give the meaning of “particulars” or “individuals. We shall further deﬁne “individuals” or “particulars” as the objects that can be named by proper names. we may say. This is part of the old scholastic deﬁnition of substance. similarly if it is false of them.) Now it often happens that. when they occur in propositions. the apparent subjects are found to be not really subjects. XIII. however. forms no part of the notion with which we are concerned. the only result of this. If this be the case. can only occur as subjects. We are thus led to the conception of terms which. But whether the axiom is true or false. which belonged to that notion. is that new subjects take their places. on a closer scrutiny. (A relation may have more than two terms: e.Chap. (It would be better to deﬁne them directly. . of course. The Axiom of Inﬁnity and Logical Types is the grammatical subject. there seems no known method of discovering. and in a straightforward statement.) It is. it must be theoretically possible for analysis to reach ultimate subjects. and classes of classes of them. but to be capable of analysis. Hence it is natural to enunciate the axiom concerning them rather than concerning any other stage in the hierarchy. rather than by means of the kind of symbols by which they are symbolised. and so on. but in order to do that we should have to plunge deeper into metaphysics than is desirable here. the axiom of inﬁnity must of course be true. We shall deﬁne “proper names” as those terms which can only occur as subjects in propositions (using “subject” in the extended sense just explained). | Thus in the simpler sort of proposition we shall have an attribute or relation holding of or between one. Brutus and Cæsar will appear as the subjects and killing as the verb. it is true | of classes of them.” But in such cases the grammar is misleading. but persistence through time. on closer scrutiny. following the rules that should guide philosophical grammar. two or more “subjects” in the extended sense. and never in any other way. But if it be not the case. a class or some kind of complex. possible that there is an endless regress: that whatever appears as a particular is really.g.g. If it is true of them.

() descriptions. somewhat hastily it is true. Before we can undertake it. namely: () the theory of deduction. it is no defence to urge that common sense shows the result to be correct. that part of the philosophy of mathematics which does not demand a critical examination of the idea of class. however. that will concern us in the present chapter.” The whole trend of modern . ought to be needed in mathematics after the premisses have been laid down. the third is not logically presupposed in the theory of classes. In a synthetic treatment. which we have hitherto ignored. the proof is defective. | rather than bring fallacy to the rescue of com. it would be better to dispense with argument altogether. No appeal to common sense. it is true also that perfect rigour is a scarcely attainable ideal.CHAPTER XIV INCOMPATIBILITY AND THE THEORY OF DEDUCTION We have now explored. It is the ﬁrst topic. in so far as rigour is lacking in a mathematical proof. but always requires the support of what is called “intuition.” or anything except strict deductive logic. but required an appeal to the ﬁgure. Of these. by a strict process of deduction. Kant. we found ourselves confronted by problems which make such an examination imperative. Nevertheless. it arrives. at the various theorems which constitute it. for if we were to rely upon that. invented a theory of mathematical reasoning according to which the inference is never strictly logical. Three topics will concern us before we reach the theory of classes. mon sense. having observed that the geometers of his day could not prove their theorems by unaided argument. the theory of deduction. in the past. Mathematics is a deductive science: starting from certain premisses. It is true that. In the preceding chapter. but it is a simpler example of the kind of theory that is needed in dealing with classes. mathematical deductions were often greatly lacking in rigour. () propositional functions. the parts which we shall now be concerned with come ﬁrst: they are more fundamental than anything that we have discussed hitherto. or “intuition. we must consider certain other parts of the philosophy of mathematics.

with its increased pursuit of rigour. In order to be able validly to infer the truth of a proposition. The positive grounds for this thesis are to be found in Principia Mathematica. We cannot here do more than refer the reader to those works. i.” so that the knowledge that the one is false allows us to infer that the other is true. (We shall deﬁne this relation shortly. we have one or more propositions called premisses. It is this relation that is chieﬂy of interest in the logical theory of deduction.e. passim. q. In deduction. so as to be able to speak of the premiss as well as of the conclusion. from which we infer a proposition called the conclusion. r. All these four are cases of inference. since this is the relation which must We shall use the letters p. in just the same circumstances in which the truth of the other might have been inferred from the truth of the one. through the senses or through experience in some form. the conclusion. and proceed to inquire as to what is involved in deduction. we shall assume that all mathematics is deductive.” i. Meanwhile.e. the axiom of parallels. XIV. we must know that some other proposition is true.Chap. What else is to belong to human knowledge must be ascertained otherwise— empirically. provided we know that the two are “incompatible. in mathematics and by mathematical methods. This may be inferred from the truth of another proposition. the other is false. When our minds are ﬁxed upon inference. Again. it will be convenient. Incompatibility and the Theory of Deduction mathematics. unless there is such a relation between premiss and conclusion that we have a right to believe the conclusion | if we know the premiss to be true.e.e. has been against this Kantian theory. Thus we may regard deduction as a process by which we pass from knowledge of a certain proposition. What can be known.) Or we may know that a certain other proposition is false. a controversial defence of it is given in the Principles of Mathematics.” expressed by “p or q. and that there is between the two a relation of the sort called “implication. not its truth. when there are originally several premisses. t to denote variable propositions. and that there is a relation between the two of the sort called “disjunction. It may also be inferred from the falsehood of another proposition. cannot be known—for example. that if one is true. But we shall not regard such a process as logical deduction unless it is correct. is what can be deduced from pure logic. to amalgamate them into a single proposition. the premiss. The things in the mathematics of Kant’s day which cannot be proved. s. from the falsehood of p we may infer the falsehood of q. . i. it seems natural to take “implication” as the primitive fundamental relation. For our purposes.” i. that (as we say) the premiss “implies” the conclusion. since the subject is too vast for hasty treatment. when q implies p. but not a priori. what we wish to infer may be the falsehood of some proposition. to knowledge of a certain other proposition.

i. disjunction.” or “if p. Before proceeding to primitive ideas and deﬁnitions. i. But all ﬁve agree in this.” but the above ﬁve will suﬃce. whereas the others are functions of two. this is the meaning which is convenient for us. Last take implication. It is convenient to speak of the truth of a proposition.” The whole meaning of a truth-function is exhausted by the statement of the circumstances under which it is true or false. likewise if q is true. otherwise it has falsehood for its truth-value. truth is the “truth-value” of a true proposition.e. its truth-value is falsehood when p and q are both true. or implication. i. and falsehood of a false one.” This is a function whose truth-value is truth when p is true and also when q is true. A function of propositions which has this property is called a “truth-function.e.” Its truth-value is truth when p is false and likewise when q is false. Thus we interpret it as meaning: “Unless p is false. Given the truth or falsehood of p.” This has truth for its truth-value when p and q are both true. incompatibility. “p and q. “not-p. disjunction. it is “not-p or not-q. or of p and q (as the case may be).e. i. or its falsehood. that their truth-value depends only upon that of the propositions which are their arguments.” This is that function of p which is true when p is false. We have thus ﬁve functions: negation. incompatibility. and is to be falsehood if p is true and q is false.” This is the negation of conjunction. “p implies q. “p implies q” is to mean “not-p or q”: its truth-value is to be truth if p is false. it is also the disjunction of the negations of p and q. “p or q. and implication.e. Take next incompatibility.Chap. Thus not-p has the opposite truth-value to p. and false when p is true. Next we may take conjunction. conjunction. conjunction. as its “truth-value” . “p and q are not both true. “not-p and not-q. Incompatibility and the Theory of Deduction hold between p and q if we are to be able to infer the truth of q from the truth of p. “Not-p.” (The fact that “implies” is capable of other meanings does not concern us. but is falsehood when both p and q are false. let us consider further the various functions of propositions suggested by the above-mentioned relations of propositions.” This term is due to Frege.” This is to be understood in the widest sense that will allow us to infer the truth of q if we know the truth of p. we are given the truth or falsehood of the negation. But for technical reasons this is not the best primitive idea to choose. We might have added others. then q. XIV. q is true. for example. Negation diﬀers from the other four in being a function of one proposition. joint falsehood.” or “either p is false or q is true. . | We may take next disjunction.) That is to say. The simplest of such functions is the negative.

and by Nicod that this enables us to reduce the primitive propositions required in the theory of deduction to two non-formal principles and one formal one. is a certain truth-function called “incompatibility. to be true for all values of p and q. Math. XIV. which is Trans. the two chosen in Principia Mathematica are negation and disjunction. January .e. –.” which we will denote by p /q. But it has been shown by Sheffer that we can be content with one primitive idea for all ﬁve.. Negation can be at once deﬁned as the incompatibility of a proposition with itself.Chap. which is known. i. There is no great diﬃculty in reducing the number to two. Incompatibility and the Theory of Deduction for example. it is (p /q) | (p /q). For example. p | (q /q). | If we know that p is true and that p implies q. We can deﬁne some of them in terms of others. we can proceed to assert q. pp. We will choose the former. Conjunction is the negation of incompatibility. “p and q” is the negation of “not-p or not-q” and vice versa. in virtue of its form. xiv. is simply that function of p which is true when p is false.e. Soc. In mathematical practice. conjunction as the negation of incompatibility. . What we are concerned with is the connection of this subject with inference. but the actual passage from the assertion of p to the assertion of q is a psychological process. Implication is then deﬁned as “not-p or q”. vol. “not-p” is deﬁned as “p /p. it is (p /p) | (q /q). thus either of these may be deﬁned as the negation of the other. either by introducing more arguments or by repeating arguments.” Disjunction is the incompatibility of not-p and not-q. now. i. we have always some expression containing variable propositions. It is clear that the above ﬁve truth-functions are not all independent. It is obvious that there is no limit to the manufacture of truthfunctions. and false when p is true: there is no further | meaning to be assigned to it. Am. i. There is always unavoidably something psychological about inference: inference is a method by which we arrive at new knowledge. Camb. say p and q.e. xix. For this purpose. i. The same applies to “p or q” and the rest. we have also some other expression. There is no further meaning in a truth-function over and above the conditions under which it is true or false. Phil. Implication is the incompatibility of p and not-q. and what is not psychological about it is the relation which allows us to infer correctly. Our primitive idea.e.. incompatibility as “not-p or not-q”.. vol. when we infer. and we must not seek to represent it in purely logical terms. It follows that two truth-functions which have the same truth-value for all values of the argument are indistinguishable. i. part of the former. Soc.. Proc. we may take as our one indeﬁnable either incompatibility or joint falsehood. Thus all our four other functions are deﬁned in terms of incompatibility.

in any convenient form. then if q implies r it follows that p implies r.e. It has a use as the premiss of an inference. inevitably introduce an order. not only as rules. () If either p is true or “q or r” is true.e. and assert what is left. then “p or q” implies “p or r. Put p = p implies q. () q implies “p or q”—i. .) These ﬁve propositions are as follows:— () “p or p” implies p—i. for example.’” i. We have here a relation of three propositions which state implications.” Call this proposition A. then ‘p implies q’ implies ‘p implies r.” but also as substantive premisses. Let us assume that we know the ﬁve formal principles of deduction enumerated in Principia Mathematica. i. p = p implies r.) | () If q implies r.” These are the formal principles of deduction employed in Principia Mathematica. as the p of our schema. we are able to drop this part of our original expression.” from which we infer q. Now when we are concerned with the principles of deduction. and in virtue of the principles of inference. Nicod has reduced these to one. A formal principle of deduction has a double use. which is their use for establishing “p implies q. () “p or q” implies “q or p. Then we have to prove that p implies that p implies p . the disjunction “p or q” is true when one of its alternatives is true. In the schema of an inference we have a proposition p. “p implies that p implies p .e. then either q is true or “p or r” is true. and it is in order to make this clear that we have cited the above ﬁve propositions. That is to say. p = q implies r. our rules of deduction are to be used. but as it is a complicated proposition. XIV. then p is true. and remember that “not-p or q” is by deﬁnition the same as “p implies q. we will begin with the ﬁve. (M. Incompatibility and the Theory of Deduction also known to be true for all values of p and q. and a proposition “p implies q. (The twist in this proposition serves to increase its deductive power. Now take the ﬁfth of our above principles. our apparatus of primitive propositions has to yield both the p and the “p implies q” of our inferences. since in the conception of disjunction there is no order involved. But since our symbols. and a use as establishing the fact that the premiss implies the conclusion. if either p is true or p is true. so that “p or q” and “q or p” should be identical. substitute not-p for p. we wish to prove that if p implies q.Chap. Suppose.” Thus our ﬁfth principle yields: “If q implies r. This somewhat abstract account may be made clearer by a few examples. we need suitable assumptions for showing that the order is irrelevant.e.” This would not be required if we had a theoretically more perfect notation.

| Now we proved by means of our ﬁfth principle that “p implies that p implies p . For this purpose we will ﬁrst show how certain truth-functions can be deﬁned in terms of incompatibility. are to be regarded as asserted for all possible values of the variable propositions p.” which was what we called A. the adaptation of our ﬁfth principle. for p and q. Incompatibility and the Theory of Deduction But the fourth of our principles. e. Nicod has reduced the ﬁve given above. becomes: “If p implies that q implies r. The legitimacy of substitutions of this kind has to be insured by means of a non-formal principle of inference. but which itself can hardly be called deduction. occurs as a substantive premiss. while the adaptation of our fourth principle. is used to give the form of the inference. then p implies that p implies p .” which was the proposition to be proved. The primitive propositions. but from a practical point of view we obtain what are virtually new propositions. “s implies t. “p implies that p implies p . whatever they may be. which yields B. provided we realise that they are in theory distinct.” | such principle is enunciated in Principia Mathematica or in M. not-q. In this proof. Nicod’s article mentioned above. namely. p in place of q. this becomes: “If p implies that p implies p . which yields A. and B represents the “p implies q.” Call this B.Chap. since A represents the p of our scheme. not-p.” Hence we arrive at q. and remember the deﬁnition of implication.” Writing p in place of p. r which occur in them. then q implies that p implies r.g. We may now state the one formal principle of inference to which M. We saw already that p | (q /q) means “p implies q. and p in place of r. By means of such substitutions we really obtain sets of special cases of our original proposition. The formal and material employments of premisses in the theory of deduction are closely intertwined.” and so on. But this would seem to be an omission. Thus we have here an instance of the schema of inference. and it is not very important to keep them separated. The earliest method of arriving at new results from a premiss is one which is illustrated in the above deduction. XIV. We may therefore substitute for (say) p any expression whose value is always a proposition. No . when we substitute not-p. q.

