You are on page 1of 414

SEMICENTENNIAL PUBLICATIONS

OF THE

UNIVERSITY OF CALIFORNIA

1868-1918

A SURVEY OF

SYMBOLIC LOGIC

BY
C.
I.

LEWIS

UNIVERSITY OF CALIFORNIA PRESS

BERKELEY
1918

PRESS OF THE NEW ERA PRINTING COMPANY LANCASTER, PA.

TABLE OF CONTENTS
PREFACE
v
I.

CHAPTER
SECTION

THE DEVELOPMENT OF SYMBOLIC LOGIC


I.

The Scope
and

Symbolic Logic. Symbolic Logic Summary Account of their Logistic.


I
5-

of

Development
SECTION
SECTION SECTION
II.

Leibniz

III.

From Leibniz to De Morgan and Boole

18

IV.

De Morgan
Boole

37
51

SECTION
SECTION

V.
VI.

Jevons
Peirce

72
79
107

SECTION VII.
SECTION VIII.

Developments since Peirce

CHAPTER
SECTION SECTION SECTION
SECTION

II.

THE CLASSIC, OR BOOLE-SCHRODER AL GEBRA OF LOGIC


I.

118

General Character of the Algebra.

The Postulates
118

and
II.

their Interpretation

Elementary Theorems
General Properties of Functions

122

III.

132
of Equations.
. .

SECTION IV.
V.

Fundamental Laws Fundamental Laws

of the

Theory

144

SECTION VI.

of the Theory of Inequations. 166 Note on the Inverse Operations, "Subtraction" 173 and "Division"

CHAPTER

III.

APPLICATIONS OF THE BOOLE-SCHRODER

ALGEBRA
SECTION
SECTION SECTION
I.

175
175

Diagrams

for the Logical Relations of Classes ....

II.

III.

SECTION IV.

The Application The Application The Application

to Classes

184

to Propositions

213 219

to Relations

CHAPTER
SECTION

IV.

SYSTEMS BASED ON MATERIAL IMPLI CATION


The Two-Valued Algebra
iii

222
222

I.

IV

Table of Contents

SECTION
SECTION

II.

III.

The Calculus of Prepositional Functions. Func 232 tions of One Variable Functions of Two or More Variables. 246 Prepositional
Derivation of the Logic of Classes from the Calcu
lus of Propositional

SECTION IV.

Functions

260 269

SECTION

V.

SECTION VI.

The Logic The Logic

of Relations of Principia

Mathematica

279
291

CHAPTER
SECTION
SECTION

V.
I.

THE SYSTEM OF STRICT IMPLICATION...


Primitive Ideas, Primitive Propositions, and

Im
292
299

II.

mediate Consequences Strict Relations and Material Relations

SECTION

III.

The Transformation
of Consistencies

{-/-)

306

SECTION IV.

Extensions of Strict Implication.

The Calculus
of

and the Calculus

Ordinary
316

Inference

SECTION

V.

The Meaning

of

"Implies"

324

CHAPTER
SECTION

VI.

SYMBOLIC LOGIC, LOGISTIC, AND MATHE MATICAL METHOD


General Character of the Logistic Method.
"Orthodox"

340

I.

The
340

View

SECTION

II.

Two

Varieties of Logistic

Method

Peano

Formu343

laire

and Principia Mathematica.

The Nature
of

of Logistic Proof

SECTION

III.

"Heterodox"

View
of

of the

Nature

Mathe
354
362

matics and of Logistic

SECTION IV. SECTION


V.

The

Logistic

Method

Kempe and Royce


.

Summarv and

Conclusion

367

APPENDIX.

TWO FRAGMENTS FROM

LEIBNIZ

373 389 407

BIBLIOGRAPHY
INDEX.

PREFACE
The student who has completed some elementary study
of symbolic
logic and wishes to pursue the subject further finds himself in a discouraging He has, perhaps, mastered the contents of Venn s Symbolic situation.

The Algebra of Logic, or the chapters concerning this subject in Whitehead s Universal Algebra. If he read German with sufficient ease, he may have made some excursions into
Logic or Couturat
s

admirable

little

book,

Schroder

Vorlesungen uber die Algebra der Logik.

These

all

concern the
is

classic, or Boole-Schroder algebra,

and

his

knowledge

of symbolic logic

probably confined to that system. His further interest leads him almost inevitably to Peano s Formulaire de Mathematiques, Principia Mathematica
of of the

Whitehead and Russell, and the increasingly numerous shorter studies same sort. And with only elementary knowledge of a single kind of

difficult

development of a small branch of the subject, he must attack these most and technical of treatises, in a new notation, developed by methods
which are entirely novel to him, and bristling with logico-metaphysical If he is bewildered and searches for some means of further difficulties.
preparation, he finds nothing to bridge the gap.

Schroder

work would

be of most assistance here, but this was written some twenty-five years
ago; the most valuable studies are of later date, and radically

new methods

have been introduced.

What
ject

such a student most needs


will familiarize

is

a comprehensive survey of the sub


single

one which

him with more than the

system

which he knows, and

not only the content of other branches and the alternative methods of procedure, but also the relation of these to
will indicate

the Boole-Schroder algebra and to one another.

The

present book

is

an

attempt to meet this need, by bringing within the compass of a single


volume, and reducing to a
of the requirements of a

common

notation (so far as possible), the most


logic.
If,

important developments of symbolic


"handbook"

in addition to this,

some

are here fulfilled, so

much

the better.

But
in

this

survey does not pretend to be encyclopedic.

A
is

gossipy recital

of results achieved, or a superficial account of

methods,

of

no more use

symbolic logic than in any other mathematical discipline.

What

is

presented must be treated in sufficient detail to afford the possibility of real


insight

and grasp.

This aim has required careful selection of material.

vi

Preface
historical

The
of

summary

in

Chapter

attempts to follow the main thread

development, and no reference, or only passing mention, is given to those studies which seem not to have affected materially the methods of
later researches.

In the remainder of the book, the selection has been

governed by the same purpose. Those topics comprehension of which seems most essential, have been treated at some length, while matters less

fundamental have been

set forth in outline only, or

omitted altogether.

My

own contribution to symbolic logic, presented in Chapter V, has not earned the right to inclusion here; in this, I plead guilty to partiality. The discussion of controversial topics has been avoided whenever possible
and, for the
rest, limited to

the simpler issues involved.

Consequently,

the reader must not suppose that any sufficient consideration of these
questions
accurate.
is

here given, though such statements as are

made

will be, I

hope,

Particularly in the last chapter, on


",

"Symbolic

Logic, Logistic,

and Mathematical Method


adequate account of the
size of this one.

it

is

not possible to give anything like an

facts.
I

Rather,

That would require a volume at least the have tried to set forth the most important and
arbitrarily

critical considerations
is

somewhat
and

and dogmatically,

since there

not space for argument

to provide such a

map

of this difficult terri

tory as will aid the student in his further explorations.


Proofs and solutions in Chapters II, III, and IV have been given very
fully.

Proof

is

of the essence of logistic,


fair

and
of

it is

my

observation that stu

dents

even those with a

knowledge

mathematics

seldom

command

the technique of rigorous demonstration.

In any case, this explicitness can do no harm, since no one need read a proof which he already understands.
I

am

indebted to

many

friends

and colleagues

for valuable assistance in

preparing this book for publication: to Professor


tions of

W.

A. Merrill for

emenda

my

translation of Leibniz, to Professor J.

H. McDonald and

Dr. B. A. Bernstein for important suggestions and the correction of certain errors in Chapter II, to Mr. J. C. Rowell, University Librarian, for assistance
in securing a

number

of rare volumes,

and to the

officers of the

University

Press for their patient helpfulness in meeting the technical difficulties of


printing such a book.

manuscript, eliminated

Mr. Shirley Quimby has read the whole book in many mistakes, and verified most of the proofs.
indebted to

But most

of all, I

am

my

friend

and teacher, Josiah Royce,


failed to give
is

who

first

aroused

my

interest in this subject,

and who never


that

me encouragement and
due to him.
BERKELEY, July
10,

wise counsel.

Much

best in this
C>

book

is

L LEWIS.

1917.

CHAPTER
I.

THE DEVELOPMENT OF SYMBOLIC LOGIC


THE SCOPE OF SYMBOLIC LOGIC. SYMBOLIC LOGIC AND SUMMARY ACCOUNT OF THEIR DEVELOPMENT
The
to as

LOGISTIC.

subject with which


"

we

are concerned has been variously referred


"algebra
logic",

"symbolic

logic",

logistic",

of

logic",

"calculus

of

logic",

"mathematical

logic",

"algorithmic

And none
because

of these

is

satisfactory.

We
in

and probably by other names. have chosen "symbolic logic"


in this country,

it is

the most

commonly used
is

England and
understood.

and
is

because

its

signification

pretty

well

Its

inaccuracy

obvious: logic of whatever sort uses symbols. We are concerned only with that logic which uses symbols in certain specific ways those ways which are exhibited generally in mathematical procedures. In particular,
logic to
relations,

be called

symbols for the logical and must so connect various relations that they admit of "trans
"symbolic"

must make use

of

formations"

and

"operations",

according to principles which are capable

of exact statement.
If

we must
is

give

some

definition,

we

shall

hazard the following: Symbolic

Logic

the development of the most general principles of rational pro

cedure, in ideographic
of these principles

symbols, and in a form which exhibits the connection


Principles which belong exclusively
e.

one with another.


of rational procedure

to

some one type

g.

to dealing with
is

number and

quantity

are hereby excluded,

and generality

designated as one of the

marks
form.

of symbolic logic.

Such general principles are likewise the subject matter

of logic in

any

To

be sure, traditional logic has never taken possession of more


field

than a small portion of the

are unnecessarily restricted.

As we

which belongs to it. The modes of Aristotle shall have occasion to point out, the

reasons for the syllogistic form are psychological, not logical: the syllogism,

made up of the smallest number of propositions (three), each with the small est number of terms (two), by which any generality of reasoning can be attained, represents the limitations of human attention, not logical necessity. To regard the syllogism as indispensable, or as reasoning par excellence, is
2
1

Survey of Symbolic Logic

the apotheosis of stupidity.

And

the procedures of symbolic logic, not


to

being thus arbitrarily restricted,

matter between

it

and the traditional

accidental, not essential,


logic
all
is

mark a difference of subject But any such difference is logic. and the really distinguishing mark of symbolic

may seem

the approximation to a certain form, regarded as ideal.

There are

degrees of such approximation; hence the difficulty of drawing any hard


fast line

between symbolic and other logic. But more important than the making of any such sharp distinction is the comprehension of that ideal of form upon which it is supposed to

and

depend.

The most convenient method which the human mind has


is

so far

devised for exhibiting principles of exact procedure


call,

the one which

we

in

general terms, mathematical.


(1)

The important
which

characteristics of

this

form are:

the use of ideograms instead of the phonograms of


(2)

ordinary language;
to

the deductive method

may

here be taken
is

mean simply

that the greater portion of the subject matter

derived

from a relatively few principles by operations which are


(3)

"exact";

and

the use of variables having a definite range of significance.

Ideograms have two important advantages over phonograms. In the first place, they are more compact, + than 3 than "three", etc.
"plus",

This

is

no inconsiderable gain, since


of

it

makes

possible the presentation of a

formula in small enough compass so that the eye


glance and the image
reference with a
it

(in visual or other terms)

may apprehend it at a may be retained for

minimum

of effort.

None but

a very thoughtless person,

or one without experience of the sciences, can fail to understand the enor

mous advantage
notation
is

of such brevity.

In the second place, an ideographic

superior to

any other in precision.