Incompatibility and the Theory of Deduction We now observe that p | (q /r) means “p implies both q and r. s and q are both true. in other words. it is clear that q must be true. It follows that (s /q) | p /s expresses the incompatibility of s /q with the conjunction of p and s. between propositions. as we saw. Whenever p is false.” which we shall consider in the next chapter. XIV. “p implies that q and r are both true”—for. Now. and p is true. π = t | (t /t). Let us write p for the negation of p. P implies both π and Q. which requires that p should . in other words. in virtue of which an inference is valid. it will mean the conjunction of p and s. but is useless for inference. in still simpler words.” This is a particular case of p | (q /q). In order that it may be valid to infer q from p. s /q is false. it is only necessary that p should be true and that the proposition “not-p or q” should be true. This principle is: “If p | (r /q) is true.” i. it states that p and s jointly imply s and q jointly. if I am not mistaken.e.” From this apparatus the whole theory of deduction follows.” For this expression means “p is incompatible with the incompatibility of q and r.Chap. But inference will only in fact take place when the proposition “not-p or q” is known otherwise than through knowledge of not-p or knowledge of q.e. There is. He employs in addition one non-formal principle belonging to the theory of types (which need not concern us). i. Whenever this is the case. the conjunction of q and r is the negation of their incompatibility. it states that if p and s are both true. Nicod’s sole formal principle of deduction is P | π /Q.e.” i. Observe next that t | (t /t) means “t implies itself. “p implies that q and r are not incompatible. we can assert q. except in so far as we are concerned with deduction from or to the existence or the universal truth of “propositional functions. i. then q is true. “not-p or q” is true. thus p /s will mean the negation of p /s. Q = (s /q) | p /s.e. a certain confusion in the | minds of some authors as to the relation. Then M. given p. and one corresponding to the principle that. put P = p | (q /r). and given that p implies q.

” He urges that the wider relation. . therefore. The essential point of diﬀerence between the theory which I advocate and the theory advocated by Professor Lewis is this: He maintains that. inference only arises when “not-p or q” can be known without our knowing already which of the two alternatives it is that makes the disjunction true. –. See Mind. For example. Incompatibility and the Theory of Deduction be true. it is in any case one that mathematics does not need. I conclude. It is under such circumstances that the relation of implication is practically useful for drawing inferences. pp. it is the case that we can see that either the ﬁrst is false or the second true. whenever the relation of “formal deducibility” holds between two propositions. holding only when there are certain formal connections between p and q. but is again useless for inference. In fact. the circumstances under which this occurs are those in which certain relations of form exist between p and q. . –.” which is not the relation expressed by “not-p or q” but a narrower relation. and that. that expressed by “not-p or q. “not-p or q” is of course also known to be true. that there is no need to admit as a fundamental notion any form of implication not expressible as a truth-function. a matter of words. without having ﬁrst to know that the ﬁrst is false or to know that the second is true. since q is already known. xxiii.” should not be called “implication. xxi. on general grounds of economy. But this formal relation is only required in order that we may be able to know that either the premiss is false or the conclusion is true. It is the truth of “not-p or q” that is required for the validity of the inference. Whenever q is already known to be true. and therefore one that. the reasons of detail which Professor Lewis adduces against the view which I advocate can all be met in detail. and vol. ﬁnally. whether or not there be such a relation as he speaks of. I. Lewis has especially studied the narrower.. however. pp. formal relation which we may call “formal deducibility. and that nothing beyond this fact is necessary to be admitted into our premisses. ought not to be admitted into our apparatus of fundamental notions. then s implies the negation of r. when one proposition q is “formally deducible” from another p. Now. | Provided our use of words is consistent. and depend for their plausibility upon a covert and unconscious assumption of the point of view which I reject. what is required further is only required for the practical feasibility of the inference. Professor C.Chap. XIV. I maintain that. Between “r implies not-s” and “s implies not-r” there is a formal relation which enables us to know that the ﬁrst implies the second. and therefore does not need to be inferred. the relation which we perceive between them is one which he calls “strict implication. vol..” That is. it matters little how we deﬁne them. we know that if r implies the negation of s. . that.

so long as x remains undetermined.” and further to such symbols as give expression to truth and falsehood.” because I do not wish to exclude other than verbal symbols. the values must actually enunciate propositions. but the bare formula “(a + b) = a + ab + b ” alone is not.” But although the word cannot be formally deﬁned. in the enunciation of mathematical formulæ. and so will “Socrates is a man” and “Socrates is not a man.” A “propositional function. we were discussing propositions. Examples of propositional functions are easy to give: “x is human” is a propositional function. that a and b are to have all possible values. (a + b) = a + ab + b ” is a proposition. in the preceding chapter. in order to avoid the very common confusion with “propositional functions. it is a function whose values are propositions. | such that. But I think the word “proposition” should be limited to what may. is an expression containing one or more undetermined constituents.” in fact. I say “primarily. it is . e.g. or even mere thoughts if they have a symbolic character. or are to have such-and-such values. or led to suppose. we did not attempt to give a deﬁnition of the word “proposition. the expression becomes a proposition.CHAPTER XV PROPOSITIONAL FUNCTIONS When.” The statement: “Whatever numbers a and b may be. although its values are propositions. as a rule. since it asserts nothing deﬁnite unless we are further told. But in such a case the propositions are only described: in a propositional function. be called “symbols. We mean by a “proposition” primarily a form of words which expresses what is either true or false.” will not be a propositional function. they would be “propositional functions. A descriptive function. “the hardest proposition in A’s mathematical treatise. The former of these is tacitly assumed. when values are assigned to these constituents. Thus “two and two are four” and “two and two are ﬁve” will be propositions. In other words. in some sense. But this latter deﬁnition must be used with caution. which thus become propositions.” which are to be the topic of the present chapter. it is necessary to say something as to its meaning. but if no such assumption were made.

x is followed by thunder. false for other values.” This phrase is only applicable to propositional functions. broadly speaking. but only .e. true for values of the co-ordinates belonging to points on the curve or surface. in an analogous way. XV. Propositional Functions neither true nor false. the question: “What is a propositional function?” A propositional function standing all alone may be taken to be a mere schema. i. Any mathematical equation is a propositional function. The equation to a curve in a plane or to a surface in space is a propositional function.” The process of generalisation (with whose validity we are | fortunately not concerned) consists in passing from a number of such instances to the universal truth of the propositional function: “If x is a ﬂash of lightning.” it is clear that the “something” involved cannot be a proposition. We do not need to ask. When we say that something is “always true” or “true in all cases. for example. in two ways: ﬁrst. A proposition is just true or false. the ﬁrst must occupy us now. it becomes true when the variable is made equal to a root of the equation. but when a value is assigned to x it becomes a true or false proposition. We are concerned with propositional functions. and there is an end of the matter. The notion of “cases” or “instances” depends upon propositional functions. the sort of thing that is often said when causation is being discussed. or attempt to answer. (We are not concerned with the truth or falsehood of what is said. Take.” We have a number of “instances” of this. as involved in the notions “true in all cases” and “true in some cases”. but if it is an “identity” it will be true when the variable is any number. for example. an empty receptacle for meaning. If it is an equation containing one variable. The second of these topics we will postpone to a later chapter. “lightning is followed by thunder. as involved in the theory of classes and relations. There are no instances or cases of “Socrates is a man” or “Napoleon died at St Helena. a number of propositions such as: “this is a ﬂash of lightning and is followed by thunder. propositional functions are always involved whenever we talk of instances or cases or examples. a mere shell. secondly. and it would be meaningless to speak of their being true “in all cases. So long as the variables have no deﬁnite value.Chap. otherwise it becomes false.” and let us take some very primitive example.” These are propositions.” It will be found that. the kind of process suggested by what is called “generalisation. Expressions of traditional logic such as “all A is B” are propositional functions: A and B have to be determined as deﬁnite classes before such expressions become true or false. not something already signiﬁcant. the equation is merely an expression awaiting determination in order to become a true or false proposition. x is followed by thunder.” What are these occurrences “instances” of? They are instances of the propositional function: “If x is a ﬂash of lightning. Consider. say.

whatever x may be. as being true only of this or that particular p or q. “‘p or q’ implies ‘q or p. A must be some general concept of which it is signiﬁcant to say “x is A. All the other uses of propositional functions can be reduced to these two. “if a is human.” then φa is to be true. But the separate ﬂashes are particulars. followed by B. in the last analysis. to return to our example. though it cannot itself be true or false. We do not assert the above principle. without any temporal suggestion). is true in certain instances and false in certain others. assuming that there is to be no necessary implication of a plurality of cases).” “x is A. When we say that a propositional function is true “in all cases. x are particulars which are not identical one with another. for example.g.” “some” require propositional functions for their interpretation. that is. This applies. We say that lightning (A) is followed by thunder (B).” Or.” Sentences involving such words as “all.” “every.” “a. however a may have been chosen. namely. in fact. e. in every instance.g.) We are told that A is. The study .” “the.” When. if x is an A. a is mortal” is true whether a is human or not. or in some cases (as we shall say. x .” and so on. XV. x is mortal” is “always true. For example. unless it is “always true” or “always false. but sharing the common property of being lightning.” or “true in all cases. the other to assert that it is true in at least one case. Thus the propositional function “if x is human. we mean that. we mean that all its values are true. every proposition of this form is true. but as being true of any p or q concerning which it can be made signiﬁcantly..” “x is A. The condition that a function is to be signiﬁcant for a given argument is the same as the condition that it shall have a value for that argument. In this case all the objects are “instances” of the truth of the propositional function—for a propositional function. If “φx” is the function. again.” There are. to our previous case of lightning. the statement “there are no unicorns” is the same as the statement “the propositional function ‘x is not a unicorn’ is true in all cases. we say that A is in every instance followed by B. e. and a is the right sort of object to be an argument to “φx. not identical. “all” and “some. we are asserting that a certain propositional function is “always true.” or “always” (as we shall also say. only two things that can be done with a propositional function: one is to assert that it is true in all cases. where x . Propositional Functions with its logical analysis. The way in which propositional functions occur can be explained by means of two of the above words. it is followed by a B. The only way of expressing a | common property generally is to say that a common property of a number of objects is a propositional function which becomes true when any one of these objects is taken as the value of the variable.’” are really assertions | that certain propositional functions are true in all cases. Now if there are “instances” of A.” The assertions in the preceding chapter about propositions. either true or false.Chap.

it is true for some values of α that α is a class of n individuals’ is true for all possible values of n.” contains two variables. is: “The propositional function ‘if n is an inductive number. Thus the statement “all men are mortals” is the negation of the statement that the function “x is an immortal man” is sometimes true.e. namely. in respect of α. When we say “cannibals still exist in Africa.Chap. We shall return in our ﬁnal chapter to the discussion of propositional functions containing no constant terms. sometimes true. α and n. or what not—and clearly it is not the province of logic to make assertions which are true concerning one such thing or concept but not concerning another. the assertion that it is “sometimes true. To say “there are at least n individuals in the world” is to say that the propositional function “α is a class of individuals and a member of the cardinal number n” is sometimes true. they would have to mention particular things or concepts—Socrates.” i. or. or redness. i. Propositional Functions of the conditions of signiﬁcance belongs to the doctrine of types.” that means that the propositional function “x is a man” is sometimes true.e. Not only the principles of deduction.” that means that the propositional function “x is a man and a Greek” is sometimes true.e. “α is a class of n individuals. If this were not the case. true in at least one instance. When we say “there are men. they all consist of the assertion that some propositional function containing no constant terms is always true. but all the primitive propositions of logic. The statement that a function φx is always true is the negation of the statement that not-φx is sometimes true.” Here there is a subordinate function. in the language of propositional functions. This form of expression is more convenient when it is necessary to indicate which is the variable constituent which we are taking as the argument to our propositional function. consist of assertions that certain propositional functions are always true. For the present we will proceed to the other thing that is to be done with a propositional function. is true for some values of x. The axiom of inﬁnity. or east and west.” which is said to be. When we say “some men are Greeks. always true. and the assertion that this happens if n is an inductive number is said to be. XV. which we shall not pursue beyond the sketch given in the preceding chapter.” that means that the propositional function “x is a cannibal now in Africa” is sometimes true. in respect of n. the above propositional function. and the statement that φx is sometimes true is the negation of the statement that not-φx is always true. is true for certain | values of α. For example. which we may shorten to “α is a class of n individuals. It is part of the deﬁnition of logic (but not the whole of its deﬁnition) that all its propositions are completely general. as we may say. And the statement “there are unicorns” is the negation of . i.

to begin with. if P is mortals. Take. to the theory of truth-functions as applied to propositions containing one. we deﬁne: “The negation of ‘φx always’ is ‘not-φx sometimes’. That is to say. Propositions containing no apparent variables are called “elementary propositions. or any number up to n. take one of the pair “always.Chap. We can. it is often convenient to say “φx is not always false” rather than “φx sometimes” or “φx is sometimes true. three .g. We may put the matter another way. and the negation of ‘φx sometimes’ is ‘not-φx always. assuming that we have already | deﬁned (or adopted as a primitive idea) the negation of propositions of the type to which φx belongs. that if x is an S. if we choose. in terms of the deﬁnitions and primitive ideas for propositions containing no apparent variables. as applied to propositions containing apparent variables. and all involve the assertion of all values or some values of a compound propositional function. and deﬁne the other by means of the one and negation. vol. for the essence of this method consists in using implications in cases where (as it afterwards turns out) the hypothesis is false. i. The forms which are taken as simplest in traditional formal logic are really far from being so. provided we know what is meant by being an S and what by being linguistic reasons.” The method of deduction is given in Principia Mathematica. φx will be “x is human”. then x is a P. namely. and deﬁne by their means the negation of propositions in which they occur. still. variables. Propositional Functions the statement that the function “x is not a unicorn” is always true.” We will take it that S is deﬁned by a propositional function φx. and P by a propositional function ψx. ∗ . we can deﬁne: “‘φx is always true’ is to mean ‘it is false that not-φx is sometimes true. ψx will be “there is a time at which x dies.” “sometimes” as a primitive idea. And this is every bit as true when x is not an S as when x is an S. two. if S is men. “all S is P. our statement “all S is P” tells us something about x. XV.’” In like manner we can re-deﬁne disjunction and the other truth-functions. it says something equally about terms which are not S’s. to avoid suggesting either the plural or the singular. Suppose we come across an x of which we do not know whether it is an S or not. using such methods as have just been indicated. Thus if we choose “sometimes” as our primitive idea.” From these we can mount up step by step.” Then “all S is P” means: “‘φx implies ψx’ is always true. where n is any assigned ﬁnite number.’” But for reasons connected with the theory of types it seems more correct to take both “always” and “sometimes” as primitive ideas.. the reductio ad absurdum would not be a valid method. . . In order to understand “all S is P.” it is not necessary to be able to enumerate what terms are S’s. For . If it were not equally true in both cases. E.” It is to be observed that “all S is P” does not apply only to those terms that actually are S’s. We say that φx is “never true” or “always false” if not-φx is always true.

This shows that it is not merely the actual terms that are S’s that are relevant in the statement “all S is P. whether true or false.” | and then we will regard “Socrates” as replaced by a variable x wher. all for which it has a value at all. ever “Socrates” occurs. i. but truth-functions of φx and ψx for the same argument x. yet it is to have the same value in “φx” as in “ψx” when we are asserting that “φx implies ψx” is always true. but from φa and ψa. i. all the terms that are S’s.” It will be observed that the propositional functions which are here asserted for all or some values are not φx and ψx themselves. .” “Some S is P” means “‘φx and ψx’ is sometimes true.) Then: “All S is P” means “‘φx implies ψx’ is always true.e.Chap. we can understand completely what is actually aﬃrmed | by “all S is P. The easiest way to conceive of the sort of thing that is intended is to start not from φx and ψx in general. and P is those for which ψx is true. Suppose we are considering “all men are mortal”: we will begin with “If Socrates is human. XV. means that “x is human” is true for some values of x. We may now proceed with our interpretation of the traditional forms of the old-fashioned formal logic.e.” “Some S is not P” means “‘φx and not-ψx’ is sometimes true. We assume that S is those terms x for which φx is true. and not only those that in fact are human. together with all the terms that are not S’s—i.g. (As we shall see in a later chapter. Propositional Functions a P. all values for which “x is human” is signiﬁcant. Here all values of x (i. “There are men. for if we start with two separate functions we can never secure that the x. all classes are derived in this way from propositional functions.e.) Every assertion about “all” or “some” thus involves not only the arguments that make a certain function true. The object to be secured is that. whether true or false) are relevant.. Socrates is mortal. without any deﬁnite value. (This becomes obvious if we consider how we could prove such a statement to be false.” e. while remaining undetermined. the whole of the appropriate logical “type.” but all the terms concerning which the supposition that they are S’s is signiﬁcant. This requires that we shall start with a function whose values are such as “φa implies ψa. where a is some constant.” however little we may know of actual instances of either. shall have the same value in both.” “No S is P” means “‘φx implies not-ψx’ is always true.” rather than with two separate functions φx and ψx. although x remains a variable.e.” What applies to statements about all applies also to statements about some. but all that make it signiﬁcant.

e. It will be seen that “all S is P” and “no S is P” do not really diﬀer in form.e. but the fundamental form is that which is derived immediately from the notion of “sometimes true. e. “φx implies ψx” means “not-φx or ψx. and some moods of the syllogism are fallacious. For. i.Chap. which was eﬀected by Peano and Frege. with our deﬁnitions. that such propositions as “all S is P” do not involve the “existence” of S’s. this name is given equally if there are several variables. which is the only technically tolerable one. it treats “all men are mortal” as of the same form as “Socrates is mortal.” while the second is of the form “ψx.” which is needed a hundred times for once that the other is needed. except by the substitution of not-ψx for ψx.” with which traditional logic begins. then “all S is P” and “no S is P” will both be true. thus conversion per accidens becomes invalid. XV.” for then we should have no language left for “φx always implies ψx. The above deﬁnitions show how far removed from the simplest forms are such propositions as “all S is P. and that the same applies to “some S is P” and “some S is not P. this result might lead the reader to desire diﬀerent deﬁnitions.. one of which will occupy us in the next chapter. But.” which fails if there is no M.” The emphatic separation of these two forms.g. this is the same sense in which the roots of an equation are said to satisfy the equation. i.” Other meanings . do not require that there should be terms which are S’s. if we adopt the view. Darapti: “All M is S. was a very vital advance in symbolic logic. and it would be very awkward to give this as the deﬁnition of “all S is P.” since the ﬁrst allows the non-existence of S and the second does not.” We say that an argument a “satisﬁes” a function φx if φa is true. The notion of “existence” has several forms. therefore some S is P. “all S is P” does not imply “some S is P. all M is P. Now if φx is sometimes true. The above deﬁnitions lead to the result that.g. if there are no S’s. according to the deﬁnition in the last chapter. At the ﬁrst moment. or we may say “arguments satisfying φx exist.” It should also be observed that the traditional rules of conversion are faulty.” This is the fundamental meaning of the word “existence. It is typical of the lack of analysis involved that traditional logic treats “all S is P” as a proposition of the same form as “x is P”— e. if φx is always false. and φx is sometimes true” is essentially composite. we may say there are x’s for which it is true.” As we have just seen. The proposition “φx always implies ψx. | whatever P may be. Propositions of the form “φx always implies ψx” are called “formal implications”. Propositional Functions For brevity we say “φx always implies ψx” when we mean that “φx implies ψx” is always true. the ﬁrst is of the form “φx always implies ψx.” which is always true if not-φx is always true. but a little practical experience soon shows that any diﬀerent deﬁnitions would be inconvenient and would conceal the important ideas.

Chap. It will be found that by bearing in mind this simple fallacy we can solve many ancient philosophical puzzles concerning the meaning of existence. therefore Socrates exists. In the case of propositional functions. if none. (Sometimes contingent or assertoric is used instead of possible. namely. The fallacy is closely analogous to that of the argument: “Men are numerous. or embody mere confusion of thought. the threefold division is obvious. Socrates is a man. those whose contradictories were necessary. as in regard to modality in general. Generally.” This is a situation which is general in probability problems and not uncommon in practical life—e. it is possible. since “Socrates” is not. or rather meaningless. the propositional function is relevant. If “φx” is an undetermined value of a certain propositional function. In all such | cases. Socrates is a man. Another set of notions as to which philosophy has allowed itself to fall into hopeless confusions through not suﬃciently separating propositions and propositional functions are the notions of “modality”: necessary. for example. possible if it is sometimes true. there was never any clear account of what was added to truth by the conception of necessity. to ascribe existence to a given particular x who happens to be a man. This sort of situation arises in regard to probability. while among false propositions some were impossible. in many very diverse directions. it will be necessary if the function is always true.” In this case it is obvious that the conclusion is nonsensical. the habit of keeping propositional functions sharply separated from propositions . We may correctly say “men exist.” merely an undetermined argument to a given propositional function. some were necessary. if some are white. possible. Suppose a ball x is drawn from a bag which contains a number of balls: if all the balls are white. devoid of signiﬁcance. Here all that is known about x is that it satisﬁes a certain propositional function.g. and impossible. for reasons which will appear more fully in the next chapter. For clear thinking. “terms satisfying φx exist” means “φx is sometimes true”.) The traditional view was that. “x is white” is necessary. For the present let us merely note the fact that.” we are talking nonsense. “x was a ball in the bag. In fact. while others were merely contingent or assertoric.” meaning that “x is a man” is sometimes true. But if we make a pseudo-syllogism: “Men exist. however. while others merely happened not to be true. but | in the case of existence it is not obvious. XV. and impossible if it is never true. though it is correct to say “men exist. it is impossible. but “a exists” (where a is a term satisfying φx) is a mere noise or shape. therefore Socrates is numerous. Propositional Functions are either derived from this. like “men.” it is incorrect. when a person calls of whom we know nothing except that he brings a letter of introduction from our friend so-and-so. among true propositions. namely.

and the failure to do so in the past has been a disgrace to philosophy. XV. .Chap. Propositional Functions is of the utmost importance.

” “That is a very indeﬁnite description.” We are therefore not departing from usage in our terminology.” A “description” may be of two sorts. what is the deﬁnition of .e. and that in fact I met Jones. though I lie. This becomes obvious when the statement is false.e. and in the next chapter we shall consider the word the in the plural. such expressions as “the father of x” or “the sine of x. even if there were no man at all. but it was not Jones”.CHAPTER XVI DESCRIPTIONS We dealt in the preceding chapter with the words all and some. if we know what it would be to be a unicorn or a sea-serpent.” I may say “I met a man. An indeﬁnite description is a phrase of the form “a so-and-so. in that case. “Who did you meet?” “I met a man. that my assertion is true. as I should do if when I say I met a | man I really mean that I met Jones. but to the philosophical mathematician it is a word of very great importance: like Browning’s Grammarian with the enclitic δ . Let us begin with the former. enters into my statement. Indeed the statement would remain signiﬁcant. “I met a unicorn” or “I met a sea-serpent” is a perfectly signiﬁcant assertion. We have already had occasion to mention “descriptive functions.” These are to be deﬁned by ﬁrst deﬁning “descriptions. Our question is: What do I really assert when I assert “I met a man”? Let us assume. deﬁnite and indeﬁnite (or ambiguous). It may be thought excessive to devote two chapters to one word. though it could not possibly be true. for the moment. But we may go further: not only Jones. but no actual man.” i. since then there is no more reason why Jones should be supposed to enter into the proposition than why anyone else should. I would give the doctrine of this word if I were “dead from the waist down” and not merely in a prison. even if he is a foreigner and has never heard of Jones. It is clear that what I assert is not “I met Jones. I do not contradict myself. i. It is clear also that the person to whom I am speaking can understand what I say. in this chapter we shall consider the word the in the singular.” and a deﬁnite description is a phrase of the form “the so-and-so” (in the singular).

” For want of the apparatus of propositional functions. Similarly. . for logic is concerned with the real world just as truly as zoology. exists in his own world. while the second involves a propositional function. I should maintain. hence they must have some kind of logical being. in fact. Jones. Thus it is only what we may call the concept that enters into the proposition. And they have not known what diﬀerences in grammatical form are important.” (It will be remembered that we adopted the convention of using “sometimes” as not implying more than once.” for example.) This proposition is obviously not of the form “I met x.g. since otherwise the propositions in which they occur would be meaningless. it seems to me. To say that unicorns have an existence in heraldry. What exists is a picture. “I met Jones” and “I met a man” would count traditionally as propositions of the same form. Misled by grammar. and becomes. since it is signiﬁcant (though false) to say “I met a unicorn. must no more admit a unicorn than zoology can. is to say something deliberately confusing. somewhere among the shades. and the thoughts that Untersuchungen zur Gegenstandstheorie und Psychologie. or in imagination.” Therefore.” and so on.” which accounts | for the existence of the proposition “I met a unicorn” in spite of the fact that there is no such thing as “a unicorn. Logic. They have regarded grammatical form as a surer guide in analysis than. Descriptions these fabulous monsters. is a most pitiful and paltry evasion. does not contain a constituent “a unicorn. made of ﬂesh and blood. There is only one world. something unreal which may be called “a unicorn.” it is clear that this proposition. we can make true propositions of which these are the subjects. many logicians have been driven to the conclusion that there are unreal objects. that we can speak about “the golden mountain. it is. or a description in words.” “the round square. is a very important one. In the case of “unicorn. In such theories. to maintain that Hamlet. just as truly as (say) Napoleon existed in the ordinary world.Chap.” though it does contain the concept “unicorn.” which confronts us at this point. e. but in actual fact they are of quite diﬀerent forms: the ﬁrst names an actual person. What exists in heraldry is not an animal.” The question of “unreality. namely. rightly analysed. in the world of Shakespeare’s imagination. for example. the “real” world: Shakespeare’s imagination is part of it. or in literature. the great majority of those logicians who have dealt with this question have dealt with it on mistaken lines. when made explicit: “The function ‘I met x and x is human’ is sometimes true. there is only the concept: there is not also. or else confused to a degree which is scarcely credible. It is argued. by Meinong. moving and breathing of its own initiative. XVI. . there is a failure of that feeling for reality which ought to be preserved even in the most abstract studies. though with its more abstract and general features.

e. Thus if we falsely attribute meaning to these two words. deﬁnite or indeﬁnite. When you have taken account of all the feelings roused by Napoleon in writers and readers of history. it is in any case not a constituent of the proposition in which it occurs. like “a unicorn” just now.” and with the problem how there can be such a thing in a world where there are no unicorns. A robust sense of reality is very necessary in framing a correct analysis of propositions about unicorns.” But the two words “a unicorn” do not form a subordinate group having a meaning of its own. i. namely. and whoever juggles with it by pretending that Hamlet has another kind of reality is doing a disservice to thought. if we take “a man” . The sense of reality is vital in logic. Descriptions he had in writing Hamlet are real. etc.. and the word “unicorn” by itself is signiﬁcant.g. sitions which contain ambiguous descriptions. So are the thoughts that we have in reading the play. but is always signiﬁcant and sometimes true. an objective Hamlet. round squares. but in the case of Hamlet you have come to the end of him. “x is unreal” or “x does not exist” is not nonsense. in the analysis of propositions. those objects x for which the propositional function φx is true. feelings. we are dealing in the ﬁrst instance with symbols. It is not an indeﬁnite description which describes something unreal. If no one thought about Hamlet. in Shakespeare and his readers are real. in dealing with propositions. we ﬁnd ourselves saddled with “a unicorn. it is not a subordinate group having a meaning of its own. he would have soon seen to it that some one did. if no one had thought about Napoleon. In the proposition “I met a unicorn. But it is of the very essence of ﬁction that only the thoughts. (E.Chap. as objects described. there would be nothing | left of him. Such a proposition as “x is unreal” only has meaning when “x” is a description.” the whole four words together make a signiﬁcant proposition. “A unicorn” is an indeﬁnite description which describes nothing. when “x” is a description. But. if there is nothing unreal. in addition to them. how. and if we attribute signiﬁcance to groups of symbols which have no signiﬁcance. in that case the proposition will be true if “x” is a description which describes nothing. after all. Suppose we wish to make some statement about “a so-and-so. in the only sense in which this is possible. nothing “unreal” is to be admitted. you have not touched the actual man. and other such pseudo-objects. golden mountains.” where “so-and-so’s” are those objects that have a certain property φ. we shall insist that. | We may now proceed to deﬁne generally the meaning of propo. All this results from the fact that. But whether the description “x” describes something or describes nothing. could we admit anything unreal? The reply is that. XVI. and that there is not. in just the same sense as the word “man. we shall fall into the error of admitting unrealities. it may be asked. In obedience to the feeling of reality.