Many

ideas which are

quite simply expressible in mathematical symbols can only with the greatest

be rendered in ordinary language. Without ideograms, even arithmetic would be difficult, and higher branches impossible.
difficulty

The deductive method, by which a considerable array

of facts

is

sum

marized in a few principles from which they can be derived, is much more than the mere application of deductive logic to the subject matter in It both requires and facilitates such an analysis of the whole question.

body

of facts as will

most

precisely exhibit their relations to one another.


is

In fact, any other value of the deductive form

largely or wholly fictitious.

presentation of the subject matter of logic in this mathematical form constitutes what we mean by symbolic logic. Hence the essential
characteristics of our subject are the following:
(1) Its

The

subject matter

is

The Development of Symbolic Logic


the subject matter of logic in any form

that

is,

the principles of rational

or reflective procedure in general, as contrasted with principles which


(2) Its belong exclusively to some particular branch of such procedure. medium is an ideographic symbolism, in which each separate character

represents a relatively simple and entirely explicit concept.


all

And,

ideally,

non-ideographic symbolism or language

is

excluded.
"

(3)

Amongst the
system)

ideograms, some will represent variables (the

terms"

of the

having a definite range of significance.

Although
"

it

is

non-essential, in
"individuals",

any system
they

so far developed the variables will represent

or classes, or relations, or propositions, or


will represent

prepositional

functions",

or

ambiguously some two or more of these. (4) Any system of symbolic logic will be developed deductively that is, the whole body of its theorems will be derived from a relatively few principles, stated
in symbols,

We

by operations which are, or at least can be, precisely formulated. have been at some pains to make as clear as possible the nature of
logic,

symbolic

because

its

distinction

from

"ordinary"

logic,

on the one

hand, and, on the other, from any mathematical discipline in a sufficiently abstract form, is none too definite. It will be further valuable to comment
briefly

on some of the alternative designations

for the subject

which have

been mentioned.

would not have served our purpose, because "logistic" is used to denote symbolic logic together with the application of commonly its methods to other symbolic procedures. Logistic may be defined as
"Logistic"

the science

which deals with types of order as such. It is not so much a subject as a method. Although most logistic is either founded upon or makes large use of the principles of symbolic logic, still a science of order

in general does not necessarily presuppose, or begin with,

symbolic

logic.

Since the relations of symbolic logic, logistic, and mathematics are to be

the topic of the last chapter,

that matter here.

We
is

have mentioned have


in the

we may postpone any further discussion of it only to make clear the meaning
pages which follow.
It

which

"logistic"

to

comprehends

symbolic logic and the application


fies

of such

methods

as symbolic logic exempli


is

to other exact procedures.


"Algebra

Its subject

matter

not confined to

logic.

of

logic"

is

hardly appropriate as the general

name

for our

subject, because there are several quite distinct algebras of logic,

and

because symbolic logic includes systems which are not true algebras at all. "The algebra of logic" usually means that system the foundations of

which were

laid

by Leibniz, and

after

him independently by

Boole,

and

Survey of Symbolic Logic

which was completed by Schroder.


"

We

shall refer to this

system as the
a

Boole-Schroder Algebra
"Calculus"

".

is

a more general term than

"algebra".

By

"calculus"

will

be meant, not the whole subject, but any single system of assumptions
their consequences.
for symbolic logic

and

The program both


clear form,

and

for logistic, in anything like a

sketched by Leibniz, though the ideal of logistic seems 1 Leibniz left frag to have been present as far back as Plato s Republic.

was

first

mentary developments of symbolic logic, and some attempts at logistic which are prophetic but otherwise without value. After Leibniz, the two
interests

somewhat

diverge.

Contributions to symbolic logic were

made by

Ploucquet, Lambert, Castillon

and others on the continent.

This type of
contribution

research interested Sir William Hamilton and, though his

own

was

slight

and not

essentially novel, his papers were, to

some extent at

least, responsible for

the renewal of investigations in this field which took

place in

Boole.

England about 1845 and produced the work of De Morgan and Boole seems to have been ignorant of the work of his continental

predecessors, which

proved so
subject,

much

probably fortunate, since his own beginning has more fruitful. Boole is, in fact, the second founder of the
is

and

all later

work goes back

to his.

The main

line of this

develop

ment runs through Jevons, C.


Vorlesungen
iiber die

S. Peirce,

and MacColl to Schroder whose


I,

Algebra der Logik (Vol.

1890) marks the perfection

of Boole s algebra

and the logical completion of that mode of procedure. In the meantime, interest in logistic persisted on the continent and
fostered

by the growing tendency to abstractness and rigor in mathe matics and by the hope for more general methods. Hamilton s quaternions

was

and the Ausdehnungslehre of Grassmann, which was recognized as a con tinuation of the work begun by Leibniz, contributed to this end, as did also
the precise logical analyses of the nature of
Also, the elimination

number by Cantor and Dedekind.


of all

from

"modern

geometry"

methods

.of

proof

dependent upon

"intuitions

of

space"

or

"construction"

brought that

subject within the scope of logistic treatment, and in 1889

such a treatment in I Principii di Geometria.


Begri/sschrift of

Frege 1879 to the Grundgesetze der Arithmetik (Vol.

Peano provided works, from the


I,

1893;

Vol. II, 1903) provide a comprehensive


logistic
1

development

of arithmetic

by the

method.

or philosophic development of mathematics in Bk. vi, Step. 510-11

See the criticisms of contemporary mathematics and the program for the dialectic and Philebus, Step. 56-57.

The Development of Symbolic Logic

In 1894, Peano and his collaborators began the publication of the

Formulaire de Mathematiques, in which

all

branches of mathematics were to


logistic.

be presented in the universal language of


logic

In this work, symbolic

and

logistic are once

more brought

together, since the logic presented

in the early sections provides, in a

way, the method by which the other

branches of mathematics are developed.


production.
logistic,

The Formulaire

is

a monumental

But

its

mathematical interests are as much encyclopedic as


the possibilities of the method are utilized or

and not
It

all

made
in

clear.

remained for Whitehead and Russell, in Principia Mathematica,

to exhibit the perfect union of symbolic logic and the logistic

method

mathematics.

The

publication of this

in the history of the subject. of the algebra of logic

work undoubtedly marks an epoch The tendencies marked in the development


in the

development of the and in the foundations algebra of relatives from De Morgan to Schroder, for number theory of Cantor and Dedekind and Frege, are all brought 2 Further researches will most likely be based upon the together here.
formulations of Principia Mathematica.

from Boole to Schroder,

We
history

must now turn back and


3

trace in

more

detail the

development
if

of

symbolic
is

logic.

history of the subject will not be attempted,

by
are

meant the report

of facts for their

own

sake.

Rather, we

interested in the cumulative process


interest us

by which those

results

which most

today have come to be. Many researches of intrinsic value, but lying outside the main line of that development, will of necessity be neglected. Reference to these, so far as we are acquainted with them, will
be found in the bibliography. 4
II.

LEIBNIZ
logistic properly begins

The
In the
say:
6

history of symbolic logic

and

with Leibniz. 5

New
"I

Essays on the Human Understanding, Philalethes is made to begin to form for myself a wholly different idea of logic from
I

that which I formerly had.

regarded

it

as a scholar
it,

diversion, but I

now
2

see that, in the

way you understand

it is like

a universal mathe-

Hilbert, Fieri,

Perhaps we should add "and the modern development of abstract geometry, as by and others", but the volume of Principia which is to treat of geometry has

not yet appeared. 3 The remainder of this chapter is not essential to an understanding of the rest of the book. But after Chapter i, historical notes and references are generally omitted. 4 Pp. 389-406. 5 Leibniz regards Raymond Lully, Athanasius Kircher, John Wilkins, and George Dalgarno (see Bibliography) as his predecessors in this field. But their writings contain
little
6

which is directly to the point. Bk. iv, Chap, xvn, 9.

A
As

Survey of Symbolic Logic

matics."

this passage suggests, Leibniz correctly foresaw the general


it

character which logistic was to have and the problems


to solve.

would

set itself

But though he caught the

large outlines of the subject

and

actually delimited the field of work, he failed of


of the difficulties to be met,

any and he contributed comparatively


Perhaps
partial explanation,
is

clear understanding
little

to

the successful working out of details.

this is characteristic of the

man.

But another explanation, or

possible.

Leibniz

expected that the whole of science would shortly be reformed by the appli cation of this method. This was a task clearly beyond the powers of any

one man, who could, at most,


plan.

offer
life,

only the

initial

stimulus and general

And
and

so,

throughout his

he besought the assistance of learned

societies

titled patrons, to

the end that this epoch-making reform might

be instituted, and never addressed himself very seriously to the more


limited tasks which he
7 might have accomplished unaided.

Hence
of

his

studies in this field are scattered through the manuscripts,


still

many

them

unedited, and out

of five

hundred or more pages, the systematic

results

attained might be presented in one-tenth the space. 8

Leibniz

conception of the task to be accomplished altered somewhat

during his

life,

but two features characterize

all

the projects which he


language"

entertained:
language"

(1) a universal
"universal

medium

("universal

or
of

"rational

or

characteristic")

for the expression

science;

and

(2) a calculus of reasoning (or

"universal calculus")

designed to display

the most universal relations of scientific concepts and to afford some sys

tematic abridgment of the labor of rational investigation in all fields, much as mathematical formulae abridge the labor of dealing with quantity and

number.
that
is

"The

true

method should furnish us with an Ariadne

thread,

and palpable medium, which will guide the mind as do the lines drawn in geometry and the formulae for 9 operations which are laid down for the learner in arithmetic."
to say, with a certain sensible

This universal medium is to be an ideographic language, each single character of which will represent a simple concept. It will differ from
existing ideographic languages, such as Chinese, through using a
7

combina-

editor s introduction to "Scientia Generalis. in Gephardt s Characteristic Philosophischen Schriften von Leibniz (Berlin, 1890), vn, gives an excellent account of Leibniz s correspondence upon this topic, together with other material of historic interest. (Work hereafter cited as G. Phil.}
a"

The

See Gerhardt, op.

tit.

especially iv
will

(1901), gives a survey study of the sources.


9

which

and vn. But Couturat, La logique de Leibniz prove more profitable to the general reader than any
vii, 21.

Letter to Galois, 1677, G. Phil,

The Development of Symbolic Logic


tion of symbols, or

some

similar device, for a

compound

idea, instead of

having a multiplicity of characters corresponding to the variety of things.

So that while Chinese can hardly be learned in a lifetime, the universal 10 The fundamental char characteristic may be mastered in a few weeks.
acters of the universal language will be few in

number, and

will represent

the

"alphabet

of

human

thought":

The

fruit of

many

analyses will be the

catalogue of ideas which are simple or not far from


catalogue of primitive ideas
of science
is

n
simple."

With

this

this alphabet of

human thought

the whole

to be reconstructed in such wise that its real logical organiza


its

tion will be reflected in

symbolism.

In spite of fantastic expression and some hyperbole,


the program of logistic.
If

we

recognize here
is

the reconstruction of

all

science

a project too
desirability

ambitious,

still

we should maintain the


its

ideal possibility

and the

of such a reconstruction of exact science in general.

And

the ideographic

language finds
matica, and

realization in

Peano

Formulaire, in Principia Mathe-

in all successful applications of the logistic

method.

Leibniz stresses the importance of such a language for the more rapid

and orderly progress


least effect of it
".

of science
. .

and

of

human thought

in general.

The

will

be the universality and communication of

different nations.

Its true use will

be to paint not the word

but the
.
. .

thought, and to speak to the understanding rather than to the eyes.

Lacking such guides, the mind can make no long journey without losing with such a medium, we could reason in metaphysics its way
.

and

in ethics

very

much
fix

as

we do

characters would

our ideas,
12

geometry and in analytics, because the which are otherwise too vague and fleeting
in
it

in such matters in

which the imagination cannot help us unless

be by

the aid of

characters."
"The

The

lack of such a universal

medium

prevents

cooperation.

human

race, considered in its relation to the sciences

which serve our welfare, seems to


in confusion in the darkness,

comparable to a troop which marches without a leader, without order, without any

me

word or other
one another.
of the road,

signs for the regulation of their

Instead of joining hands to

march and the recognition of guide ourselves and make sure


another."
13

we run

hither

and yon and


complex
(?),

interfere with one

The
10
11
"

"alphabet

of

human
of

thought" is

more

visionary.