It is characteristic of ambiguous (as opposed to deﬁnite) descriptions that there may be any number of true propositions of the above form—Socrates is a . as applied to ambiguous descriptions.g. when there is at least one true proposition of the form “x is a so-and-so. however.” ψx will be “I met x. The important point is that. this is the same proposition as might be expressed by “some φ’s are ψ’s”. we wish to assert that “a so-and-so” has that property which x has when ψx is true.” φx will be “x is human. when rightly analysed.” So far as logic goes. The identity in “Socrates is a man” is identity between an object named (accepting “Socrates” as a name. (E. We say that “men exist” or “a man exists” if the | propositional function “x is human” is sometimes true. Descriptions as our instance of “a so-and-so.” If it were.Chap. “a so-and-so” would have to be identical with x for a suitable x. We may put this in other language.e.” i. The deﬁnition of existence. The proposition “Socrates is a man” is no doubt equivalent to “Socrates is human.”) Let us now wish to assert the property ψ of “a so-and-so. and in the other case of plurality. it is certainly not true in such a case as “a unicorn. “unreal. and although (in a sense) this may be true in some cases.”) Now the proposition that “a so-and-so” has the property ψ is not a proposition of the form “ψx.” It is just this fact. that the statement that a so-and-so has the property ψ is not of the form ψx. The is of “Socrates is human” expresses the relation of subject and predicate. in a certain clearly deﬁnable sense. in the case of “I met a man. and generally “a so-and-so” exists if “x is so-and-so” is sometimes true.” where “x” is a name.e. propositions verbally about “a so-and-so” are found to contain no constituent represented by this phrase.” but it is not the very same proposition. i. but rhetorically there is a diﬀerence. And that is why such propositions can be signiﬁcant even when there is no such thing as a so-and-so. is not the important point. the is of “Socrates is a man” expresses identity. This. which makes it possible for “a so-and-so” to be. It is a disgrace to the human race that it has chosen to employ the same word “is” for these two entirely diﬀerent ideas—a disgrace which a symbolic logical language of course remedies. subject to qualiﬁcations explained later) and an object ambiguously described.” The deﬁnition is as follows:— The statement that “an object having the property φ has the property ψ” means: “The joint assertion of φx and ψx is not always false. XVI. results from what was said at the end of the preceding chapter. An object ambiguously described will “exist” when at least one such proposition is true. because in the one case there is a suggestion of singularity.

when we have enumerated all the men in the world. namely. Plato is a man. though it has parts (namely.” It is of course quite clear that whatever there is in the world is deﬁnite: if it is a man it is one deﬁnite man and not any other. though at ﬁrst sight less obvious. because. but we cannot infer that “a man” means the same as “Socrates” means and also the same as “Plato” means and also the same as “Aristotle” means. namely. Thus “a man exists” follows from Socrates. “the author of Waverley” is not a simple symbol.Chap. “Scott. If.” this is fairly obvious: no one could suppose that “a man” was a deﬁnite object. The distinction between a name and all other symbols may be explained as follows:— A name is a simple symbol whose meaning is something that can only occur as subject. as opposed to speciﬁc men. With deﬁnite descriptions. Nevertheless.” and a description.” We have here a name. but it is the ‘a man. separate letters). but rather more complicated. on the other hand. by a consideration of the diﬀerence between a name and a deﬁnite description. Take the proposition. and not only so. not a deﬁnition of the phrase itself in isolation. “x is the soand-so” (where “x” is a name). XVI.. “This is a man. “the author of Waverley. In the case of “a so-and-so. Descriptions man. we shall have to content ourselves with what may be . One very important point about the deﬁnition of “a so-and-so” applies equally to “the so-and-so”. because the separate words that compose the phrase are parts which are symbols. the deﬁnition of the word the (in the singular). “Scott is the author of Waverley. And accordingly it is natural that we do not deﬁne “a man” itself. can only be true for one value of x at most. but only the propositions in which it occurs. which could be deﬁned by itself.’ the quintessential entity that is just an indeﬁnite man without being anybody in particular.e. or anyone else. the deﬁnition to be sought is a deﬁnition of propositions in which this phrase occurs.” which are asserted to apply to the same person. i. there is nothing left of which we can say. these parts are not symbols. On the other hand. | Socrates is a man. etc. as may be the case.” And a “simple” symbol is one which has no parts that are symbols. We may demonstrate that this must be the case. since these three names have diﬀerent meanings. Plato is a man. we deﬁned as an “individual” or a “particular. This brings us to the subject of deﬁnite descriptions. in Chapter XIII. whatever seems to be an “individual” is really capable of further analysis. Aristotle is a man. We come now to the main subject of the present chapter. Thus there cannot be such an entity as “a man” to be found in the world. the corresponding form of proposition. In the case of “the so-and-so” this is equally true. which are to be deﬁned in a way analogous to that employed for ambiguous descriptions. Thus “Scott” is a simple symbol. something of the kind that. or Plato.

and having this meaning in its own right.” which will be terms that. “Scott is the author of Waverley” is obviously a diﬀerent proposition from “Scott is Scott”: the ﬁrst is a fact in literary history. therefore. are never analysed and never occur | otherwise than as subjects. two things to compare: () a name. Descriptions called “relative individuals.” in which two names are said to apply to the same person. if “Scott is Sir Walter” really means “the person named ‘Scott’ is the person named ‘Sir Walter.” From the standpoint of our present problem. is being described as the person having that name. And if we put anyone other than Scott in place of “the author of Waverley.” whereas we have to compare such couples as “Scott” and “the author of Waverley. treat names as capable of being absolute. throughout the context in question. When a name is used directly. and not merely into the language used in making the assertion. But. or of the falsehood if our assertion happens to be false: it is merely part of the symbolism by which we express our thought. merely to indicate what we are speaking about. when we make a proposition about “the person called ‘Scott. it may be said. instead of being named.Chap. XVI. but of which they are no part. nothing that we shall have to say will depend upon this assumption. since it concerns diﬀerent stages in the hierarchy of “types. as a rule. Our proposition will now . but the wording may be a little shortened by it.’” then the names are being used as descriptions: i.” which both apply to the same object. And in that case we shall have correspondingly to content ourselves with “relative names. independently of the meanings of all other words. namely. may be ignored.” our proposition would become false. it is no part of the fact asserted. be nothing in the phraseology to show whether they are being used in this way or as names. The reply is that. it is something for which the actual words are a vehicle. and there will. the individual. A proposition containing a description is not identical with what that proposition becomes when a name is substituted. whether these are absolute names or only relative names. What we want to express is something which might (for example) be translated into a foreign language. whose meanings are already ﬁxed.’” the actual name “Scott” enters into what we are asserting. for the moment. which is a simple symbol. We may. and from which results whatever is to be taken as the “meaning” of the description. then. this problem. which consists of several words. This is a way in which names are frequently used | in practice. the second a trivial truism. and would therefore certainly no longer be the same proposition. even if the name names the same object as the description describes. directly designating an individual which is its meaning. On the other hand. the deﬁnition of descriptions. our proposition is essentially of the same form as (say) “Scott is Sir Walter. and do not raise the problem of types. We have.e. () a description.

Suppose. that φx is “always true”.” This completes the proof that “Scott is the author of Waverley” is not the same proposition as results from substituting a name for “the author of Waverley. propositional functions which are “always true” may become false.” x = x. that the author of Waverley is the author of Waverley. Thus so long as names are used as names.” assuming that φ is a function which has individuals for its arguments. the process of applying general statements about φx to particular cases will consist in substituting a name for the letter “x. the “law of identity.” because inhabiting London is an attribute which is not unique. and speak of a propositional function. Then we may substitute for “x” any name we choose. “x = x” is not the same proposition as “the author of Waverley is the author of Waverley. or that the round square is the round square. let it be. but we can speak about “the present King of England.” and “Aristotle” are names (a very rash assumption).” “Plato. The only thing that distinguishes “the so-andso” from “a so-and-so” is the implication of uniqueness. When we use a variable. XVI. and we shall obtain a true proposition. for example. When we substitute a description for a name. if the description describes nothing. without further premisses. There is no mystery in this as soon as we realise (what was proved in the preceding paragraph) that when we substitute a description the result is not a value of the propositional function in question. that the author of Waverley is the author of Waverley. the proposition we obtain is a diﬀerent one. This results | from what we have just proved. say. we can infer from the law of identity that Socrates is Socrates. if we substitute a name for “the author of Waverley” in a proposition. We cannot speak about “the present King of France. That is to say. propositions of the form “the so-and-so is the so-and-so” are not always true: it is necessary that the so-and-so should exist (a term which will be explained shortly).Chap.” Thus propositions about “the so-and-so” . Assuming for the moment that “Socrates. “Scott is Sir Walter” is the same trivial proposition as “Scott is Scott.” no matter what name “x” may be. But we shall commit a fallacy if we attempt to infer. and Aristotle is Aristotle.” no matter what name may be substituted.’” But so long as we are using names as names. We cannot speak of “the inhabitant of London. φx say. without more ado. Descriptions be a diﬀerent one if we substitute “the person called ‘Sir Walter. applying the result to our present case: If “x” is a name. Plato is Plato. that. It is false that the present King of France is the present King of France.” because there is none. We are now in a position to deﬁne propositions in which a deﬁnite description occurs. whether we say “Scott” or whether we say “Sir Walter” is as irrelevant to what we are asserting as whether we speak English or French. Thus from the fact that all propositions of the form “x = x” are true we cannot infer. In fact.

or if several people had written it. () “if x and y wrote Waverley.” We may somewhat simplify these three propositions. XVI.” We may say that “the author of Waverley” means “the value of x for which ‘x wrote | Waverley’ is true. () “if x wrote Waverley.” with the addendum that there is not more than one so-and-so. x was Scotch” is always true.’” . two functions of x.’” (Two propositions are “equivalent” when both are true or both are false. Our deﬁnition is as follows:— “The term satisfying the function φx exists” means: “There is a term c such that φx is always equivalent to ‘x is c. “The author of Waverley” is “the term satisfying the function ‘x wrote Waverley. that it is true for at least one value of c.) These two conditions together are deﬁned as giving the meaning of “the author of Waverley exists. the three together (but no two of them) imply that the author of Waverley was Scotch. The ﬁrst and second together are equivalent to: “There is a term c such that ‘x wrote Waverley’ is true when x is c and is false when x is not c. Hence the three together may be taken as deﬁning what is meant by the proposition “the author of Waverley was Scotch.” We may now deﬁne “the term satisfying the function φx exists. All these three are implied by “the author of Waverley was Scotch.” This is the general form of which the above is a particular case. Such a proposition as “Scott is the author of Waverley” could not be true if Waverley had never been written. translated into ordinary language.” Thus the proposition “the author of Waverley was Scotch. Descriptions always imply the corresponding propositions about “a so-and-so. () at most one person wrote Waverley.” In other words. and no more could any other proposition resulting from a propositional function φx by the substitution of “the author of Waverley” for “x. state: () at least one person wrote Waverley.” i. () whoever wrote Waverley was Scotch. x and y are identical” is always true. “x wrote Waverley” and “x is c.” for example. to begin with.Chap. “There is a term c such that ‘x wrote Waverley’ is always equivalent to ‘x is c. that which deﬁnes the property that makes a thing a so-and-so.” Conversely. namely.e. we then proceed to assert that the resulting function of c is “sometimes true.) We have here. (It obviously cannot be true for more than one value of c. These three propositions.” and we form a function of c by considering the equivalence of these two functions of x for all values of x.’” And “the so-and-so” will | always involve reference to some propositional function. involves: () “x wrote Waverley” is not always false.

of which the meaning is derived from that of its constituent symbols.e. if “a” is a name. It is only of descriptions | —deﬁnite or indeﬁnite—that existence can be signiﬁcantly asserted. A description has a “primary” . i. and therefore. in the hope that ultimately they will suﬃce to demonstrate that it was A who did the deed. is a symbol devoid of meaning. without knowing any proposition of the form “x is the so-and-so. whereas a description. we are using the word “Homer” as an abbreviated description: we may replace it by (say) “the author of the Iliad and the Odyssey. The abstract distinction is as follows. When descriptions occur in propositions. Thus “the author of Waverley was Scotch” is: “There is a term c such that () ‘x wrote Waverley’ is always equivalent to ‘x is c.” does not become incapable of occurring signiﬁcantly merely on the ground that it describes nothing. The proposition “the so-and-so exists” is signiﬁcant. the reason being that it is a complex symbol. We may inquire signiﬁcantly whether Homer existed. We may even go so far as to say that.” And generally: “the term satisfying φx satisﬁes ψx” is deﬁned as meaning: “There is a term c such that () φx is always equivalent to ‘x is c. but if a is the so-and-so (where “a” is a name). to know many propositions concerning “the so-and-so. which we could not do if “Homer” were a name. namely.” where “x” is a name. whether true or false.Chap.” we have still to take account of the third of our three propositions. occur. if intended to be a name. for. the words “a exists” are meaningless. it is necessary to distinguish what may be called “primary” and “secondary” occurrences.” This will be satisﬁed by merely adding that the c in question is to be Scotch.’ () c is Scotch. when we ask whether Homer existed.” This is the deﬁnition of propositions in which descriptions occur. but what seem like names are really descriptions. i. It is possible to have much knowledge concerning a term described. In a detective story propositions about “the man who did the deed” are accumulated.e.” The same considerations apply to almost all uses of what look like proper names.” without actually knowing what the so-and-so is. in the strict sense. Descriptions In order to deﬁne “the author of Waverley was Scotch. like “the present King of France. it must name something: what does not name anything is not a name. in all such knowledge as can be expressed in words—with the exception of “this” and “that” and a few other words of which the meaning varies on diﬀerent occasions—no names. “Whoever wrote Waverley was Scotch. And so. XVI.’ () ψc is true.

Every proposition in which a description which describes nothing has a primary occurrence is false.e.” This is ambiguous. and have therefore been omitted in the above account. and the proposition is false. But for purposes of mathematics.” or “the R of y” as we may say. i.” then substitute “the present King of France” for “x. An instance will make this clearer. . is to say that the following propositional function of c: “c is rich. To say “the father of y is rich. a description has a “secondary” occurrence when the result of substituting the description for x in φx gives only part of the proposition concerned.Chap.” and then deny the result. “the term having the relation R to y. is true for at least one value of c. tive functions. is of the utmost importance both in logic and in theory of knowledge. but if we are to take “x is not bald” and substitute “the present King of France” for “x.” for example. Consider “the present King of France is bald.’” is “sometimes true. Descriptions occurrence when the proposition in which it occurs results from substituting the description for “x” in some propositional function φx. | Descriptions occur in mathematics chieﬂy in the form of descrip. which has conﬁned itself to the barest mathematical requisites. The theory of descriptions. It obviously cannot be true for more than one value. XVI.” Here “the present King of France” has a primary occurrence.” then “the present King of France” has a primary occurrence and the proposition is false.e. brieﬂy outlined in the present chapter. and ‘x begat y’ is always equivalent to ‘x is c. the more philosophical parts of the theory are not essential. But now consider “the present King of France is not bald. on the analogy of “the father of y” and similar phrases.” i. Confusion of primary and secondary occurrences is a ready source of fallacies where descriptions are concerned. the occurrence of “the present King of France” is secondary and our proposition is true. If we are ﬁrst to take “x is bald.