The

possibility

of constructing the whole of a


Letter to the

science from a few primitive conG. Phil., vn, 24-25.

Duke

Hanover, 1679

G. Phil., vii, 84. G. Phil., vii, 21.

13

G. Phil, vii, 157.

A
is,

Survey of Symbolic Logic

cepts

indeed, real

vide the

few primitives of Principia Mathematica.

But we should today recognize a certain arbitrariness in the selection of The these, though an arbitrariness limited by the nature of the subject.
secret of Leibniz s faith that these primitive concepts are fixed in the nature
of things will

believes that

all

be found in his conception of knowledge and of proof. He predicates are contained in the (intension of the) subject

and may be discovered by analysis. Similarly, all truths which are not absolutely primitive and self-evident admit of reduction by analysis into
such absolutely
thing
"real"

first

truths.

And

finally,

only one real definition of a


is
14

as opposed to

"nominal"

possible;

that

is,

the result
in

of the correct analysis of

any concept

is

unambiguously predetermined

the concept

itself.

The
and he

construction, from such primitives, of the complex concepts of


"synthesis"

the various sciences, Leibniz speaks of as


is

or

"invention",

concerned about the


is

"art

of

invention".

But while the

result of

analysis

synthesis,

character.

always determined, and only one analysis is finally correct, like inverse processes generally, has no such predetermined In spite of the frequent mention of the subject, the only im
of a calculus of reasoning.

portant suggestions for this art have to do with the provision of a suitable

medium and
essay,

To be

sure there are such obvious

counsels as to proceed from the simple to the complex,

and

in the early

De

Arte Combinatoria, there are studies of the possible permutations


"syntheses"

and combinations or
later regarded this

of

fundamental concepts, but the author


value.

Specimina he says that the utmost which we can hope to accomplish at present, toward the general art of invention, is a perfectly orderly and finished reconstruction of existing science in terms of the
Scientice novce Generalis,

study as of

little

And

in Initia et

absolute primitives which analysis reveals. 15

After

two hundred

years,

we

are

fields

without any general method by which logistic may be used in as yet unexplored, and we have no confidence in any absolute primi
still

tives for such investigation.

The
for the

calculus of reasoning, or universal calculus,

is

to be the instrument

development and manipulation


to get its complete generality

of systems in the universal language,

and

it is

from the fact that


of

all

science will be

expressed in the ideographic

symbols

that universal medium.

The

calculus will consist of the general principles of operating with such ideo14

See G. Phil, vn, 194, footnote.


G. Phil, vii, 84.

15

The Development of Symbolic Logic


graphic symbols:
16
"All

our reasoning

stituting of characters,
images."

is nothing but the relating and sub whether these characters be words or marks or

Thus while the

characteristica universalis

is

the project of the

logistic

treatment of science

in general, the universal calculus is the pre

cursor of symbolic logic.

The plan

for this universal calculus

changed considerably with the

development of Leibniz s thought, but he speaks of it always as a mathe matical procedure, and always as more general than existing mathematical methods. 17 The earliest form suggested for it is one in which the simple
concepts are to be represented by numbers, and the operations are to be

merely those of arithmetical multiplication, division, and factoring. When, later, he abandons this plan of procedure, he speaks of a general calculus which will be concerned with what we should nowadays describe as "types
with combinations which are absolute or relative, symmetrical or unsymmetrical, and so on. 1 * His latest studies toward such a calculus
of
order"

form the

earliest presentation of

what we now

call

the

"algebra

of

logic".

But

it

is

doubtful

if

Leibniz ever thought of the universal calculus as

restricted to our algebra of logic:

we can only say

that

it

was intended

to

be the science of mathematical and deductive form in general (it is doubtful whether induction was included), and such as to make possible the appli
cation of the analytic
scientific

method

of

mathematics to

all

subjects of

which

knowledge is possible. Of the various studies to this end our

chief interest will be in the early

essay,

in the fragments which attempt to an algebra of logic. 20 develop Leibniz wrote De Arte Combinatoria when he was, in his own words,

De Arte Combinatorial and

vix egressns ex Ephebis,

and before he had any considerable knowledge


tells

of

mathematics.
consent.

It

was published, he

us,

without his knowledge or

The
it

intention of the work, as indicated

by

its title, is
it.

to serve the

general art of rational invention, as the author conceived

As has been

mentioned,

seems that this end

is

to be accomplished

by a complete
is

analysis of concepts of the topic under investigation and a general survey


of the possibilities of their combination.

large portion of the essay

concerned with the calculation of the possible forms of this and that type
16

G. Phil, vii, 31.

17

18 19

See New Essays on the Human Understanding, Bk. iv, Chap, xvn, 9-13. See G. Phil, vn, 31, 198 jf., and 204. G. Phil, iv, 35-104. Also Gerhardt, Leibnizens mathematische Schriften (1859),
Scientia Generalis.
Characteristica,

V,

1-79.
20

xv-xx, G.

Phil.,

vn.

10

Survey of Symbolic Logic


etc.,

of logical construct:

the various dyadic, triadic,


of elements;

complexes which

can be formed with a given number


of the syllogism;

of the

moods and

figures

of the possible predicates of a given subject (the

com

the key to the predicates plexity of the subject as a concept being itself which can be analyzed out of it); of the number of propositions from a

given

number

of subjects, given
21
\

number

of predicate relations,

and given

number of quaestiones terms, and so on. In


of permutations to discover

of the variations of order with a given

number

of

fact so

much

space

is

occupied with the computation


of his contemporaries failed

and combinations that some

any more important meaning

of the essay,

and

it is

most

fre
22

quently referred to simply as a contribution to combinatorial analysis. Beyond this the significance of the essay lies in the attempt to devise a

symbolism which

will preserve the relation of

analyzable concepts to their

primitive constituents.

The

particular device selected for this purpose


is

representation of concepts
itself is of interest.

by numbers

unfortunate, but the attempt

Leibniz makes application of this method to geometry

and suggests

it

for other sciences. 23

In the geometrical illustration, the


Class
1

concepts are divided into classes.

consists of concepts or terms

regarded as elementary and not further analyzable, each of which is given a number. Thereafter, the number is the symbol of that concept. Class 2
consists of concepts analyzable into (definable in terms of) those of Class 1.

By

the use of a fractional notation, both the class to which a concept


its

belongs and

place in that class can be indicated at once.

nator indicates the

number

of the class

and the numerator

is

The denomi the number of


is

the concept in that class.


represented by 7/2.
those in Class
analysis of
1

Thus the concept numbered


2,

7 in Class 2

Class 3 consists of concepts definable in terms of

and Class
is

and

so on.

By

this

method, the complete


its

any concept

supposed to be indicated by

numerical symbol. 24

21 Leibniz tells us that he takes this problem from the Ars Magna of Raymond Lully. See G. Phil., v, 62. 22 See letter to Tschirnhaus, 1678, Gerhardt, Math., iv, 451-63. Cf. Cantor, Geschichte d. Math., m, 39 ff. 23 See the Synopsis, G. Phil, iv, 30-31.

24

See Couturat, op.

tit.,

appended Note

vi, p.

554 ff.

The concepts

are arranged as follows (G. Phil., iv, 70-72):

"Classis I; 1. Punctum, 2. Spatium, 3. intervallum, 4. adsitum seu contiguum, 5. dissitum seu distans, 6. Terminus seu quae distant, 7. Insitum, 8. inclusum (v.g. centrum est insitum circulo, inclusum peripheriae), 9. Pars, 10. Totum, 11. idem, 12. diversum, 13. unum, 14. Numerus, etc. etc. [There are twenty-seven numbered concepts in this class.] "Classis II; 1. Quantitas est 14 T&V 9 (15). [Numbers enclosed in parentheses have their usual arithmetical significance, except that (15) signifies an indefinite number .] 2. Indudens est 6.10. III. 1. Intervallum est 2.3.10. 2. Aequale A rijs ll.. 3. Continuum est A ad B, si TOV A j 9 est 4 et 7 B.; etc.
TU>

etc."

The Development of Symbolic Logic

11

In point of fact, the analysis (apart from any merely geometrical defects) Leibniz uses not only the inflected Greek falls far short of being complete.
article to indicate various relations of

concepts but also modal inflections


etc.

indicated

by

et, si,

quod,

quam faciunt,
its

In later years Leibniz never mentions this work without apologizing for
it,

yet he always insists that

main intention

is

sound.

This method

of assuming primitive ideas which are arbitrarily symbolized, of introducing

other concepts by definition in terms of these primitives and, at the same


time, substituting a single symbol for the complex of defining symbols
this
is,

in fact, the

method

of logistic in general.

Modern

logistic differs

from

this

logistic

two respects: (1) modern attempt of Leibniz most would insist that the relations whereby two or more concepts are
notably in
while Leibniz regards his set of primitive concepts as the

united in a definition should be analyzed precisely as the substantives are

analyzed;

(2)

necessary result of any proper analysis, modern logistic would look upon

them

as arbitrarily chosen.

Leibniz

s later

work looks toward the elimina

tion of this first difference, but the second represents a conviction from

which he never departed.

At a much

later date

come various

studies (not in Gerhardt),

w hich
r

attempt a more systematic use of number and of mathematical operations in logic. 25 Simple and primitive concepts, Leibniz now proposes, should be
symbolized by prime numbers, and the combination of two concepts (the qualification of one term by another) is to be represented by their product.

Thus

if

3 represent

"rational"

and 7

"animal",

"man"

will be 21.

No
a

prime number
animal.
cation:
"man",

will enter

more than once

into a given combination


is

rational rational animal, or a rational animal animal,

simply a rational

Thus

logical synthesis

is

represented by arithmetical multipli

logical analysis

by

resolution into prime factors.


its

The

analysis of
"rational",

21,

would be accomplished by finding


7.

prime factors,
s

3,

and

"animal",

In accordance with Leibniz

conviction that

all

knowledge

is

analytic
"All

and

all
P"

valid predicates are contained in the subject,


will

the proposition
the concept
25

is

be true

if

the

number which

represents

is

divisible

by that which represents P.

Accordingly the

Dated

April, 1679.
"Elementa

as follows:

Couturat (op. cit., p. 326, footnote) gives the titles of these Characteristicae Universalis (Collected manuscripts of Leibniz in

the Hanover Library, PHIL., v, 8 b); Calculi universalis Elementa (PHIL., v, 8 c); Calculi universalis investigations (PHIL., v, 8 d); Modus examinandi consequentias per numeros (PHIL., v, 8 e); Regulae ex quibus de bonitate consequentiarum formisque et modis syllogis-

morum

many

categoricum judicari potest per numeros (PHIL., v, 8f)." These fragments, with others, are contained in Couturat s Opuscules et fragments inedits de Leibniz.

12

Survey of Symbolic Logic

universal affirmative proposition

(where y

is

a whole number)

= Py may be symbolized by S/P = y or S By the plan of this notation, Py will represent


within the genus P,
is y.

some

species whose

"difference",

Similarly

Sx

Hence the particular affirmative, "Some will represent a species of S. S is P," may be symbolized by Sx = Py, or S/P = y x. Thus the uni versal is a special case of the particular, and the particular will always be
true

when the

universal

is

true.
this

There are several objections to

scheme.

In the

first

place,

it

the class as genus. presumes that any part of a class is a species within defensible on the ground This is far-fetched, but perhaps theoretically that any part which can be specified by the use of language may be treated
as a logical species.

worse defect

lies in

the fact that Sx

= Py

will

always be true.
will satisfy

For a given S and P, we can always find x and y which = P, the equation Sx = Py. If no other choice avails, let x

or

some multiple
"man-angel"

and y = S, or some multiple of S. although no men are angels. "Spineless


of P,

"Angel-man"
man"

"ra

tional

invertebrate",

but

it

is

false that

some men are invertebrates.

third difficulty arises because of the existential import of the particular a difficulty which later drew Leibniz s attention. If the particular affirma
tive
is

true,

then for some x and


4=

y,

Sx = Py.

The
But

universal negative should,

then, be Sx

Py.

And

since the universal affirmative

particular negative should be


practically
for all
will

4=

Py>

this

S = Py, the symbolism would be


is

unworkable because the inequations would have to be verified values of x and y. Also, as we have noted, the equality Sx = Py
4=

always hold and Sx

Py, where x and y are arbitrary, will never be


further,

true.

Such

difficulties led

Leibniz to complicate his symbolism

still

introducing negative numbers and finally using a pair of numbers, one But this scheme also breaks positive and one negative, for each concept. down, and the attempt to represent concepts by numbers is thereafter

abandoned.

in

Of more importance to symbolic logic are the later fragments included the plans for an encyclopedia which should collect and arrange all known
Leibniz cherished the

26 science as the proper foundation for future work.


26

G. Phil., vii, xvi-xx. Of these, xvi, without title, states rules for inference in terms of inclusion and exclusion; Difficultates quaedam logicae treats of subalternation and conversion and of the symbolic expression for various types of propositions; xvm, Specimen Calculi universalis with its addenda and marginal notes, gives the general prin ciples of procedure for the universal calculus; xix, with the title Non inelegans specimen

The Development of Symbolic Logic

13

notion that this should be developed in terms of the universal characteristic.


In these fragments, the relations of equivalence, inclusion, and qualification by another, or combination, are defined and used. These

of one concept

relations are always considered in intension

when

it is
is

ing the calculus to formal logic.


concepts, not simply of

"Equivalence"

a question of apply the equivalence of


"for

two

classes

which have the same members;

to include

or

to be included in
27

is

to affirm the predicate

universally

However, Leibniz evidently considers the calculus to have many applications, and he thinks out the relations and illustrates them frequently in terms of extensional diagrams, in which A, B, etc., are
of the subject
A".

represented by segments of a right

line.

Although he preferred to treat


If

logical relations in intension, he frequently states that relations of intension

are easily transformed into relations of extension.


in intension,

is

included in

is

included in

in extension;

and a calculus

may

be inter

preted indifferently as representing relations of concepts in intension or


relations of individuals

and

classes in extension.

Also, the inclusion rela

tion

may

be interpreted as the relation of an antecedent proposition to a

consequent proposition.
just as the subject

The hypothesis A
all

includes

includes the predicate B.~ B

its consequence B, This accords with his


is

frequently expressed conviction that


these studies are by no

demonstration

analysis.

means

to be confined to the logic of intension.

Thus As

one

title

suggests, they are studies demonstrandi in abstracts.

demonsirandi in abstractis struck out, and xx, without title, are deductive developments of theorems of symbolic logic, entirely comparable with later treatises.
place of symbolic logic in Leibniz s plans for the Encyclopedia is sufficiently by the various outlines which he has left. In one of these (G. Phil., vu, 49), divisions 1-6 are of an introductory nature, after which come

The

indicated

scientiarum instauratione, ubi de Systematibus et Repertoriis, et de Encyclo paedia demonstrativa codenda. Elementa veritatis aeternae, et de arte demonstrandi in omnibus disciplinis ut in Mathesi.
"7.

De

"8.

De novo quodam Calculo generali, cujus ope tollantur eos qui in ipsum consenserit; est Cabala sapientum. "10. De Arte Inveniendi.
"9.

omnes disputationes

inter

"11.

"12.

"13.

De Synthesi seu Arte combinatoria. De Analysi. De Combinatoria speciali, seu scientia formarum,

sive qualitatum in genere (de

Characterismis) sive de simili et dissimili. "14. De Analysi speciali seu scientia quantitatum in genere seu de "15. De Mathesi generali ex duabus praecedentibus composita."

magno

et parvo.

Then

cine, psychology, political science, economics, military science, jurisprudence, theology, in the order named.
27 28

various branches of mathematics, astronomy, physics, biological science, medi and natural

G. Phil, vii, 208.


"

"Generates

Inquisitiones

(1686): see Couturat, Opuscules

etc.,

pp. 356-99.

14
It is a

Survey of Symbolic Logic

failed to

frequent remark upon Leibniz s contributions to logic that he accomplish this or that, or erred in some respect, because he

chose the point of view of intension instead of that of extension.


facts are these:

The

Leibniz too hastily presumed a complete, or very close,

analogy between the various logical relations. It is a part of his sig nificance for us that he sought such high generalizations and believed in their validity. He preferred the point of view of intension, or connotation,
partly from habit and partly from rationalistic inclination.

As a conse

quence, wherever there

is

a discrepancy between the intensional and exis

tensional points of view, he

likely to overlook

it,

and to follow the former.

This led him into some


opposite inclination

difficulties

and choice

of example, but it also led

which he might have avoided by an him to make which has since been overlooked and
his

some

distinctions the importance of

to avoid certain difficulties into

which

commentators have

fallen. 29

In Difficultates quaedam logicae, Leibniz shows that at last he recognizes


the difficulty in connecting the universal and the corresponding particular.

He

sees also that this difficulty

is

connected with the disparity between the

intensional point of view


tions.

existential import of particular proposi In the course of this essay he formulates the symbolism for the four

and the

propositions in

two
All

different ways.

The

first

formulation

is:

30

Univ.

aff.;

is

B:
is

AB =
not B;

A, or

non-5 does not

exist.

Part, neg.;

Some

A
is

AB

4=

A, or
exist.

non-5

exists.

Univ. neg.;
Part,
aff.;

No A
Some A

B;

is

AB does not 5; AB exists.


men

AB = A
man"

and

AB

=}=

A may

be interpreted as relations of intension or of


are mortal, the intension of
"man",

extension indifferently.
is

If all

"mortal

the same as the intension of


is

and likewise the

class of

mortal

men

identical in extent

with the class of men.

The statements

concerning existence are obviously to be understood in extension only. The interpretation here put upon the propositions is identically that of

contemporary symbolic logic. With these expressions, Leibniz infers the subaltern and the converse of the subaltern, from a given universal, by
impossible,

For example, it led him to distinguish the merely non-existent from the absurd, or and the necessarily true from the contingent. See G. Phil, vn, 231, footn6te; and "Specimen certitudinis seu de conditionibus," Dutens, Leibnitii Opera, iv,
Part in, pp. 92
footnote.
30
.

29

,C

ff.,

also Couturat,

La Logique

de Leibniz, p. 348, footnote,

and

p. 353,

G. Phil, vii, 212.

The Development of Symbolic Logic

15

means

of the hypothesis that the subject,

A,

exists.

Later in the essay, he


31

gives another set of expressions for the four propositions:


All

is

B:
is

AB =
not B:

A.

Some A

AB

4=

A.
Ens.

No A

is

B:
is

Some A

B:

AB does not exist, or AB 4= AB AB exists, or AB = AB Ens.


"in

AB before the sign of equality represents the AB s or the AB the region of possible represents existing AB s, or actual members of the class AB. (Read AB Ens, which AB = AB Ens thus represents the fact that the class AB has members; AB AB Ens, that the class AB has no members. A
In the last two of these,
ideas";
"AB

Ens"

"AB

exists".)

=f=

logical species of the genus

A,

"some

A",

may

be represented by
"some

YA;
exist

YA Ens
ing
--

will represent existing

members

of that species, or
if

A".

Leibniz correctly reasons that

AB = A

(All

is

B),

YAB

A is B); but if AB 4= A, it does not follow that YAB 4= YA, YA. Again, if AB 4= AB Ens (No A is B), YAB B, YAB 4= YAB Ens (It is false that some A is B); but if AB = AB Ens (Some A is B), YAB = YAB Ens does not follow, because Y could assume values incompatible with A and B. For example, some men are wise, but it does
YA
if

(Some
=

for

=--

not follow that foolish

men
32

are foolish wise persons, because

"foolish"

is

incompatible with
division of

"wise".

The

distinction here between

AB,

a logical

A
s

or of B,

and

AB

Ens, existing

AB s,

is

ingenious.

This

is

our author
intension,

most

successful treatment of the relations of extension

and
are

and

of the particular to the universal.


"principles

In Specimen calculi universal, the announced as follows: 33


1)
"Whatever is

of the

calculus"

concluded in terms of certain variable letters

may

be

concluded in terms of any other letters which satisfy the same conditions; for example, since it is true that [all] ab is a, it will also be true that [all]
be
is

and that

[all]

bed

is be.

2)

"Transposing

coincides with ba,


3)
"Repetition
"One

letters in terms changes nothing; animal rational with rational animal

for
.

example ab

of a letter in the

same term

is

useless.

4) proposition can be made from any number by joining all the subjects in one subject and all the predicates in one predicate: Thus, a is b and c is d and e is /, become ace is bdf.
.

31 32

G. Phil,, vii, 213-14. G. Phil., vii, 215: the illustration


G. Phil, vii, 224-25.

is

mine.

33

16

A
5)
"From

Survey of Symbolic Logic

any proposition whose predicate is composed of more than one term, more than one proposition can be made; each derived proposition having the subject the same as the given proposition but in place of the
given predicate some part of the given predicate. u a is b and [all] a is c and [all] a is
d."

If [all]

is

bed,

then

[all]

If

we add
ab
is

to the
of

number

of these,

two

principles
(1)

which are announced


a
is

under the head

"self-evident

propositions"

included in a;

and

(2)

included in a

we have
logic.

here the most important of the funda


Principle 1 or of
is

mental principles of symbolic

usually qualified
of

by

some doctrine

of the

"universe

of

discourse"

"range

significance",

but some form of

it is is

indispensable to algorithms in general.


call

The law
the

numbered 2 above
3, the
"principle

what we now

the

"principle

of permutation";
5,

of

tautology";

4, the

"principle

of composition";

"principle

of

division".

And

the two

"self-evident propositions"

are often

included in sets of postulates for the algebra of logic.

There remain

for consideration the

two fragments which are given

in

translation in our Appendix,


acteristica.
sir andi

XIX

and

XX

of Scientia Generalis:

Char-

The

first

of these, with the title

Non

inelegans specimen
is

demoninter

in abstractis, stricken out in the manuscript,

rather the
or ab
is

more

esting.

Here the relation previously symbolized by

AB

represented

by A+B. And A+B = L signifies that A is contained or included in A scholium attached to the definition of this inclusion relation (est in} B.
distinguishes

from the part-whole relation. Comparison of this and other passages shows that Leibniz uses the inclusion relation to cover
it

(1)

the relation of a

member

of the class to the class itself;

(2)

the relation
(3)

of a species, or subclass, to its genus

a relation in extension;
a relation of intension.

the rela
first

tion of a genus to one of its species

The

of

these

is

our e-relation;
is

(2) is

the inclusion relation of the algebra of logic;

and

(3)

the analogous relation of intension.


it is

Throughout both these


theorems
in

fragments,

clear that Leibniz thinks out his

terms of

extensional diagrams, in which classes or concepts are represented by segments of a line, and only incidently in terms of the intension of concepts.

The

different interpretations of the


If

tinguished.

is

"rational"

and

is

symbols must be carefully dis "animal", and A and B are taken

m intension,

then

A+B

will represent

"rational animal".

But

if

and

are classes taken in

extension^ then

A B
+

is

the class

made up

of those

things which are either


34

or

(or both).

Thus the

inclusion relation,

4.

and

5.

ties of

universal affirmative propositions.

are stated without qualification because this study is confined to the proper 4. is true also for universal negatives.