Hitherto we have. in fact. a deﬁnition which will assign a meaning to propositions in whose verbal or symbolic expression words or symbols apparently representing classes occur. and so on.. they are deﬁned in a way closely analogous to that used for descriptions. treated “class” as a primitive idea. Some further subtlety appears to be required.e. as we should say but for the vicious circle. when the number is deﬁned as the class of all unit classes. in dealing with arithmetic. We shall then be able to say that the symbols for classes are mere conveniences. if there is a term c such that x will be a member of α when x is c but not otherwise. But.e. We must seek a deﬁnition on the same lines as the deﬁnition of descriptions. In other words.e. that the number is to be deﬁned as the class of all unit classes. for the reasons set forth in Chapter XIII. if for no others. not representing objects called “classes. and there are reasons (which we shall give in outline) for regarding the deﬁnition of classes that will be suggested as not ﬁnally satisfactory. like descriptions.” The theory of classes is less complete than the theory of descriptions. we cannot accept “class” as a primitive idea. that a cardinal number is to be deﬁned as a class of classes. in more ordinary language. but which will assign a meaning that altogether eliminates all mention of classes from a right analysis | of such propositions. .. namely: A class α is said to be a “unit” class if the propositional function “‘x is an α’ is always equivalent to ‘x is c’” (regarded as a function of c) is not always false. Of course. the sons of rich men. but the reasons for regarding the deﬁnition which will be oﬀered as being approximately correct and on the right lines are overwhelming. and in Chapter III. “unit classes” must be deﬁned so as not to assume that we know what is meant by “one”. i. We saw in Chapter II. This gives us a deﬁnition of a unit class if we already know what a class is in general.” and that classes are in fact. i. of all that have just one member. we shall be concerned with classes.CHAPTER XVII CLASSES In the present chapter we shall be concerned with the in the plural: the inhabitants of London. or (as we say) “incomplete symbols. i. logical ﬁctions.

On the other hand. as we .). we shall have no means of expressing the proposition which we expressed by means of “a” as opposed to the one that | we expressed by means of the description. Thus descrip. We might try to avoid this conclusion by the use of descriptions. but one consequence which it implies may be used to elucidate its meaning. We shall come much nearer to a satisfactory theory if we try to identify classes with propositional functions. that there are such entities as “heaps. We have already seen that classes cannot be regarded as a species of individuals. If we were to attempt to do that. It is diﬃcult to explain precisely what one means by this statement. I am not called upon to have an opinion on this point. I do not mean to assert. Our ﬁrst business is to give the reasons for this opinion.” As a mathematical logician. which has no members at all and cannot be regarded as a “heap”. We cannot take classes in the pure extensional way as simply heaps or conglomerations. we should ﬁnd it impossible to understand how there can be such a class as the null-class. we might use it as (in one perfectly legitimate sense) a deﬁnition of that particular. XVII. tions would not enable a perfect language to dispense with names for all particulars. the undeﬁned symbols in this language would represent symbolically what I mean by “the ultimate furniture of the world. or to deny.Chap. we are maintaining. and because we can prove that the number of classes is greater than the number of individuals.” This is a description of some particular. we cannot identify them with the classes composed of their constituents. If we had a complete symbolic language. and need not be represented by undeﬁned symbols.” I am maintaining that no symbols either for “class” in general or for particular classes would be included in this apparatus of undeﬁned symbols.” or some other name for the same particular. on account of the contradiction about classes which are not members of themselves (explained in Chapter XIII.” If our language does not contain the name “a. we should also ﬁnd it very hard to understand how it comes about that a class which has only one member is not identical with that one member. All that I am maintaining is that. Classes The ﬁrst thing is to realise why classes cannot be regarded as part of the ultimate furniture of the world. if there are such things as heaps. and an undeﬁned symbol for everything indeﬁnable. with a deﬁnition for everything deﬁnable. all the particular things there are in the world would have to have names which would be included among undeﬁned symbols. Every class. a proposition in which “a” occurs is not (as we saw in the preceding chapter) identical with what this proposition becomes when for “a” we substitute “the last thing Cæsar saw before he died. classes diﬀer from particulars. In this respect. Take (say) “the last thing Cæsar saw before he died. But if “a” is a name for the same particular.

ψx are “formally equivalent” when φx is always equivalent to ψx. it becomes very diﬃcult to see what they can be. . we must not be supposed to be asserting dogmatically that there are none. It is the fact that there are other functions formally equivalent to a given function that makes it impossible to identify a class with a function. namely. and therefore two formally equivalent functions will have to determine the same class. we can imagine Socrates replaced by Plato or Aristotle or a gorilla or the man in the moon or any other individual in the world. When we have decided that classes cannot be things of the same sort as their members. it can equally well be deﬁned by any other which is true whenever the ﬁrst is true and false whenever the ﬁrst is false. and that every propositional function determines an appropriate class. “je n’ai pas besoin de cette hypoth`se. Classes explained in Chapter II. Given any proposition (true or false).” All that | we are observing at present is that a class is rendered determinate by a propositional function. And if we can ﬁnd any way of dealing with them as symbolic ﬁctions. that they cannot be just heaps or aggregates. We are merely agnostic as regards them: like Laplace. there are always many others which are true when it is true and false when it is false. “entities are not to be multiplied without necessity. The class determined will consist of all those substitutions that give a true one. etc.” But when we refuse to assert that there are classes. some of these substitutions will give a true proposition and some a false one.” e Let us set forth the conditions that a symbol must fulﬁl if it is to serve as a class. This is an example of Occam’s razor. In general. But if a class can be deﬁned by one propositional function.Chap. Of course. since we avoid the need of assuming that there are classes without being compelled to make the opposite assumption that there are no classes.. we have still to decide what we mean by “all those which. say about Socrates. Two propositions are | “equivalent” when both are true or both false. and also that they cannot be identiﬁed with propositional functions. We say that two propositional functions are “formally equivalent” when this happens. we increase the logical security of our position. we can say. for we wish classes to be such that no two distinct classes have exactly the same members. consisting of those arguments for which the function is true. We merely abstain from both assumptions. For this reason the class cannot be identiﬁed with any one such propositional function rather than with any other—and given a propositional function. is deﬁned by some propositional function which is true of the members of the class and false of other things. XVII. two propositional functions φx. if they are to be more than symbolic ﬁctions. I think the following conditions will be found necessary and suﬃcient:— () Every propositional function must determine a class.

namely. a class is determined by its membership. This results from the contradiction which we discussed in Chapter XIII. since I may believe falsely that the Phœnix is an immortal rational animal. () Lastly—and this is the condition which is most diﬃcult of fulﬁlment—it must be possible to make propositions about all the classes that are composed of individuals. Now. “The combinations of n things m at a time” represents a class of classes.) () We must ﬁnd some way of deﬁning not only classes. but classes of classes. Classes () Two formally equivalent propositional functions must determine the same class. “I believe that all men are mortal” may be true. () It must under all circumstances be meaningless (not false) to suppose a class a member of itself or not a member of itself.g. to begin with. mathematical logic would break down. we say that a is a “member” of the class if φa is true. e. for each group of formally equivalent propositional functions. when we say that two formally equivalent propositional functions may be not identical. The reason there is a diﬃculty about this condition is that it can be proved to be impossible to speak of all the propositional functions that can have arguments of a given type.Chap. that cardinal numbers are to be deﬁned as classes of classes. although they deﬁne the same class. Without some symbolic method of dealing with classes of classes. we need to be able to say that a member of the posterity belongs to all hereditary classes to which the given term belongs. whereas others . They state that there is to be one class. We saw in Chapter II. and two which are not formally equivalent must determine diﬀerent classes. the class of men is to be the same as that of featherless bipeds or rational animals or Yahoos or whatever other characteristic may be preferred for deﬁning a human being. Thus we are led to consider statements about functions. ignore this last condition and the problems which it raises. e. or about all the classes that are composed of objects of any one logical “type. The ordinary phrase of elementary mathematics.” If this were not the case. XVII.g. no more and no less. we may prove the truth of the assertion by pointing out that a statement may be true of the one function and false of the other. or (more correctly) functions of functions. That is. Some of the things that may be said about a function may be regarded as said about the class deﬁned by the function. while “I believe that all rational animals are mortal” may be false. the class of all classes of m terms that can be selected out of a given class of n terms. mathematical induction. (If a class is determined by a function φx. The ﬁrst two conditions may be | taken together. many uses of classes would go astray—for example. We will. and this requires the sort of totality that is in question. In deﬁning the posterity of a given term. and no two diﬀerent classes can have the same membership.

Chap. Given a function of a function which may or may not be extensional. the condition. XVII. for example. but when the original function is intensional.” and “φx is sometimes true” is equivalent to “α has members” or (better) “α has at least one member. the new one is more often true than the old one.” Thus extensional functions of a function φx may. a class which is a member of . then consider the assertion “there is a function having the property f and formally equivalent to φx. “φx is always true” is equivalent to “everything is a member of α. It is equivalent to the assertion that the class deﬁned by the function φx is a unit class. again. a class having one member. it is true when our original statement is true.” The condition is that there is a term c such that φx is always equivalent to “x is c.” regarded as a function of “x is human. Classes cannot. in other words. The statement “all men are mortal” involves the functions “x is human” and “x is mortal”.e. consider again “I believe that all men are mortal. because its truth-value may be changed by the substitution of a formally equivalent function (which leaves the class unchanged). while intensional functions cannot be so regarded. or.” This is an extensional function of φx. i.” The derived extensional function is: “There is a func- . the two fundamental functions of functions are: “φx is always true” and “φx is sometimes true. and it is formally equivalent to the original function of φx if this original function is extensional. It is to be observed that all the speciﬁc functions of functions that we have occasion to introduce in mathematical logic are extensional. if it is like “all men are mortal.” Take.” i. Thus. we can always derive from it a connected and certainly extensional function of the same function. and when a function of a function is not extensional. by the following plan: Let our original function of a function be one which attributes to φx the property f . We can interpret the statement in either way. dealt with in the preceding chapter.” so that “I believe that all men are mortal” is an intensional function of “x is human” or “x is mortal. for practical | purposes. But.” This is obviously extensional. because its truth-value is unchanged if we substitute for “x is human” or for “x is mortal” any formally equivalent function. we can say that it involves the classes men and mortals.e. for the existence of “the term satisfying φx. if we choose. if its truth-value is unchanged by the substitution of any formally equivalent function. be regarded as functions of the class determined by φx. In the language of classes. the statement “I believe that all men are mortal” cannot be regarded as being about the class determined by either function. We will call a statement involving a function φx an “extensional” function of the function φx. For example. if α is the class determined by φx.” Each of these has its truth-value unchanged if any formally equivalent function is substituted for φx. as we have just seen. we will call it “intensional.

Take such an every-day statement as “a is a typical Frenchman. is somewhat technical. these we will call “predicative a-functions. also ∗ . We give the name of “derived extensional function” to the function constructed as above. See Principia Mathematica. and as asserting f of this class. We saw in Chapter XIII. I. namely. This may be taken as the deﬁnition of a proposition about a class. But this condition—at once the most important and the most diﬃcult—is not fulﬁlled in virtue of anything we have said as yet.. and must be brieﬂy discussed. .” even if I believe falsely that the Phœnix is rational and immortal. to the function: “There is a function having the property f and formally equivalent to φx. it is explained in Principia Mathematica. we might regard our task as completed. XVII. The diﬃculty is connected with the theory of types. chap. Classes tion formally equivalent to ‘x is human’ and such that I believe that whatever satisﬁes it is mortal. vol. Introduction. that there is a hierarchy of logical types. namely. but may be taken for granted here. Let us call them all afunctions. It results that. pp. and that it is a fallacy to allow an object belonging to one of these to be substituted for an object belonging to another. and it will be found that technically it yields the results which are required in order to make a theory symbolically satisfactory. This gives a meaning to any statement about a class which can be made signiﬁcantly about a function. We may take ﬁrst those among them which do not involve reference to any collection of functions. we may deﬁne: To assert that “the class determined by the function φx has the property f ” is to assert that φx satisﬁes the extensional function derived from f.” We may regard the derived extensional function as having for its argument the class determined by the function φx. and the impossibility of a class being or not being a member of itself. ii.” But unless The reader who desires a fuller discussion should consult Principia Mathematica.” This remains true when we substitute “x is a rational animal” | for “x is human.” If we now proceed to functions involving reference to the totality of predicative a-functions. we shall incur a fallacy if we regard these as of the same type as the predicative a-functions. What we have said just now as regards the deﬁnition of classes is suﬃcient to satisfy our ﬁrst four conditions. | Now it is not diﬃcult to show that the various functions which can take a given object a as argument are not all of one type.Chap. – and ∗ .” where the original function was “the function φx has the property f. i. but for our ﬁfth condition. The way in which it secures the third and fourth.e. the possibility of classes of classes.” How shall we deﬁne a “typical Frenchman”? We may deﬁne him as one “possessing all qualities that are possessed by most Frenchmen.