The Development of Symbolic Logic

17

A +B =
in

L,

may
is

".

This

is be interpreted either in intension or in extension as little confusing to us, because we should nowadays invert
"A

the inclusion relation

when we pass from


meaning
+
of

intension to extension;

instead
B"

of this, Leibniz changes the

A +B

from
If

"both

and

(in

intension) to
mal",

"either

or

B"

(in extension).

A
If

is

"rational",

"ani

and

"man",

then

B = L

is

true in intension,
"man".

"rational animal"

"man"

or

"rational"

is

contained in

A, B, and

L are

classes

of points, or segments of a line, then


class of points

comprising the points

A + B = L will mean that L is the in A and the points in B (any points

common
segments

to

and

counted only once), or the segment made up of

and B.

The

relation
If

A +B
is

exclusive.

does not require that A and B should be mutually a line, A and B may be overlapping segments; and, in

intension,

and

B may

be overlapping concepts, such as

"triangle"

and

"equilateral",

Leibniz also introduces the relation

each of which contains the component "figure". L A, which he calls


that

detractio.

A =
+

N signifies
is

contains

and that
[ ]

if

be taken from

the
if

remainder

N.

The

relations [+] and

are not true inverses:

B =

L,

it

does not follow that


Leibniz
s

==

B, because
.

overlapping

(in

terms, communicantia)

If

A and B may be A = N, A and N


if

must be mutually exclusive (incommunicantia} A and B have a common part, M, L A = B


.

Hence

A+B

= L and

M.

(If

the reader will

take a
clear.)

line,

L, in which

and

are overlapping segments, this will be

This makes the relation of detractio somewhat confusing.

In

extension,
it
is

A may
and
calls

be interpreted

"L

which

is

not

".

In intension,
"rational"

more

difficult.

Leibniz offers the example:

"man"

=
is

"brute",

our attention to the fact that


or
"man"

"man"

"rational"

not

"non-rational man"

"non-rational".

35

In intension, the

relation seems to indicate an abstraction; not a negative qualification.

But there
"

are difficulties, due to the overlapping of concepts.


"woodworking"

Say that
pp. 377-78)

man"
35

r
"carpenter"

w^~

and

"man"

"white-skinned"

-^^
&

G. Phil., vn, 231, footnote.

Couturat in commenting on

this (op.

cit.,

says:
"Ailleurs

Leibniz essaie de preciser cette opposition en disant:

"Mais il oublie que le ne"ant Nihilum. Sed A non-A est Absurdwn. (non-Ens) n est pas autre chose que ce qu il appelle 1 absurde ou 1 impossible, c est-a-dire

est

le contradictoire."

Non-existence may It may be that Couturat, not Leibniz, is confused on this point. be contingent, as opposed to the necessary non-existence of the absurd. And the result of abstracting A from the concept A seems to leave merely non-Ens, not absurdity.
-

18

A
"Caucasian".

Survey of Symbolic Logic

Then

"

Caucasian

"

"

"

carpenter

"man"

"white-

skinned"
penter"

"woodworking".
"white-skinned",

Hence

("Caucasian"

"carpenter")

"car

because the

common

constituent
is,

"man"

has

been abstracted in abstracting


"carpenter"

"carpenter".

That
which

the abstraction of
"Caucasian"

from

"Caucasian

carpenter"

leaves, not
is

but

only that part of the concept


"carpenter".

"Caucasian"

wholly absent in
because
is
"man"

We

cannot here say

"white-skinned

man"

is

abstracted, nor
"man":

"white-skinned animal"

because

"animal"

contained

in

we can only say


is difficult

"white-skinned"

as a pure abstraction.

Such

abstraction

to carry out

and

of very little use as

an instrument

of logical analysis.

manuscript, and
his

theorems

in

Leibniz s illustration is scribbled in the margin of the seems clear that at this point he was not thinking out terms of intensions.
it

Fragment by
[

XX differs from XIX in that it lacks the relation


is

].

This

a gain rather than a


]

loss,
[ ]

and because [+ and XX is more carefully developed more of the simple theorems are proved, and more illustrations are given. Otherwise the definitions, relations, and
interpretation
:

symbolized both because of the difficulty of are not true inverses. Also

methods

of

proof are the same.

In both fragments the fundamental


is

operation by which theorems are proved


expressions.
If

the substitution of equivalent

the successors of Leibniz had retained the breadth of view which


7

characterizes his studies

and aimed to symbolize

relations of a like generality,

these fragments might well have proved sufficient foundation for a satis

factory calculus of logic.

III.

FROM LEIBNIZ TO DE MORGAN AND BOOLE


w ere made to develop a
r

After Leibniz, various attempts


logic.

calculus of

Segner, Jacques Bernoulli, Ploucquet, Tonnies, Lambert, Holland,

Castillon,

and

others, all

made

studies toward this end.

Of these, the

most important are those


of Holland s
in extension.
is

of Ploucquet,

Lambert and
it

Castillon, while one

of particular interest because


this

intends to be a calculus
success,

But
is

attempt was not quite a


either

and the net

result

of the others

to illustrate the fact that a consistent calculus of logical


is

relations in intension

most

difficult or quite impossible.


s

Of Segner
36

w ork and Ploucquet


r

36 copies of these writings are available.

we can give no account, since no Venn makes it clear that Ploucin this country,

secure

There seem to be no copies of Ploucquet s books them from the continent have so far failed.

and attempts to

The Development of Symbolic Logic

19

quet

calculus

was a calculus

of intension

and that

it

involved the quanti

fication of the predicate.

Lambert 37 wrote voluminously on the subject


important contribution to symbolic procedure
is

of logic, but his

most

contained in the Seeks

Versuche einer Zeichenkunst in der Vernunftlehre.* 8


separate studies,

These essays are not


later essays presuppose

made from

different beginnings;

those which precede and refer to their theorems; and yet the development
is

not entirely continuous.

Material given briefly in one

will

be found

set forth

more

at length in another.

And

discussion of

lems of the theory of knowledge and of scientific introduced. But the important results can be presented as a continuous

more general prob method are sometimes

development which follows

in general the order of the essays.


list

Lambert

gives the following

of his symbols:

The symbol

of equality (Gleichgultigkeit)

=
+

addition (Zmeizung)

abstraction (Absonderung)
opposition (des Gegentheils)
universality
particularity

copula
given concepts (Begriffe)
a, b, c, d, etc.

undetermined concepts

n,

m,

I,

etc.

unknowns
the genus
the difference

x, y, z.

7
5

developed entirely from the point of view of intension: the letters represent concepts, not classes, + indicates the union of two
calculus
is
[

The

concepts to form a third,

represents the withdrawal or abstraction of

some part
term

of the connotation of a concept, while the product of a

and

represents the

common

part of the two concepts.

"multiplied"

into them.

difference of a.

Much

use

is

qualify any Thus ay represents the genus of a, ad the made of the well-known law of formal logic
8

7 and

that the concept (of a given species) equals the genus plus the difference.
(1)
37

ay + ad = a(y +

6)

Johann Heinrich Lambert (1728-77), German physicist, mathematician, and astrono He is remembered chiefly for his development of the equation x n +px = q in an infinite series, and his proof, in 1761, of the irrationality of TT. 38 In Logische und philosophische Abhandlungen; ed. Joh. Bernoulli (Berlin, 1782),
mer.
vol.
i.

20

A
is

Survey of Symbolic Logic


a.

ay + a8

the definition or explanation (Erkldrung) of

As immediate

consequences of (1),
(2)

we have

also
(3)
]

ay = a
it

ad

ad
]

ay

Lambert takes
tions.
If

for granted that [+

and

are strictly inverse opera

We
+
b.

two concepts,
(a
b)

then

have already noted the difficulties of Leibniz on this point. a and b, have any part of their connotation in common, b will not be a but only that part of a which does not belong
"European"
"+

also to

If

and

"carpenter"

have the

common
is

part

"man",

then

("

European

"carpenter")
"man".

minus

"carpenter"
]

not

"European"

but

"European"

minus

And

[+

and

will

not here be true

inverses.

But this difficulty may be supposed to disappear where the terms of the sum are the genus and difference of some concept, since genus anddifference may be supposed to be mutually exclusive. We shall return
to this topic later.

More complex laws


fact that the genus of

of

genus and difference


is

may

be elicited from the


"explained,"

any given a

also a concept

and can be

as can also the difference of a.


(4)

a
Proof:

= a(y+

2
<5)

= ay

+ ayd + ady + ad 2
ad

ay = ayy + ayd
ad.

and

= ady +

add

But a = ay +
That
is

Hence Q.E.D.

to say:

giving

its

one wish to define or explain a, one need not stop at genus and difference, but may define the genus in terms of its
if

genus and

difference,

and define the difference

similarly.

Thus

is

equiva

lent to the genus of the genus of a plus the difference of the genus of a plus

the genus of the difference of a plus the difference of the difference of

a.

This

may

be called a

"higher"

definition or

"explanation"

of a.

Obviously, this process of higher and higher "explanation" may be carried to any length; the result is what Lambert calls his "Newtonian
formula".

We

shall best

understand this

if

we take one more preliminary

step. Suppose the explanation carried one degree further and the resulting terms arranged as follows:

3 a(7 + 776 + ydd +

6 )

+ 767 + dyd

+ dyy + ddy

The

three possible arrangements of

two y

s,

and one

might be summarized

The Development of Symbolic Logic

21

the three arrangements of two 6 s and one 7 by 8752. With this convention, the formula for an explanation carried to any degree, n, is:

by 3y

d;

This

"Newtonian formula" is

Two
(6)

a rather pleasant mathematical conceit. further interesting laws are given


:

a
Proof:

ad + ayd + ay 2 d + ay*d +

etc.

= ay +

ad

But
and

2 ay = ay + ayd 2 2 ay = ay* + ay d

ay*
(7)

= ay 4 + ay*d,
2

etc. etc.
. .

a
Proof:

= ay + ady + ad y + ad*y + a = ay + ad
ad

etc.

But
and

ad 2
ad*

= ady + ad = ad y + ad*
2 2

=
is

ad*y + ad*, etc. etc.

Just as the genus of a

represented by ay, the genus of the genus of

a by ay 2
n

etc., so a species of

which a

is

genus

may

be represented by ay~
2
,

l ,

and a species

ay

genus of the genus by ay~ etc. In general, as a. genus above a, so a species below a may be represented by represents
of
is

which a

ay~

a
or

yn
Similarly afny concept of which a
is

difference of the difference of the differ

ence

etc.,

may

be represented by
ad~ n
or

--

Also, just as a
tion",

a(y +

n d)
,

where a

is

a concept and a(y +

d)

its "explana

so-

a,

where

^n ls ^ ne concept and a

the

"explanation"

of

it.

Certain cautions in the transformation of expressions, both with respect


to
"multiplication"
39

and with respect to


5.

"division,"

need to be observed. 40

40

Seeks Versuche, p. Ibid.j pp. 9-10.

22

A
is

Survey of Symbolic Logic

The concept ay 2 + ady


(8)

very different from the concept (ay + ad)y, because

(ay + ad)y
2 7 + ady

= a(y+d)y = ay(y+d) = ay

while
of a.

is

the genus of the genus of a plus the genus of the difference


distinguished from
i.

Also

- y must be
7
which a
is

77
.

- y

is

the genus of any

species x of

the genus,

e.,

(9)

-7 =

But ay/y

is any species of which the genus of a is the genus, x such that a and x belong to the same genus. species

i.

e.,

any

We

turn

now
is

to consideration of the relation of concepts

which have a
and

common
in

part.

Similarity
so far as,

identity of properties.

Two

concepts are similar

if,

they comprehend

identical properties.

In respect to the

41 remaining properties, they are different.

ab represents the

common

properties of a and
a.

b.

a
a + b

ab represents the peculiar properties of


ab

ab represents the peculiar properties of a together with


b.

the peculiar properties of


It is evident

from

this last that

Lambert does not wish

to recognize in
b

his

system the law a + a a; else he need only have written a + If x and a are of the same genus, then

ab.

xy = ay
If

and

ax

= ay = xy
is

now we symbolize by a

b that part of a

which

different

from

42

6,

then

(10)

a\b + b\a + ab + ab

=
x

a+ b

Also

x
ax
a

x a

ay,

or

= ay + x\a

ad

ax

ad

ax + ad
a
ad

ax
41 42

= ay = xy
sometimes a
:

Ibid., p. 10.

Lambert sometimes uses a

b for this,

b.

The Development of Symbolic Logic

23

And

since

x
ax + a x ax

= ay

ax + x\a

a\x
a ax

x\a
x\a

a\x

=
is

ax

The
y

fact that y

a property comprehended in x

= xy

or

by y + x y

x.

The manner
first is

in

may be expressed by which Lambert deduces the


43

second of these expressions from the of x, then y x is null. But by (10),

interesting.

If

is

a property

2xy + x\y + y\x

x+y

Hence

in this case,

2xy + x y
\

=
=

x+y x +y

And

since y

xy,

2y + x\y

Hence

y+x\y =
This
is

He
2y

has subtracted y from both


y

sides, in the last step,

and we observe that


it

=
y>

rather characteristic of his procedure;

follows,

throughout, arithmetical analogies which are quite invalid for logic.

With the complications


little

of this calculus, the reader will

probably be

concerned.

There

is

no general type of procedure

for elimination or

solution.

Formulae

of solution for different types of equation are given.

They

are highly ingenious, often complicated,

and

of dubious application.

It is difficult to

of the development, so far as outlined, there

judge of possible applications because in the whole course is not a single illustration of a

solution which represents logical reasoning,

and very few

illustrations of

any kind.

The shortcomings

of this calculus are fairly obvious.

There

is

too

much
and

reliance

upon

the analogy between the logical relations symbolized

their arithmetical analogues.

Some

of the operations are logically

uninterpretable, as for

example the use of numerical coefficients other than


in the "Newtonian
formula",

and

1.

These have a meaning

but 2y either

has no meaning or requires a conventional treatment which is not given. And in any case, to subtract y from both sides of 2y = x + y and get y = x
represents no valid logical operation.
of the relations
b,
43

employed

is

lacking,

Any adequate study of the properties x = a + b is transformed into a = x


b

regardless of the fact that a


Seeks Versuche, p. 12.

and

may have

common

part and that

24
x

A
man =

Survey of Symbolic Logic

b represents the abstraction of the whole of b

from

x.
s

Suppose, for
procedure,

example, should have also rational


out this
difficulty,

rational + animal.

Then, by Lambert
animal.

we

= man

Since Leibniz had pointed

that addition and subtraction


it is

meanings) are not true inverses, should err in this.

(with exactly these the more inexcusable that Lambert

There

is

still

deeper difficulty here.


"

As Lambert himself remarks, 44

no two concepts are so completely dissimilar that they do not have a


part.
"

common
"

One might say that the concept


is

thing"

ing

common to every pair of concepts.


a<5;

word) or be This being the case, + and

(Lambert

are never really inverse operations.

Hence the
and a

difficulty will

not really
a8

disappear even in the case of ay and


will

ay =

ad, a

= ay

not be strictly valid.


"subtraction"

use of
of a

in a calculus

In fact this consideration vitiates altogether the based on intension. For the meaning
[ ]

ventional inverse of

becomes wholly doubtful unless be treated as a wholly con + ], and in that case it becomes wholly useless. The method by which Lambert treats the traditional syllogism is only
b
[

remotely connected with what precedes, and

its

value does not entirely

depend upon the general validity of his calculus.

He

reconstructs the

whole of Aristotelian logic by the quantification of the predicate. 45 The proposition "All A is has two cases:
B"

(1)

A =

B, the case in which

it

has a universal converse, the concept


particular, the concept

is

identical with the concept B.

(2)

>

B, the case in which the converse


in the

is

comprehended concept A. The particular affirmative similarly has two cases (1) A B, the case in which the converse is a universal, the subject comprehended within the predicate B.
:
<

(2)

The

case in which the converse

is

particular.

In this case the

comprehended within a subsumed species of the predicate and the predicate within a subsumed Lambert says species of the subject.
subject
is

this

may

be expressed by the pair:

mA
not

>

and

<

nB
must
<

Those who are more accustomed to

logical relations in extension

make the mistake


is
44

here of supposing that

A
is

>

mA, and

mA

A.

mA

a species of A, and in intension the genus


pp. 93
jf.

contained in the species,

Ibid., p. 12.

45

7m,

The Development of Symbolic Logic


not vice versa.
at first glance.
species

25

Hence

mA
in

>

does not give

>

B, as one might expect


"Some

We

see that

comprehended

by mA, a A, making the same assumption which occurs in


or portion of a class
is

Lambert here

translates

"

Leibniz, that

any subdivision

capable of being treated

as

some

species

comprehended under that

class as its genus.

Lambert says the subject and predicate each have peculiar properties by virtue of whose comprehension neither is contained in the other. But if the peculiar properties of the subject be taken away, then what remains is contained in the predicate; and if the peculiar properties of the predicate be taken away, then what
remains
is

In a universal negative proposition

contained in the subject.

Thus the universal negative

is

repre

sented by the pair

m
The
is

and

A>

B n

particular negative has

two

cases:
i.

(1) When it has not B but all B is

a universal affirmative converse,

e.,

when some

A.

This

is

expressed by
.4
<B

(2)

When

it

has not a universal affirmative converse.


is

In this case a

contained in the predicate, and a sub sumed species of the predicate in the subject.

subsumed

species of the subject

mA
Either of the signs,
terms.
<

>

B
>,

and

<

nB
by transposing the
"

and
>

may

be reversed
/,

And

if

<

Q,

P, then for some

P =

IQ.

Also,

multi

plication"

and

"division"

are strict inverses.

Hence we can transform

these expressions as follows:

>

is

equivalent to

A = mB nA = B
kB]
]

mA =
IA A_

= nB

or

pA = qB

_B
k
or

m
A =
"

B_

It is

evident from these transformations and from the prepositional equiva-

26
lents of the
"

A
inequalities"

Survey of Symbolic Logic


that the following
is

the

full

expression of these

equations:
(1)

A = mB:

All

is

and some

is

not A.

(2)
(3)

nA = B: Some A is not B and all B is A. mA = nB: Some, but not all, A is B, and some, but

not

all,

is

A.

(4)

-=-

No A

is

B.
that

The

first

noticeable defect here


(4)

is

A/m = B/n

is

transformable into

nA = mB and
fact,

can mean nothing different from (3). Lambert has, in four different propositions, if he sticks to the laws of his calculus: only

(1)
(2) (3) (4)

A = B: All A is all B. A = mB: All A is some B. nA == B: Some A is all B. mA = nB: Some A is some

B.

These are the four forms which become, in Hamilton s and De Morgan s A little scrutiny will show that treatises, the four forms of the affirmative.

Lambert
all, it is

treatment of negatives
necessary that
"

is

a failure.

For

it

to be consistent at

should not be transformed. But Lambert constantly makes such transformations, though he carefully re frains from doing so in the case of expressions like A/m = B/n which are supposed to represent universal negatives. His method further requires that m and n should behave like positive coefficients which are always
fractions"

greater than

and such that

4= n.

This

is

unfortunate.

It

makes

it

impossible to represent a simple proposition without


If

"entangling alliances".

he had taken a leaf from Leibniz

as affirmatives

book and treated negative propositions with negative predicates, he might have anticipated the
s

calculus of

De Morgan.

In symbolizing syllogisms, Lambert always uses A for the major term, for the middle term, and C for the minor. The perfectly general form of
is:

proposition

mA
P
Hence the

nB
(

1
46
:

perfectly general syllogism will be

AT
46

mA
p

nB
q

Major
Ibid., pp. 102-103. Ibid., p. 107.
47

The Development of Symbolic Logic

27

Minor

nC

vB

Conclusion

fj.n

C =

mv
pp

irq

The indeterminates
letters.

in the

minor are always represented thus by Greek


as follows:

The conclusion

is

delved from the premises

The major premise

gives

B =
np

A.

The minor

gives

B =
irv

C.

Hence
np

A-C.
irv

mv and therefore -- C = - A. irq pp


The above being the general form of the moods in the first figure is the following:
classification only so far as indicated

syllogism,
it

Lambert

scheme of

coincides with the traditional

by the use of the traditional names:

28

Survey of Symbolic Logic

nB = A
IV.
Fideleo

C
7T

X.
Pilosos

nB = A V.C = B

nC

= A

XL
Romano

nB = mA C = B nC = mA
nB = mA C = B
7ijjiC

XII.

Somnio

= mA

The
since

difficulty

about

"division"

it is

only required that

if

does not particularly affect this scheme, one of the premises involve fractions",
"

the conclusion must also.


cal in
lies in

It will

be noted that the


Lilii

mood

Hilario

is

identi

form with Romano, and


the fact that

with Somnio.

The reason

for this

nB =

mA

has two partial meanings, one affirmative


Hilario and
Lilii

and one negative

(see

above).

take the affirmative

interpretation, as their

the negative. indicate; Into the discussion of the other three figures, the reader will probably not care to go, since the manner of treatment is substantially the same as
in the above.

names

Romano and Somnio,

method

There are various other attempts to devise a convenient symbolism and for formal logic; 48 but these are of the same general type, and

they meet with about the same degree and kind of success. Two brief passages in which there is an anticipation of the logic of
relatives possess
attributes",

some

interest. 49

Relations,

Lambert

says, are

"

external

by which he means that they do not belong


(i.

to the object

an

"

sich.

Metaphysical"

e.,

non-logical) relations are represented

by

Greek

letters.

For example

if

=
/

fire,

h
: :

=
h

heat,

and a = cause,

The symbol
48

represents a relation which behaves like multiplication:


vi.

See in Seeks Versuche, v and

Also fragments
i,

"Uber

die Vernunftlehre", in

Logische und Philosophische Abhandlungen,


p.

xix and xx; and Anlage zur Architektonik,

190 jf.
49

ec/is

Versuche, pp. 19, 27 ff.

The Development of Symbolic Logic


h is in fact what Peirce and Schroder later a Lambert transforms the above equation into:
: :

29
"relative product".

called a

Fire
h

is

to heat as cause to effect.

Fire

is

to cause as heat to effect.

-*

<*

Heat

is

to

fire

as effect to cause.

The dot

here represents Wirkuny (it might be, Wirklichkeit, in consonance with the metaphysical interpretation, suggestive of Aristotle, which he It has the properties of 1, as is illustrated elsewhere 50 gives to Ursache).

by the fact that 7 may be replaced by this symbol. Lambert also uses powers of a relation.
If

=
<p

b,

and

=
<p

c,

=
<p

(p

(p-

And

if

=
<p~

c,

o
<p-

a c

j and

V7

a \|-

And more to the same effect. No use is made of this symbolism;


Lambert could have used
it.

indeed

it

is

difficult to see

how
felt

Yet

it is

interesting that he should have

that the powers of a relation ought to be logically important, and that he here hit upon exactly the concept by which the riddles of "mathematical
induction"

were later to be solved.


s

Holland
bert.
51

attempt at a
it

logical calculus

is

contained in a letter to

Lam

He

himself calls

an

"unripe thought",

and

in a letter

some three

52 years later he expresses a doubt

if

logic

is

really a purely formal discipline

capable of mathematical treatment.

But

this

study

is

of particular interest

because

it

treats the logical classes in extension

the only attempt at a

symbolic logic from the point of view of extension from the time of Leibniz
to the treatise of Solly in 1839.