Then ψ appears as a . thus we escape the vicious circle. i. if I say “Napoleon had all the qualities that make a great general. construct a variable which would run through the ﬁrst n of these types. where n is ﬁnite. the type is not rendered determinate by that of the argument. This is not a logical contradiction. but what has been said may suﬃce to make it clear that the functions which can take a given argument are of an inﬁnite series of types. It would require a much fuller discussion to set forth this point fully. Whenever. if it were. This is fairly obvious.” It is necessary to decide upon the type of our function. that mere fact would at once generate a new type of function with the same arguments. XVII. we may suppose that “qualities” is to mean “predicative functions. since there is no reason why there should be any typical Frenchmen. and is the principle which leads to the theory of types by which vicious-circle paradoxes are avoided. Any decision will do. the functions in question must be limited to one type if a vicious circle is to be avoided. and so on. and.” I must deﬁne “qualities” in such a way that it will not include what I am now saying. but some decision is unavoidable. and therefore the deﬁnition shows that to be not typical is essential to a typical Frenchman.. and. Classes we conﬁne “all qualities” to such as do not involve a reference to any totality of qualities. and we should be involved in a vicious circle.” I mean | “Napoleon satisﬁed all the predicative functions. “having all the qualities that make a great general” must not be itself a quality in the sense supposed. the totality of values over which the variable could range would only be deﬁnable in terms of itself. We call predicative a-functions the ﬁrst type of a-functions. this new object must not be among the values which our previous variable could take. As applied to a-functions. we generate a new object. We there spoke of “a function formally equivalent to φx. by statements about “all” or “some” of the values that a variable can signiﬁcantly take. etc. etc. and would set the whole process going again. But wherever “all functions which” occurs. but not a predicative property.” Then when I say “Napoleon had all the qualities. No variable a-function can run through all these diﬀerent types: it must stop short at some deﬁnite one.e. afunctions involving reference to the totality of the ﬁrst type we call the second type. if we could. For example. we shall have to observe that most Frenchmen are not typical in the above sense. but it illustrates the need for separating oﬀ qualities that involve reference to a totality of qualities from those that do not. since.” This statement attributes a property to Napoleon. These considerations are relevant to our deﬁnition of the derived extensional function.Chap. as Napoleon and the typical Frenchman have shown. Let us call the supposed formally equivalent function ψ. We could. by various technical devices. but we cannot construct a variable which will run through them all.

like the multiplicative axiom and the axiom of inﬁnity. It is therefore worth while to ask whether there is any reason to suppose it true. as we have just seen. We cannot use them as hypotheses. The theory of deduction. not necessary. Statements about all a-classes (i. and deduce hypothetical consequences. like our two previous mathematical axioms. we should not merely not obtain the same results. does not determine its type. then any extensional function which holds of all a-functions of the nth type will hold of any a-function whatever. we can also deduce the consequences of supposing it false. On the other hand. for they are rules of deduction as well as premisses. is necessary for certain results. as explained in Chapter XIV. say the nth .” One particular region where this is vital is mathematical induction. but not for the bare existence of deductive reasoning. But this. it is impossible as yet to say whether there may . that is to say. and the laws for propositions involving “all” and “some. such that any a-function is formally | equivalent to some a-function of the nth type.” and may be stated as follows:— “There is a type (τ say) of a-functions such that. we use functions of this type in deﬁning our associated extensional function.” If this axiom is assumed. or else what we deduce according to them does not even follow from the premisses. this gives us in practice results which would otherwise have required the impossible notion of “all a-functions. or something like them.e. We can deduce | its consequences hypothetically. It is therefore only convenient. So long as only extensional functions of functions are involved. It is chieﬂy as a technical means of embodying an assumption leading to this result that classes are useful. If we are to be able (as our ﬁfth requisite demands) to deal with all classes whose members are of the same type as a. They must be absolutely true.. but we should not obtain any results at all. The axiom of reducibility involves all that is really essential in the theory of classes.” are of the very texture of mathematical reasoning: without them. we must be able to deﬁne all such classes by means of functions of some one type. it is formally equivalent to some function of the type in question. XVII. And in view of the complication of the theory of types. there must be some type of a-function.Chap. The assumption is called the “axiom of reducibility. and must be of some determinate type. given any afunction. could perfectly well be stated as an hypothesis whenever it is used. all classes deﬁned by a-functions) can be reduced to statements about all a-functions of the type τ. instead of being assumed to be actually true. If this is the case. Classes variable. and of the uncertainty of all except its most general principles. This axiom. All that we know necessarily about the type of φ is that it takes arguments of a given type—that it is (say) an a-function. the axiom of reducibility.

and pure mathematics (which is the same thing). But it is reasonable to regard the theory outlined in the present chapter as right in its main lines. There is need of further work on the theory of types. Classes not be some way of dispensing with the axiom of reducibility altogether. and various properties not to be reckoned as predicates. even if the axiom is empirically true. be two things which had exactly the same predicates. i. in Leibnizian phraseology. be sound in principle. assuming the correctness of the theory outlined above. which regarded all propositions as reducible to the subject-predicate form. in the narrow sense in which we have been using the word “predicate.e. Thus Leibniz’s assumption is a much stricter and narrower one than ours. in the hope of arriving at a doctrine of classes which does not require such a dubious assumption. an accident. in spite of our desire to exclude. The avoidance of classes as entities by this method must.” How does our axiom look when we pass beyond predicates in this narrow sense? In the actual world there seems no way of doubting its empirical truth as regards particulars. so far as I can see. not only in this higgledy-piggledy job-lot of a world in which chance has imprisoned us. according to his logic. I do not see any reason to believe that the axiom of reducibility is logically necessary. (Not.Chap. what can we say as to the truth or falsehood of the axiom? The axiom. as far as possible. Pure logic. However. which is what would be meant by saying that it is true in all possible worlds. aims at being true. as it were. There is a certain lordliness which the logician should preserve: he must not condescend to derive arguments from the things he sees about him. as a logical principle. we may observe. But this is. It is for this reason that the theory of classes cannot be regarded as being as complete as the theory of descriptions. Now predicates are only some among what we called “predicative functions. | Viewed from this strictly logical point of view. it would seem. however the detail may still require adjustment. in its reduction of propositions nominally about classes to propositions about their deﬁning functions. owing to spatio-temporal diﬀerentiation: no two particulars have exactly the same spatial and temporal relations to all other particulars. reduces itself to one . whatever seemed open to serious doubt.) But there is no good reason for believing his form. Leibniz assumed. of course. XVII. as a matter of abstract logical possibility. The admission of this axiom into a system of logic is therefore a defect. as above outlined. a fact about the world in which we happen to ﬁnd ourselves. is a generalised form of Leibniz’s identity of indiscernibles. It is because this seems indubitable that we have included the theory of classes. that two diﬀerent subjects must diﬀer as to predicates. in all possible worlds. There might quite well. The theory of classes.” which will include also relations to given terms.

The axiom is: There is a type τ such that if φ is a function which can take a given object a as argument. then there is a function ψ of the type τ which is formally equivalent to φ. formally equivalent to φ. and τ the type mentioned in the above axiom. . then to say that the class determined by φ has the property f is to say that there is a function of type τ.Chap. we will here repeat them. The deﬁnition is: If φ is a function which can take a given object a as argument. Classes axiom and one deﬁnition. and having the property f. For the sake of deﬁniteness. XVII.

we ﬁnd that there is no point at which a sharp line can be drawn. of course. and have then analysed the conceptions involved in the deﬁnition.CHAPTER XVIII MATHEMATICS AND LOGIC Mathematics and logic. so much of modern logic is symbolic and formal. The consequence is that it has now become wholly impossible to draw a line between the two. we may challenge them to indicate at what point. a matter of detail: starting with premisses which would be universally admitted to belong to logic. It will then be obvious that any answer must be quite arbitrary. are incapable of following a piece of symbolic reasoning. and arriving by deduction at results which as obviously belong to mathematics. having spent their time in the study of classical texts. historically speaking. Mathematics has been connected with science. In a synthetic. in the successive deﬁnitions and | deductions of Principia Mathematica. So much of modern mathematical work is obviously on the border-line of logic. In the earlier chapters of this book. they consider that logic ends and mathematics begins. we have ﬁrst deﬁned “cardinal number” and shown how to generalise the conception of number. though formally . the two are one. and the natural numbers are only reached after a long journey. Both types are now fortunately growing rarer. If there are still those who do not admit the identity of logic and mathematics. deductive treatment these fundamentals come ﬁrst. This view is resented by logicians who. in fact. have been entirely distinct studies. until we found ourselves dealing with the fundamentals of logic. Such treatment. The proof of their identity is. and by mathematicians who have learnt a technique without troubling to inquire into its meaning or justiﬁcation. They diﬀer as boy and man: logic is the youth of mathematics and mathematics is the manhood of logic. logic with Greek. that the very close relationship of logic and mathematics has become obvious to every instructed student. But both have developed in modern times: logic has become more mathematical and mathematics has become more logical. with logic to the left and mathematics to the right. starting from the natural numbers.

and similarity between classes. And in thus generalising we have. because the ultimate logical concepts and propositions with which it starts are remote and unfamiliar as compared with the natural numbers. for example: projective and descriptive geometry. in which traditional arithmetic is at once dissolved and enlarged. down to the point at which co-ordinates are introduced. We are thus brought face to face with the question: What is this subject. through the deﬁnition of cardinals. through the theory of induction and ancestral relations. On the other hand. since we thereby secure that a given process of deduction shall have more widely applicable results. of a certain kind of one-many relations. It is a principle. therefore.” i. but for the sake of argument we may replace it by the word “number. and incapable of being decided rationally. The ordinal properties of the various kinds of number-series. in all formal reasoning. The most | elementary properties of numbers are concerned with one-one relations. and the elements of the theory of continuity of functions and the limits of functions. in thus generalising the reasoning of arithmetic. beyond which is the still unknown. there are recognised branches of mathematics which have nothing to do with number—all geometry that does not use co-ordinates or measurement. created a set of new deductive systems. and the dominion of knowledge over them is not as yet very secure. XVIII. the theory of selections— is to be said to belong to logic or to arithmetic is entirely arbitrary. but whether any one of these new deductive systems—for example. which may be called indiﬀerently either mathematics or logic? Is there any way in which we can deﬁne it? . does not have to do with number. no one of which is specially concerned with numbers. we are. can be generalised so as no longer to involve any essential reference to numbers. is more diﬃcult for the reader. Multiplication is merged in the theory of “selections. and through the deﬁnitions of the arithmetical operations.” The statement that mathematics is the science of number would be untrue in two diﬀerent ways. Finitude is merged in the general study of ancestral relations. On the one hand.e. which yields the whole theory of mathematical induction. through the general theory of series. It used to be said that mathematics is the science of “quantity. to generalise to the utmost. Addition is concerned with the construction of mutually exclusive classes respectively similar to a set of classes which are not known to be mutually exclusive.” “Quantity” is a vague word. or even with quantity in the sense of greater and less. in eﬀect. it has become possible to generalise much that used to be proved only in connection with numbers. The result is that what was formerly the single study of Arithmetic has now become divided into a number of separate studies. merely following a precept which is universally admitted in mathematics. Mathematics and Logic more correct than that which we have adopted.Chap. Also they represent the present frontier of knowledge.

” We may now observe that it is intended to convey that this argument is valid in virtue of its form. then Socrates is mortal. as pure mathematicians or logicians. “the propositional function ‘if all α’s are β’s and x is an α. we should have had a non-formal argument. If our argument is one (say) which holds of all men. then x is a β’ is always true. only admissible because Socrates is in fact a man. instead of supposing x to be a man. but not that Socrates and Plato are two. not that premisses and conclusion are actually true. It is clear that. deal with particular things or particular properties: we deal formally with what can be said about any thing or any property. | we introduce something irrelevant and not formal. We are prepared to say that one and one are two. Thus we may substitute α for men. It is not open to us. because. then x is a β”. To begin with. If we had omitted “Socrates is a man” from our premisses. as above.Chap. β for mortals. we were to suppose him to be a monkey . in which no actual things or properties are mentioned. we have never heard of Socrates and Plato. We may make this clear by applying it to the case of the syllogism. because. we shall prove it concerning “x. Traditional logic says: “All men are mortal.” Here at last we have a proposition of logic—the one which is only suggested by the traditional statement about Socrates and men and mortals. to mention anything at all. even the most traditional logic points out that the actual truth of the premisses is irrelevant to logic. Socrates is a man. we do not. and x for Socrates. XVIII.” Now it is clear that what we mean to assert. Mathematics and Logic Certain characteristics of the subject are clear. A world in which there were no such individuals would still be a world in which one and one are two. if all α’s are β’s and x is an α. to begin with. Thus the ﬁrst change to be made in the above traditional syllogism is to state it in the form: “If all men are mortal and Socrates is a man. It would be ridiculous to go through a long argument about Socrates. we shall always arrive ultimately at statements like the above. in this subject. this will happen through the mere desire not to waste our time proving in a particular case what can be proved generally. nothing depends upon the terms that occur in it. We then arrive at the statement: “No matter what possible values x and α and β may have. if we do so. and then go through precisely the same argument again about Plato. the argument will retain its hypothetical validity even when x is not a man.” With | this hypothesis. But now we shall ﬁnd that our argument would still be valid if.” with the hypothesis “if x is a man. is only that the premisses imply the conclusion. the argument is formal. in other words. But when. in that case we could not have generalised the argument. if formal reasoning is what we are aiming at. where α and β are any classes whatever. in our capacity of logicians or pure mathematicians. and x is any individual. not in virtue of the particular terms occurring in it. therefore Socrates is mortal.

We can proceed to general assertions. Mathematics and Logic or a goose or a Prime Minister. XVIII. But the form is not itself a new constituent. This assertion will belong to logic (or mathematics) in the sense in which we are using the word.e.e. as a ﬁrst approximation. But in this assertion we do not mention any particular things or particular relations. Are we to infer that the general form itself is a constituent of such logical propositions? Given a proposition. those asserting relations between two terms. This is what we do when we use such a schema as “xRy.” where α is any class of individuals. also the fact that what appear to be their names are really truncated descriptions. Neither of these facts is relevant to the present issue. and that the constituents of the proposition (as well as of the corresponding fact) are simply the two terms and the relation. but no particular instance of it can occur. in fact. if it were. no particular things or relations can ever enter into a proposition of pure logic. We cannot embark upon this question now. but I propose to explain how the problem arises. Aristotle. such as “xRy is sometimes true”—i. We can. We are left with pure forms as the only possible constituents of logical propositions. (I ignore the fact that Socrates and Aristotle are not simple.” which may be read “x has the relation R to y. as we say. while keeping the form unchanged. Socrates. there are cases where dual relations hold. We shall therefore not waste our time taking as our premiss “x is a man” but shall take “x is an α. or “φx” where φ is any propositional function of some assigned type. Take (say) the proposition “Socrates was before Aristotle. we should need a new form to embrace both it and the other constituents. turn all the constituents of a proposition into variables. with conﬂicting considerations on the one side and on the other.” Here it seems obvious that we have a relation between two terms.Chap.g.) We may represent the general form of such propositions by “xRy.” At this point we ﬁnd ourselves faced with a problem which is easier to state than to solve. the form “xRy”—do actually enter into propositions of the kind we are considering. and before. the view that forms are what enter into logical propositions as their constituents.” This general form may occur in logical propositions.” which stands for any | one of a certain class of propositions. i. The problem is: “What are the constituents of a logical proposition?” I do not know the answer. The question of the analysis of such propositions is a diﬃcult one. namely. I do not wish to assert positively that pure forms—e. “purely formal.” we have certain constituents and also a certain form. such as “Socrates is before Aristotle. And we may explain (though not . Thus the absence of all mention of particular things or properties in logic or pure mathematics is a necessary result of the fact that this study is. but we may accept.

that they are to be such as can be obtained from a proposition containing no variables (i.” “is” and “than” merely indicate form. that remains unchanged when every constituent of the proposition is replaced by another. broadly speaking. can be indicated otherwise than by speciﬁc words: the order of the words can do most of what is wanted. we must have a word for it. But without even one we should ﬁnd ourselves in diﬃculties. we wish to speak about all propositions that have the form. a word for the form will usually be found not indispensable. in it. is not the point that is important for our present purpose. Take “Socrates is human. And another way of stating the same thing is to say that logic (or mathematics) is concerned only with forms. But this principle must not be pressed. as a rule. namely. Assuming—as I think we may—that the forms of propositions can be represented by the forms of the propositions in which they are expressed without any special words for forms. Thus “Socrates is earlier than Aristotle” has the same form as “Napoleon is greater than Wellington.Chap.e.” though every constituent of the two propositions is diﬀerent. a word or symbol expressing incompatibility. a.” Here “is” is not a constituent of the proposition. What is important for us is to observe that form may be the one concern of a general proposition. etc. or any variant of these forms. we should arrive at a . even when no word or symbol in that proposition designates the form. XVIII. are commonest in languages having fewest inﬂections. but merely indicates the subject-predicate form. that one word or symbol is enough for this purpose. what we call “truthfunctions”) without any word at all. the proposition is the same as “Socrates precedes Aristotle. the.) by turning every constituent into a variable and asserting that the result is always true or sometimes true. some. probably in theory it is never indispensable. or that it is always true in respect of some of the variables that the result is sometimes true in respect of the others. as in mathematics. no such words as all. as a necessary (though not suﬃcient) characteristic of logical or mathematical propositions.” in which these words have disappeared and the form is otherwise indicated. however. For example. Mathematics and Logic formally deﬁne) what we mean by the “form” of a proposition as follows:— The “form” of a proposition is that. There are in every language some words whose sole function is to indicate form. These words. We may thus lay down. but if. If we wish to speak about the form itself. Similarly in “Socrates is earlier than Aristotle. Form. it is diﬃcult to see how we could conveniently express molecular forms of propositions (i.e. and is concerned with them only in the way of stating that they are always or | sometimes true—with all the permutations of “always” and “sometimes” that may occur. This. We saw in Chapter XIV.