Holland objects to Lambert


concepts by the relation of
50
61

method

of representing the relation of

lines,

one under the other, and argues that the

Ibid., p. 21.

52

Joha n. Lamberts deutscher Gelehrten Brief wechsel, Brief m, pp. 16 ff. See Ibid., Brief xxvn, pp. 259 ff.

30
"

A
"men"

Survey of Symbolic Logic

relation of

to

mortals" is

not sub but

inter.

He

is

apparently not

aware that

this

means exchanging the point

of view of intension for that

of extension, yet all his relations are consistently represented in extension,

as

we
(1)

shall see.
If

S represent the

subject,

the predicate; and p,

TT

signify
is

unde
a part a P.
is

termined variable

numbers, S/p =

P/TT will

come

to:

part of S

of P, or certain of the

are certain of the

P s,
all

or (at least) an

is

This expression

is

the general formula of

possible judgments, as

evident by the following:


(2)

A member
If

is

either positive or negative,


shall see in
is

and

in
IT

both

cases,

is

either

finite or infinite.
(3)

We

what fashion p and

can be understood.

p =
less

attains
it

its logical

can

still

all S, and in this way S/p cannot become less than 1, Since, then, p disappear and consequently cannot become negative.
1

in S/p, then

S/p as

many

as

maximum.
TT.

The same
(4)

is

true of

Therefore p and
TT

TT

cannot but be positive and cannot be

less

than

If

or

(5)

infinite, the concept becomes negative. If /expresses a finite number 1, then the possible forms of judg
>

becomes

ment

are as follows:
All

(1)

--=

^
--=

^
j

is all

P.

(2)

All

is

some P.

Now
an

expresses negatively

what

l/oo expresses positively.


is

To

say that

infinitely small part of a curved line


is

straight,

means

exactly:

No

part

of a curved line

straight.

(3)

~ =
1

All
CO

is

not P.

(4)
j-

-.

Some S
p

is all

P.

o
(5)

- =

Some S

is

some P.

(6)

S - = -P

Some S

is not-

P.

(7)
53

oo

=^ 1

All not-S

is all

P.

See

Ibid., Brief iv.

The Development of Symbolic Logic

31

S
(8)

P j
P

All not-S

is

some P.

S
(9)

All not-S

is all

not-P.

(1),

(2),

and
and

(9)
(8),

(3), (7), (6),

says are universal affirmative propositions; universal negatives; (4) and (5), particular affirmatives;

Holland

a particular negative.

As Venn has
Boole.
If

said, this notation anticipates, in a

way, the method of


of the

instead of the fraction


it,

we take the value

numerator

indicated by

the three values are

where
Boole
is

<

<

1,

and

/co

-.=

Q-S.

But the

differences between this

and
form

procedure are greater than the resemblances.


unfortunate in that
it

The

fractional

little

suggests that the equations


results

may

be cleared

of fractions,

and
s

this

would give

which are

logically uninterpretable.

But Holland
calculus.

notation can be

That he did

made the basis of a completely successful not make it such, is apparently due to the fact that

he did not give the matter sufficient attention to elaborate the extensional
point of view.

He

gives the following examples:

Example

1.

All All

men

H are mortal M
E are men H

Europeans

P
Ergo,

[All

Europeans are mortal]

Example

2.

All plants are organisms

P =

P
All plants are

no animals

oo

Ergo,

- =

A
-

[Some organisms are not animals]

32

A
Example
3.

Survey of Symbolic Logic

All

men

are rational

//

P
T>

All plants are not rational

P =

CO

Ergo, All plants are no

men

vH P=00

In this last example, Holland has evidently transformed //

= R/p

into

pH =
"Some

R, which

is

not legitimate, as
the rational
It

men

are

all

we have noted. pH = R would be And the conclusion P = pH/ao beings".


"All

is

also misinterpreted.

should be,

plants are not

some

men".

correct reading

would have revealed the invalid operation.

Lambert

replied vigorously to this letter, maintaining the superiority

method, pointing out, correctly, that Holland s calculus would not distinguish the merely non-existent from the impossible or
contradictory (no calculus in extension can), and objecting to the use of c in this connection. It is characteristic of their correspondence that each

of the intensional

pointed out the logical defects in the logical procedure of the other, and neither profited by the criticism.
Castillon s essay toward a calculus of logic
is

contained in a paper

presented to the Berlin

Academy in

1S03. 54

The

letters S,

concepts taken in intension,


synthesis"

A,

etc.,

represent

is

an indeterminate, S +
the withdrawal or

of

S and M, S

from

S.

M thus represents a genus concept in which S


"difference"

- M,

M represents the abstraction of M


is

subsumed,

being the logical

of

in

- M.

Consonantly S + M,
to S, represents a

symbolizing the addition of some


species concept

"further specification"

which contains

(in intension)

the concept S.
is

The

predicate of a universal affirmative proposition

contained in the

subject (in intension).

Thus

"All

is

A"

is

represented by

S =

A
is

M
+
is

The

universal negative

"No

S
-

is

A"

symbolized by

= (- A)

M
withdrawn
classes,
is

The concept S
lusory".

is

something,

M, from which A

no A.

Particular propositions are divided into

two

"real"

and

"il

real particular is the converse of a universal affirmative;


sur

the

"Memoire

un nouvel algorithme

logique",

in

Memoires de V Academic des Sciences


See also his paper,
"Reflexions

de Berlin, 1803, Classe de philosophie speculative, pp. 1-14. sur la Logique", loc. cit., 1802.

The Development of Symbolic Logic


illusory particular, one
lar affirmative is

33

whose converse

also

is

particular.

The

real particu

A = S since this
is
is

M
The
illusory particular affirmative

the converse of S

= A

M.

represented by

S =

A=?

M
A
which

Castillon s explanation of this is that the illusory particular judgment gives us to understand that some S alone is A, or that S is got from A by ab
straction (S

= A - M), when
==

abstraction (S

M + A).

in reality
this
it

it is

is

Thus

put +

M]

one can, then, indicate


is,

judgment puts by S = A ^ M.

M where

drawn from S by
it

should

The

fact

of course, that

"Some

the relations of the concepts S and

This means, in intension, that

if

indicates nothing about that they are not incompatible. except one or both be further specified in proper
is

"

might well be symbolized by S + N is really suspect that Castillon s choice of S = A =F governed by the consideration that S = A + be supposed to give may
fashion, the results will coincide.
It

= A

M.
^F

We

S =

the universal to give its subaltern, and that A will also give S = A =F M, that is to say, the real particular

M,

= S which

M
is

"All

is

S"

will also give


"All

S =
A"

=F

M.

Thus
"All

"Some

is

A"

may
is

be

derived both from

is

and from

is

S",

which

a de

sideratum.

The

illusory negative particular

is,

correspondingly,
=F

M
symbolism.

Immediate inference works out

fairly well in this

The S =
is

universal affirmative and the real particular are converses.

A
-is

M gives A

= S

M, and
S+
=

vice versa.

The

universal negative

directly convertible.

S =
ticular

A+M
J/

gives

A =
-

J/,

and

vice versa.

The
S

illusory

par

also convertible.

S =

A T

gives

=F

M.

Hence

A =

=F

comes back to

S = A

M, which

M.
subaltern

A
S

universal gives

its

S =
=

A
-

gives

S =

A
=

=F J/,
-

and
=T=

A+M

gives S

A
M,

M.

And a real A = S 4

particular gives also the converse illusory particular, for


gives

S =

34

A
its

Survey of Symbolic Logic

which gives which gives

subaltern,

S =

A ^ M,
figures of the syllogism

A =

M.

All the traditional

moods and

may

be symbolized

in this calculus, those

both for the real


All
All

which involve particular propositions being valid For example: particular and for the illusory particular.

Mis A

M
.

SisM
Sis
,4
.

All

= A+N S = M+P S = A + (N + P) = - A+N S = M+P


S
=

No
All

M
S

is is

M
:.

M
A

No Sis A
All

(N + P)

is

M
.
.

Some S is Some S is A
is

= A+N S = M =F P S = (A + N)

or
=F

S =
or

M-

P
+ N)

S = (A

- P

This

the most successful attempt at a calculus of logic in intension.

The
and
not
in

difficulty

about

"subtraction"

in the

Lambert
"

calculus, arises because


is

"

or

M which
If,

XIX Fragment of P does not mean


then
[

Leibniz,
"M

but

not

P
is

".

If it

mean

this,

and

are not

true inverses.

on the other hand,

M-

P indicates the

abstraction from

the concept
difficult or

of all that

involved in the concept P, then

-P

is

impossible to interpret, and, in addition, the idea of negation

cannot be represented by [-]. How does it happen, then, that Castillon s notation works out so well when he uses [-] both for abstraction and as
the sign of negation?
It

would seem that

his calculus

ought to involve

him

in

both kinds of
is

difficulties.

method

by good luck, hit upon a which nothing is ever added to or subtracted from a determined concept, S or A, except an indeterminate, or N or P, and this indeter
in

The answer

that Castillon has, apparently

M
]

minate, just because


are not true inverses.

it is

indeterminate, conceals the fact that [+

and [-]

And when

the sign

appears before a determinate,

A,

serve as the sign of negation, because no difficulty arises from supposing the whole of what is negated to be absent, or abstracted. Castillon s calculus is theoretically as unsound as Lambert s, or more
it

may

so

if unsoundness admits of It is quite possible that it was worked degree. out empirically and procedures which give invalid results avoided.

The Development of Symbolic Logic

35

Whoever studies Leibniz, Lambert and Castillon cannot fail to be con vinced that a consistent calculus of concepts in intension is either immensely difficult or, as Couturat has said, Its main difficulty is not impossible.
the one which troubled Leibniz
s system the failure This can be avoided by treating negative propositions as affirmatives with negative predicates, as Leibniz did. The more serious difficulty is that a
[ ]

Lambert

and which constitutes the main defect in of + and [-] to behave like true inverses.

calculus of

"

"concepts

is

not a calculus of things in actu but only in possibile,

and

in a rather loose sense of the latter at that.

Holland pointed this out

admirably

in a letter to

Lambert. 55

He

gives the example according to

Lambert

method,
All triangles are figures.
All quadrangles are figures.

T =

tF

Q = qF
tQ

Whence,

T =56

or

qT =

and he then proceeds:


"In

the conclusion nA = be drawn, the calculus cannot determine whether the ideas nA and
general,
if

from

==

mC

and

B = nC

mB mB

consist of contradictory partial-ideas, as in the foregoing example, or not.

The thing must be judged according


This example also
tributed middle term.
of the concept
calls

to the

matter."

attention to the fact that Lambert

calculus,

by operations which he continually


If
"some

uses, leads to the fallacy of the undis


is

A"

simply some further specification

A, then

this

mode

is

not fallacious.

And

this observation

brings

down

the whole treatment of logic as a calculus of concepts in in

tension like a house of cards.

The
more

relations of existent things cannot be

determined from the relations of concepts alone.

The

calculus of Leibniz

is

successful than
s is

continental successors

unless Ploucquet

any invented by his an exception. That the long


successful

period between

him and De Morgan and Boole did not produce a


is

system of symbolic logic


sional point of view.

probably due to the predilection for this inten-

It is

successful after the initial interest


logical relations in extension,

no accident that the English were so quickly was aroused; they habitually think of

and when they speak

of

"intension"

it

is

usually clear that they


"intension"
55 56

do not mean those

relations of concepts

which the

of traditional logic signifies.


I,

Ibid., pp.

Deutscher Gelehrter Briefwechsel, 262-63.

Brief xxvu.

36

A
The beginning
of

Survey of Symbolic Logic


this subject in

England is marked by the proposing some modification of the traditional logic by quantifying the predicate. As Sir William Hamilton 57 notes, the period from Locke to 1833 is singularly barren of any real con
thought upon
publication of

numerous

treatises, all

tributions to logic.

quantification of the predicate.


least as Leibniz.