” i. by term-for-term substitution. in fact.| predicate form and the second expresses a three-term relation.” Logical constants may be deﬁned exactly as we deﬁned forms. And this which is in common is what we call “form. for example. not of the pure mathematician or logician. express dual relations.” In this sense all the “constants” that occur in pure mathematics are logical constants. If we are to have any words in our pure logical language. there are words that express form. after all. We should have symbols for variables. But. would be such a language. in the above manner.e. and various diﬀerent . We cannot obtain from the above prototype by term-for-term substitution such propositions as “Socrates is human” or “the Athenians gave the hemlock to Socrates. such as “is” and “than. they must be such as express “logical constants.” “Wellington” for “Aristotle. The language of | mathematical logic. x is c. any one of which can result from any other by substitution of terms one for another. In such a language we could express all the propositions of mathematics even if we did not know one single word of the language. such a proposition can be asserted in such a language by a person who knows the syntax without knowing a single word of the vocabulary.” and “logical constants” will always either be. Such words or symbols may occur in logic. because they would only be needed for giving values to the variables. such as “x” and “R” and “y.” arranged in various ways. those that can are those that are of the form “xRy. is derivative from propositions of the form: “There is a term c such that φx is true when.” and “greater” for “earlier.” And in every symbolism hitherto invented for mathematical logic there are symbols having constant formal meanings.Chap. they are in essence the same thing. It is one of the marks of a proposition of logic that. We should not need to know any words. The question is: How are we to deﬁne them? Such words or symbols express what are called “logical constants. given a suitable language. what is in common among a group of propositions derivable from each other. We may take as an example the symbol for incompatibility which is employed in building up truth-functions.” This is a function of φ. and the way of arrangement would indicate that something was being said to be true of all values or some values of the variables. or be derived from. For example. A fundamental logical constant will be that which is in common among a number of propositions. “Napoleon is greater than Wellington” results from “Socrates is earlier than Aristotle” by the substitution of “Napoleon” for “Socrates. if it were perfected. and only when. XVIII. which is the business of the applied mathematician.” because the ﬁrst is of the subject. The number . Mathematics and Logic language in which everything formal belonged to syntax and not to vocabulary.” Some propositions can be obtained in this way from the prototype “Socrates is earlier than Aristotle” and some cannot.

| cannot be asserted by logic to be true. and intended to be deﬁned. it is not the case that. Mathematics and Logic propositions result from giving diﬀerent values to φ. There does not even seem any logical . . . or that their contradictories were self-contradictory. We may take the axiom of inﬁnity as an example of a proposition which. where n is any ﬁnite integer. is not satisfactory. the characteristic of logical propositions that we are in search of is the one which was felt. there will be worlds having one. And the same will be found to be the case with all mathematical constants: all are logical constants. All the propositions of logic have a characteristic which used to be expressed by saying that they were analytic. are fallacious. . This is a more diﬃcult matter. In this way. but not of the primitive propositions from which all the propositions of mathematics can be deduced. We may (with a little omission of intermediate steps not relevant to our present purpose) take the above function of φ as what is meant by “the class determined by φ is a unit class” or “the class determined by φ is a member of ” ( being a class of classes). We have suﬃciently deﬁned the character of the primitive ideas in terms of which all the ideas of mathematics can be deﬁned. we may call tautology. We have found so far a necessary but not a suﬃcient criterion of mathematical propositions. We are left to empirical observation to determine whether there are as many as n individuals in the world. three. propositions in which occurs acquire a meaning which is derived from a certain constant logical form. Nevertheless. two. individuals. whatever number n may be. This characteristic. such proofs. Among “possible” worlds. conversely. in the Leibnizian sense. it has no special pre-eminence. owing to types. obviously does not belong to the assertion that the number of individuals in the universe is n. But although all logical (or mathematical) propositions can be expressed wholly in terms of logical constants together with variables. XVIII. or even that there are classes of ℵ terms. as to which it is not yet known what the full answer is. But. as we saw in Chapter XIII. however. and the proof that the contradictory of some proposition is self-contradictory is likely to require other principles of deduction besides the law of contradiction. it would be possible to prove logically that there are classes of n terms. by those who said that it consisted in deducibility from the law of contradiction. or symbolic abbreviations whose full use in a proper context is deﬁned by means of logical constants.. which. But for the diversity of types. though it can be enunciated in logical terms.Chap. The law of contradiction is merely one among logical propositions. for the moment. all propositions that can be expressed in this way are logical. This mode of statement.

so that it is meaningless to argue from “this is the so-and-so” and “the so-and-so exists” to “this exists.e. Although we can no longer be satisﬁed to deﬁne logical propositions as those that follow from the law of contradiction. This is a characteristic. and in fact rests upon a mistaken view of existence—i. Such propositions may occur in logic. it fails to realise that existence can only be asserted of something described. none can be of the form “the propositional function so-and-so is sometimes true. all general propositions would be true. For example.e. They all have the characteristic which.Chap. It is clear that the deﬁnition of “logic” or “mathematics” must be sought by trying to give a new deﬁnition of the old notion of “analytic” propositions. if it were valid. not as complete asserted propositions. no principle of logic can assert “existence” except under a hypothesis. a bearing upon the question what their nature may be. we agreed to call “tautology. It has. But it is generally recognised as invalid. it is not logically necessary. not of something named.) is a proposition asserting existence. if there were no universe. when they occur in logic. We only know from a study of empirical facts that Socrates is a man. or that. | combined with the fact that they can be expressed primitive propositions in Principia Mathematica are such as to allow the inference that at least one individual exists. but we know the correctness of the syllogism in its abstract form (i. We may lay it down that. however. it is always true that if p implies q and q implies r then p implies r. we can and must still admit that they are a wholly diﬀerent class of propositions from those that we come to know empirically. XVIII. a moment ago.e. but of the way in which we know them.” This. without study of the actual world. when it is stated in terms of variables) without needing any appeal to experience. we seem driven to conclude that the existence of a world is an accident—i. The complete asserted propositions of logic will all be such as aﬃrm that some propositional function is always true. and their truth is independent of the existence of the universe. The ontological proof of the existence of God. Logical propositions are such as can be known a priori.e. in fact.” Propositions of this form.” If we reject the ontological | argument. not of logical propositions in themselves. would establish the logical necessity of at least one individual. and would therefore always be false if no universe existed. if all α’s are β’s and x is an α then x is a β. i. If that be so. there should be any world at all. for the contradictory of a general proposition (as we saw in Chapter XV. will have to occur as hypotheses or consequences of hypotheses. Mathematics and Logic necessity why there should be even one individual —why. The . But I now view this as a defect in logical purity. since there are some kinds of propositions which it would be very diﬃcult to suppose we could know without experience.

and the reader is sure. to strain words into unusual meanings. logical symbolism is absolutely necessary to any exact or thorough treatment of our subject.Chap. or even whether he is alive or dead. For the moment. it is to be hoped. it is necessary. so long as we adhere to ordinary language.” so that might be thought to be an adjective qualifying “men. again. | therefore.” It would be easy to oﬀer a deﬁnition which might seem satisfactory for a while. who was working on the problem.. will. after a time if not at ﬁrst. in fact. who wish to acquire a mastery of the principles of mathematics. Moreover. as regards numbers. there are innumerable unsolved problems in the subject. I do not know how to deﬁne “tautology. Because language is misleading. e. importance of “tautology” for a deﬁnition of mathematics was pointed out to me by my former pupil Ludwig Wittgenstein.” It is the case. not shrink from the labour of mastering the symbols—a labour which is. We have now come to an end of our somewhat summary introduction to mathematical philosophy. much less than might be thought. as well as because it is diﬀuse and inexact when applied to logic (for which it was never intended). I do not know whether he has solved it. for the moment. If any student is led into a serious study of mathematical logic by this little book. “ten men” is grammatically the same form as “white men. wherever propositional functions are involved. it will have served the chief purpose for which it has been written. This is the case. XVIII. Mathematics and Logic wholly in terms of variables and logical constants (a logical constant being something which remains constant in a proposition even when all its constituents are changed)—will give the deﬁnition of logic or pure mathematics. and in particular as regards existence and descriptions. At this point. As the above hasty survey must have made evident. in spite of feeling thoroughly familiar with the characteristic of which a deﬁnition is wanted.g. Since ordinary language has no words that naturally express exactly what we wish to express. we reach the frontier of knowledge on our backward journey into the logical foundations of mathematics. and much work needs to be done. to lapse into attaching the usual meanings to words. It is impossible to convey adequately the ideas that are concerned in this subject so long as we abstain from the use of logical symbols. The . therefore. thus arriving at wrong notions as to what is intended to be said. but I know of none that I feel to be satisfactory. ordinary grammar and syntax is extraordinarily misleading. Those readers.

Associative law. . ﬀ. . . . . . . . The page numbers listed are for the original edition. . Diversity. . ﬀ. Correlators. . Contradictions. . . Counting. . Cliﬀord. . Consecutiveness.. . Converse. . . Equivalence. . . extensional and Between. . . Deduction. . . . Convergence. intensional. . . . . Cantorian. of functions. . . ﬀ. Dedekindian. . Georg. . . Derivatives. . . n. ﬀ. . Commutative law. Collections. . .. K. . Distributive law. . Continuity. . Conjunction. . Boundary. . Classes.. All. Constants. Ancestors. . . . inﬁnite. . .INDEX [Online edition note: This is a hyperlinked recreation of the original index. Domain. Extension of a relation.. . Dimensions. . . ﬀ. those marked in the margins of this edition. ﬀ. . Counterparts. . Arithmetising of mathematics. . . Existence. Deﬁnition. Euclid. i. . .e. . . Alephs.. reﬂexive. . Descriptions. logical. ﬀ. . n.. . Argument of a function. . . . . . . method of. Boots and socks.. . . . in philosophy. n. ﬀ. . similar. Construction.. Bolzano. Axioms. . Disjunction. Dedekind. . . . W. Fictions. n. Cantor. Aliorelatives. . ﬀ.] Aggregates. Analysis. . . . Exponentiation. objective.

. . Multiplicative axiom. .. Occam. .. . Incomplete symbols. of functions. . Lewis. . primary and secondary. . . irrational. . . . reﬂexive. ﬀ. ﬁnite. Order. . . n. Frege. . Cantorian. v. . Median class. Integers. real. . . Necessity. multipliable. Ontological proof. Functions. . . . Limit. . . Greater and less. I. mathematical. . . ﬀ.. . Logic. inductive. . . ﬀ. .Index Field of a relation. Nicod. . . Flux. ﬀ. .. ﬀ. . . . analytical. Irrationals. . Induction. .. Hegel. . Likeness. . Inference.. intensional and extensional. Modality. . Intuition. axiom of. and series and ordinals. predicative. . Neighbourhood. Method. .. .. . . . ﬀ. .. mathematical. . . . Mathematics. . Indiscernibles. . relation. .. Instances. ﬀ. ﬀ. . . . Limiting points. Hereditary properties. . . cyclic. Minimum.. . Inﬁnite. . . Occurrences. Oscillation. Number. ﬀ. serial. . cardinal. . . Generalisation. . Incompatibility. . . Incommensurables. . . . . C. non-inductive. ﬀ. ﬀ.. propositional. Inductive properties. . complex. n. Names. . . ﬀ. Geometry. ﬀ. Individuals. . ﬀ. ﬀ. Intervals.. ﬀ. Dedekindian. . . . [].. ﬀ. . . Logicising of mathematics. . maximum ? .. Meinong. . . . of rationals. .. . . . ultimate. ﬀ. Inﬁnity. inﬁnite. . ﬀ. ﬀ. . Form. . . descriptive. . ﬀ. . Gap. . Fractions. Maximum. of cardinals. . . Implication. . n. | Kant. Maps.. positive and negative. . . . Leibniz. . formal. Null-class.. ﬀ. . . . vi. n. . . . . . natural. . . Multiplication. ﬀ.. . . Finite.

. . . . well-ordered. . Rigour. perfect. . ﬀ. . . one-one. Similarity. reﬂexive. . Wells.. . squares of. . symmetrical. . . . . . . axiom of. . Propositions. proper. .. Progressions. . . Space. Relatum. Types. ﬀ. condensed in itself. . Weierstrass. n. . ﬀ. . . ultimate. Truth-value.. . . . Truth-function. . . Variables. Representatives. ﬀ. ﬀ. closed. ﬀ. . ﬀ. ﬀ. Plato. . Successor of a number. Referent. inﬁnite. . generation of. Segments. . Posterity. . . . Tautology. . . . Syllogism. Veblen. . . . . . Royce... . logical. of classes. Plurality. . . ﬀ. ﬀ. . . . . . . . Structure. H. Sequent. . . . . Primitive ideas and propositions. . . ﬀ. . Dedekindian. Wittgenstein.. . . Sub-classes. The. . Reducibility. Poincar´ . Quantity. Value of a function. . ﬀ. . Dedekindian. asymmetrical. . similar. elementary. ﬀ. Sheﬀer. . G. . . . Relations.. . compact. . Pythagoras. Ratios. . Peano. ... . . Section. Some. Philosophy. . . Precedent. . . . Relation-numbers. v. . Permutations. ﬀ. Zero. e Points. Particulars. serial. Series. ﬀ. . . . Unreality. . . of relations. . one-many.. . .Index Parmenides. transitive. . . . . ﬀ. . . Subjects. Zermelo. Time. mathematical. Whitehead. . . many-one. . . . n. . . . analytic. . ﬀ. Peirce. .. Postulates. Premisses of arithmetic. connected. Verbs. . . Selections. . Subtraction. .

ﬁxes. I take full responsibility for any remaining errors. McMaster University. It is based on the April so-called “second edition” published by Allen & Unwin. Since this version uses diﬀerent pagination. Brandon Young. it was necessary to number footnotes sequentially through each chapter. mentioned below. in certain respects. A tremendous debt of thanks is owed to Kenneth Blackwell of the Bertrand Russell Archives/Research Centre. was simply a second printing of the original edition but incorporating various. (February . The pagination of the Allen & Unwin edition is given in the margins. and. as are other additions to the text not penned by Russell. Pierre Grenon. for proofreading the bulk of the edition. ).CHANGES TO ONLINE EDITION This Online Corrected Edition was created by Kevin C. David Blitz. Thanks to members of the Russell-l and HEAPS-l mailing lists for help in checking and proofreading the version.umass.edu. Some are mere stylistic diﬀerences. The online edition diﬀers from the Allen & Unwin edition. John Ongley. and reprintings thereof. Rosalind Carey. and some new ﬁxes. this is version . or ﬁx small grammatical or typographical errors. Others represent corrections based on discrepancies between Russell’s manuscript and the print edition. Another large debt of gratitude is owed to Christof Gr¨ ber who compared this version a to the print versions and showed remarkable aptitude in spotting discrepancies. If you discover any. checking it against Russell’s handwritten manuscript. mostly minor. for example. This edition incorporates ﬁxes from later printings as well. please email me at klement@philos. The stylistic diﬀerences are these: • In the original. by contemporary standards. including Adam Killian. the footnote listed as note on page of this edition was listed as note on page of the original. and providing other valuable advice and assistance. Thus. which. especially. . Klement. These are in red. with page breaks marked with the sign “|”. footnote numbering begins anew with each page.

but vertical fractions x of the form “ y ” in displays. . . the mid-paragraph fractions on pages .e. .” is ﬁxed to “. provided neither m nor n is zero. . ” to match Russell’s manuscript and the obviously intended meaning of the passage. i. (page / original page ) “. and of the original (pages . . (page n. . those in the display on page of the original (page of this edition) were linear.Changes to Online Edition • With some exceptions. (page / original page ) “. (page / original page ) “. . its limit or maximum. . . . A. ” is changed to “. and then . the limit of its value for approaches either from . / original page n. but have been converted to vertical fractions in this edition. but here are horizontal. pp. This error was noted by Pfeiﬀer in but unﬁxed in Russell’s lifetime. i. . . .” Thanks to John Ongley for spotting this error. . which exists even in Russell’s manuscript. ” is changed to “. . (page / original page ) “. . either by limiting the domain to males or by limiting the converse to females” is changed to “. . . Contrary to this usual practice. ” to match Russell’s manuscript. . . of a class α. A few were originally noted in an early review of the book by G. . the limit of its . . Most of these result from Ken Blackwell’s comparison with Russell’s manuscript. The following more signiﬁcant changes and revisions are marked in green in this edition. . and the apparent meaning of the passage. . . this has been ﬁxed to . . . either by limiting the domain to males or by limiting the converse domain to females”. . . of a class α.e. / original page n. . . . Pfeiﬀer (Bulletin of the American Mathematical Society : (). . . ” is changed to “. its limits or maximum. the Allen & Unwin edition uses linear fractions of the style “x/y” mid-paragraph. . . which is how it read in Russell’s manuscript. . provided neither m or n is zero. Similarly. (page / original page ) “. or all that are less than . –).) Russell wrote the wrong publication date () for the second volume of Principia Mathematica. or all that are not less than . and here) were printed vertically in the original. and seems better to ﬁt the context. (page n.) The word “deutschen” in the original’s (and the manuscript’s) “Jahresbericht der deutschen Mathematiker-Vereinigung” has been capitalized. . and then . .

were omitted. ” to match Russell’s manuscript.Changes to Online Edition values for approaches either from . . (page / original page ) The words “and then the idea of the idea of Socrates” although present in Russell’s manuscript. the negation of propositions of the type to which φx belongs . . all terms z such that x has the relation P to x and z has the relation P to y . . . This requires a”. . . if x is the member of γ . x is a member of y . The missing words are now restored. (page / original page ) “. . and so on. x is a member of γ . . were left out of previous print editions. x is a member of γ . thereby amalgamating two sentences into one. Note that Russell mentions “all these ideas” in the next sentence. was attached in previous versions to the sentence that now refers to the ﬁrst footnote in the chapter. . the negation of propositions of the type to which x belongs . . The footnote references have been returned to where they had been placed in Russell’s manuscript. and at his request. ” is changed to “. (page / original page ) The passage “. . the reference to Principia Mathematica ∗ . and so on. . (page / original page ) The two footnotes on this page were misplaced. . . ” Russell himself hand-corrected this in his manuscript. . and is more appropriate for the meaning of the passage. ” to match Russell’s manuscript. . ”. . x is a member of y . . . all terms z such that x has the relation P to z and z has the relation P to y . . ” is changed to “. . ” is changed to “. advantage of this form of deﬁnition is that it analyses . ” to match Russell’s manuscript. (page / original page ) The ungrammatical “. . That footnote was placed three sentences below. . This is another error noted by Pfeiﬀer. it was changed in the printing. then . . The second. . . . . . . if x is the member of y . . which constitute exactly one line of Russell’s manuscript. . . (page / original page ) The words “correlator of α with β. . . and similarly for every other pair. and the obviously intended meaning of the passage. . . . . ” is ﬁxed to “. but not in a clear way. advantages of this form of deﬁnition is that it analyses . which matches Russell’s manuscript. . then . . . (page / original page ) “.

though it does appear in Russell’s manuscript. . . (page / original page ) The Allen & Unwin printings have the sentence as “How shall we deﬁne a “typical” Frenchman?” Here. ” is changed to “Suppose we are considering “all men are mortal”: we will . for practical purposes. without the indeﬁnite article. . ” has been changed to “There is a type (τ say) . ” were omitted from previous published versions. . was excluded from the Allen & Unwin printings. . ” to match the obviously intended meaning of the passage. (page / original page ) The “φ” in “. as opposed to speciﬁc man. . but do appear in Russell’s manuscript. ” to match Russell’s manuscript. . present in Russell’s manuscript. (page / original page ) “Suppose we are considering all “men are mortal”: we will . . Again thanks to John Ongley. and the placement of the opening quotation mark in Russell’s manuscript (although he here used single quotation marks. . . and has been restored. . . divided into a number of separate studies . (page / original page ) “. . Some . .” is ﬁxed to “. while intensional functions cannot . (page / original page ) “There is a type (r say) . (page / original page ) “. . as he did sporadically throughout). the closing quotation mark has been moved to make it “How shall we deﬁne a “typical Frenchman”?” Although Russell’s manuscript is not entirely clear here. . . . . the process of applying general statements about φx to particular cases . . . in the singular.” Russell sent this change to Unwin in . . . . . ” Russell’s manuscript just had “number”. . . . which had also been noted by Pfeiﬀer. . a . it appears the latter was intended. . . . . . (page / original pages –) The two occurrences of “φ” in “. Thanks to Christof Gr¨ ber for spotting this error. resulting from a propositional function φx by the substitution of . and it was made in the printing. ”. and it also seems to make more sense in context. . . be regarded as functions of the class determined by φx. . and seems necessary for the passage to make sense. (page / original page ) The “φ” in “. Thanks to John Ongley for spotting this error. . as opposed to speciﬁc men. ” has been changed to “.Changes to Online Edition . divided into numbers of separate studies . and conventions followed elsewhere in the chapter. extensional functions of a function φx may. ” was excluded from previous published versions.

” is changed to “. . . . ”. Thanks to Christof Gr¨ ber for noticing this error. (Again thanks to Christof Gr¨ ber. is the ﬁeld of Q. and which is . . . a relation number is a class of . without any special word for forms . the “s” in “y’s” in what appears here as “Form all such sections for all y’s . . . ” f) (pages – / original pages –) One misprint of “progession” for “progression”. and “[]” inserted. . and is changed to do so. c) (page / original page ) What appears in the and later printings as “. . (page / original page ) The passage “the propositional function ‘if all α’s are β and x is an α. . the discussion of Frege occurs on page . . . . ” is changed to “. and which is . ” has been changed to “. and one misprint of “progessions” for “progressions”. . . . but not marked in green. without any special words for forms .) g) (page / original page ) In the Allen & Unwin printing. e) (page / original page ) “. is the ﬁeld of Q. . and “featherless biped”—so two . . . . . . but does in Russell’s manuscript. Here. have been corrected. .”—so two . . a relation-number is a class of . however. . then x is a β’ is always true” has been changed to “the propositional function ‘if all α’s are β’s and x is an α. . ” was italicized along with the “y”. as well as to make it consistent with the other paraphrase given earlier in the sentence. which matches Russell’s manuscript and seems to ﬁt better in the context. . . . and “featherless biped.) a . b) (page / original page ) “We may deﬁne two relations .Changes to Online Edition emendation was necessary to make the passage grammatical. a) Ellipses have been regularized to three closed dots throughout. but in fact. “” is crossed out. a . . (Thanks to Christof Gr¨ ber a for noticing these errors in the original. . Nothing in Russell’s manuscript suggests it should be italicized. but the ﬁx adopted here seems more likely what was meant. Some very minor corrections to punctuation have been made to the Allen & Unwin printing. ” to match the hyphenation in the rest of the book (and in Russell’s manuscript). ” is changed to “. . . . . ” did not start a new paragraph in previous editions. then x is a β’ is always true” to match Russell’s manuscript. A similar change is made in the index. (page / original page ) “. . (page / original page ) The original index listed a reference to Frege on page . ” d) (page / original page ) “. . .

and x for Socrates. . . it is italicized here. l) (page / original page ) “.s. ” is changed to “. . (): –. “Let y be a member of β . There are. ” in the Allen & Unwin editions. however. but was marked to be in Russell’s manuscript. . . For a detailed examination of the diﬀerences between Russell’s manuscript and the print editions. ” and has been made to start a new paragraph. a number of other places where the previous print editions diﬀer from Russell’s manuscript in minor ways that were left unchanged in this edition. and it seems natural to do so for the sake of consistency. ﬀ.Changes to Online Edition h) (page / original page ) In the Allen & Unwin printing. we may substitute α for men. One of Peano’s . .” has been changed to “similar. most diﬃcult of fulﬁlment—it must . see Kenneth Blackwell.—and clearly . or what not—and clearly . k) (page / original page ) The accent on “M´taphysique”. . and clearly should not.— it must . a deﬁnition which might seem satisfactory for a while . . so it has been italicized. . ” is changed to “. n) (page / original page ) “.” to match the punctuation elsewhere. . . . . j) (page / original page ) “The way in which the need for this axiom arises may be explained as follows:—One of Peano’s . Misprints and a Bibliographical Index for Introduction to Mathematical Philosophy”. ” is changed to “The way in which the need for this axiom arises may be explained as follows. has been restored. . q) (page / original page ) Under “Relations” in the index. . “Socrates” was not italicized in “. ” Russell had marked it for italicizing in the manuscript. “similar. e included in Russell’s manuscript but left oﬀ in print. ﬀ. “Variants. . . . . . . . ” begins a new paragraph. . . p) (page / original page ) The word “seem” was not italicized in “. . . ” m) (page / original page ) Italics have been added to one occurrence of “Waverley” to make it consistent with the others. as it did in Russell’s manuscript. . β for mortals. and between the various printings themselves (including the changes from the to the printings not documented here). or what not. Russell n. but it does not in Russell’s manuscript. where . i) (page / original pages –) The phrase “well ordered” has twice been changed to “well-ordered” to match Russell’s manuscript (in the ﬁrst case) and the rest of the book (in the second). . ” o) (page / original page ) In the Allen & Unwin printings. most diﬃcult of fulﬁlment. .. . .

See http://creativecommons.Changes to Online Edition p Bertrand Russell’s Introduction to Mathematical Philosophy is in the Public Domain. United States License.org/licenses/publicdomain/ A are licensed under a Creative Commons Attribution—Share Alike . See http://creativecommons./us/ cba This typesetting (including LTEX code) and list of changes .org/licenses/by-sa/.

Sign up to vote on this title

UsefulNot useful- Bertrand Russell, The Analysis of Mind
- Bertrand Russell is There a God
- Russell, Bertrand - The.philosophical.importance.of.Mathematical.logic
- Bertrand Russell - The Problems of Philosophy
- Logic in the early 20th century (Routledge Encyclopedia of Philosophy)
- Posy - Kant's Philosophy of Mathematics.pdf
- Mathematics under the Microscope
- Bertrand Russell - The Russell Einstein Manifesto
- Gill
- How Bertrand Russell Became An Evil Man
- Bertrand Russell
- Bertrand Russell - Portraits From Memory and Other Essays (1956)
- Fermat's Last Theorem
- Mathematics for Physics
- Frege Foundations of Arithmetic
- Analysis of Mind
- Barker PhilosophyOfMathematics
- Unit 4 "Science and Human Life" by Bertrand Russel
- Russel's Introduction to mathematical philosophy
- Mathematical Philosphy
- Introduction to Mathematical Philosophy-A4x2
- Bertrand Russell, Introduction to Mathematical Philosophy
- Ultra Finit is m
- Moodle Fma1 Problems Complete
- Mathematics and Mathematical Logic
- Bertrand Russell Logicomix
- lecture1.pdf
- Mathematics Made Difficult - Carl E. Linderholm
- Bertrand Russell Introduction to Mathematical Philosophy