About that time, Hamilton himself proposed the As we now know, this idea was as old at
Both Hamilton and
his student

Ploucquet, Lambert, Holland, and Castillon also had

quantified the predicate.


tion Ploucquet;

Thomson men

but this new burst of logical study in England impresses one as greatly concerned about its own innovations and sublimely indifferent
to its predecessors.

establish his

Hamilton quarrelled at length with De Morgan to 58 This is the more surprising, since priority in the matter.

George Bentham, in his Outline of a

New

System of Logic, published in 1827,


of propositions:

had quantified the predicate and given the following table


1
.

2. 3.

4.
5.

6.

7.

8.
is
|

X in toto = Y ex parte X in toto Y ex parte; X in toto = Y in toto; X in toto Y in toto; X ex parte = Y ex parte; X ex parte Y ex parte; X ex parte = Y in toto; X ex parte 7 in toto.
;
!
| | | | |

(|

here the sign of

"diversity").

in

But Hamilton was certainly the center and inspirer of a new movement logic, the tendency of which was toward more precise analysis of logical

significances.

Thought are
there

Bayne s Essay on the New Analytic and Thomson s Laws of the most considerable permanent record of the results, but
fervid discussion of logical topics in various peri

was a continual
was

odicals; logistic

in the air.

This movement produced nothing directly which belongs to the history of symbolic logic. Hamilton s rather cumbersome notation is not made the
basis of operations, but
is essentially only an abbreviation of language. scheme of representing syllogisms was superior as a calculus. But

Solly
57

See Discussions on Philosophy, pp. 119 ff. This controversy, begun in 1846, was continued for many years (see various articles in the London Athenceum, from 1860 to It was concluded in the pages of the Con 1867).
temporary Review, 1873.

The Development of Symbolic Logic


this

37

in fact
it

movement accomplished two things for symbolic logic: though not always in name the point of view of

it

emphasized

extension,

and

aroused interest in the problem of a newer and more precise logic. These may seem small, but whoever studies the history of logic in this period
will easily

never have been revived.


Boole.
in

convince himself that without these things, symbolic logic might Without Hamilton, we might not have had

record of symbolic logic on the continent is a record of failure, England, a record of success. The continental students habitually

The

emphasized intension; the English, extension.


IV.

DE MOP CAN
of symbolic logic only

De Morgan 59
permanent value
of certain
tions.

is

known

to

most students

through
of

the theorem which bears his name.

But he made other contributions


of
60
discourse",

the idea of the

"universe

the discovery

new types

of propositions,

and a beginning

of the logic of rela

Also, his originality in the invention of


illustrations,"

new

logical forms, his

ready

wit, his pat

and the

clarity

and

liveliness of his writing did


\

yeoman
of

service in breaking

down
in

the prejudice against the introduction

His important writings on logic logic. are comprised in the Formal Logic, the Syllabus of a Proposed System of
"mathematical"

methods

Logic,

and a

series of articles in the


1

Transactions of the Cambridge Philo

sophical Society?
59

Augustus De Morgan (1806-78), A.B. (Cambridge, 1827), Professor of Mathematics London 1828-31, reappointed 1835; writer of numerous mathematical treatises which are characterized by exceptional accuracy, originality and clearness. Per haps the most valuable of these is "Foundations of Algebra" (Camb. Phil. Trans., vu, vm); the best known, the Budget of Paradoxes. For a list of his papers, see the Royal For many years an active member of the Cambridge Philosophical Society Catalogue. Society and the Royal Astronomical Society. Father of William F. De Morgan, the novelist and poet. For a brief biography, see Monthly Notices of the Royal Astronomical Society,
in the University of
xii, 112.
60

The

idea

is

introduced with these words:


It
is

"Let

us take a pair of contrary names,

as

man and

not-man.

or real, in the universe.


universe, but

plain that between them they represent everything, imaginable But the contraries of common language embrace, not the whole

idea. Thus, of men, Briton and alien are contraries: The same may be said of be one of the two, no man can be both. integer and fraction among numbers, peer and commoner among subjects of a realm, male and female among animals, and so on. In order to express this, let us say that the whole idea under consideration is the universe (meaning merely the whole of which we are considering parts) and let names which have nothing in common, _but which between them contain the whole of the idea under ^nsideratiqn, be called contraries in, or with respect to,

some one general

every

man must

that

universe."
61

after

(Formal Logic, p. 37; see also Camb. Phil. Trans., vm, 380.) Formal Logic: or, The Calculus of Inference, Necessary and Probable, 1847. to be cited as F. L.

Here

38

Survey of Symbolic Logic

Although the work of De IN [organ is strictly contemporary with that of Boole, his methods and symbolism ally him rather more with his prede
cessors than with Boole

and those who

follow.

Like Hamilton, he

is

bent

upon improving the traditional Aristotelian logic. His first step in this direction is to enlarge the number of typical propositions by considering
all

the combinations and distributions of two terms,


It is a feature of

and F, and

their

negatives.

De Morgan

notation that the distribution of

affirmative or negative of the proposition are indicated, these being sufficient to determine completely the type of

each term, 62 and the quality


the proposition.

That a term
^before or after

A"

is

distributed

is

indicated

by writing

half a parenthesis
A"),

it,

with the horns turned toward the


is
7
)A".

letter, thus:

or (X.

An

undistributed term
A"(,

marked by turning the half-parenthesis the other


A"))}
,

way, thus: that

or

for example, indicates the proposition in

which the subject, X,


is,
"All
A"

is

distributed

and the predicate, F,


63

is

undistributed,

is

F".

X()Y

indicates a proposition with both terms un


is
F".

distributed, that

is,

"Some

The negative

of a term,
is

X,

is

indi

negative proposition placed between the parenthetical curves; thus "Some


.r;

cated by

of

F by

y, etc.

indicated

by

a dot
will

is

not

Y"

an affirmative proposition. All the different forms of proposition which De Morgan uses can be from two types, the universal, "All is generated ./ and the
A"(-(F.

be

64

Two

dots, or none, indicates

particular,
tive,
.r,

"Some

F and

y.

is by using the four terms, For the universals we have:


.

.,"

X and

its

nega

Syllabus of a Proposed System of Logic, 1860. Hereafter to be cited as Syll. Five papers (the first not numbered; various titles) in Camb. Phil. Trans., vm, ix, x. The articles contain the most valuable material, but they are ill-arranged and inter

spersed with inapposite discussion. Accordingly, the best get these articles and the Formal Logic, note in a general

way way

to study

De Morgan

is

to

the contents of each, and

then use the Syllabus as a point of departure for each item in which one is interested. 62 He does not speak of "distribution" but of terms which are "universally spoken
cr
"particularly

of"

spoken

of

",

or of the

"quantity"

of a term.

63 This is the notation of Syll. and of the articles, after the first, in Camb. Phil. Trans. For a table comparing the different symbolisms which he used, see Camb. Phil. Trans.,

ix, 91.
64

It is

sometimes hard to determine by the conventional

criteria

whether

De Morgan

propositions should be classed as affirmative or negative.


rule for distinguishing of and false of

them

X and

(Syll., p. 13):

"Let

gives the following ingenious a proposition be affirmative which is true

He

Thus Every A is Y is Some things are neither Some things are neither
things are neither

or x; negative, which is true of and x, false of and X. affirmative: is is true; is x is false. But Every Every
not-A"

A s nor F s A s nor A s
is

is is

also affirmative,
true,

A s nor z s

though in the form of a denial: though superfluous in expression; Some

false."

The Development of Symbolic Logic


(1) (2) (3) (4)

39

A))
x))y

MIX
All

is

Y.
is

All not-

A"

not- 7.

X))y

is

not-7.
7.

x))Y
:

Allnot-Zis

and

for particulars

we have
(5)

X()

Some

X X

is

7.
is

(6) (7) (8)

x()y

Some not-X

not- 7.

X()y
x()Y

Some
Some

is

not- 7.
is

not-JT

7.

The
its

rule for transforming a proposition into other equivalent forms may be stated as follows: Change the distribution of either term that is, turn

parenthetic curve the other way, change that term into its negative, and change the quality of the proposition. That this rule is valid will appear if we remember that "two negatives make an affirmative", and note

that we introduce one negative by changing the term, another by changing the quality of the proposition. That the distribution of the altered term
should be changed follows from the fact that whatever proposition distrib
utes a term leaves the negative of that term undistributed,

and whatever

proposition leaves a term undistributed distributes the negative of that

term.

Using this rule of transformation, we get the following table of equivalents for our eight propositions:
(a)
(1) (2)
(3)
(b)
(c)

(d)

XY
x))y

= X)-(y
=
-

--=x((y
=

x(-)Y

x)-(Y

X((Y = X(-)y
*()?
-

X))y

X).(Y = x((Y
a-)-Q/
=

(4) .r))7
(5)

X((y

X(-)Y

T()

7 = X(-(y
*x(.(Y
-.=

-=x)(y
=

=x).)Y

(G) x()i/
(7) (8)

XQy
a-()7

Z)(7-Z)-)y = x)(Y = x)-)y X(.(Y


-

=x(.(y

X}(y -=X)-)Y
is

It will

be observed that in each line there

one proposition with both


eight different types

terms positive,

A and

Y.

Selecting these,

we have the

of propositions:

40

A
(la)
(2c)

Survey of Symbolic Logic


is

A))F

All

A"

Y.
is all

A ((F
X)-(Y
X(-)Y
r

Some

F;

or, All

is

X.

(36)

NoZis
Some
Some
Some
All

Y.
is

(4d)

Everything

either

X or

Y.

(See below.)

(5o)
(6c)

A ()F
X)(Y
X(-(Y X)-)Y

A
A"

is

7.

things are neither


is

X nor

F.

(See below.)

(76)

not Y.
or,

(Sd)

is

not some Y;
is

Some 7
any one

is

not X.

Since the quantity of each term

indicated,

of these propositions

may

be read or written backwards

that
is

is,

w ith Y
r

subject
(4d)

and

X predicate X
is

provided the distribution of terms


cult to understand.
F",

preserved.

and

(6c) are diffi

We

might attempt to read X(-)Y

"Some

not

but we hardly get from that the difference between X(-)Y and some is not (any) A (-(F, "Some Also, A"(-)F is equivalent to uniis not some would make it par and the reading, "Some versals,

F".

F",

ticular.
"Xo

A"(-)F

is

equivalent to

cc))

F,

"All

not-Z

is

F",

and

to x)-(y,

not-X
is,

is

not-F".

The only equivalent


(in

of these

with the terms


is

A"

and F

"Everything

the universe of discourse)


likely to read

either

or

F
or

(or
"

both)".

(6c),

X)(Y, we should be
";

"All

is all

F",

X and

F are equivalent

but this would be an error, 65 since


(6a),

its

equivalents
is
not-F".

are

particular

propositions.
of this in

The equivalent
neither

terms of

xQy, X and

is

"Some

not-X

is

plainly,

"Some

things are

X nor

F".

Contradictories 66 of propositions in line


of those in line (2), in line (8);

(1) will

be found in line (7);

of line (3), in line (5); of line (4), in line (6).


:

We

give those with both terms positive


(la)
(2c)

X))Y
X((Y
A>(F

contradicts
"

(76)

X(-(Y

(Sd)
"

AF
X
",

(36)

(5a)
(6c)

XQY
X)(Y

X(-)Y
65

which it might seem that De Morgan himself has fallen. See e. g., is all F SylL, p. 25, and Camb. Phil. Trans., ix, 98, where he translates X)(Y by "All or "Any one X is any one Y But this belongs to another interpretation, the "cumular", which requires X and Y to be singular, and not-X and not- Y will then have common members. However, as we shall note later, there is a real difficulty.
error into
".

An

66

De Morgan
he

calls contradictory

propositions

"contraries"

(See F. L., p. 60; Sytt.,


"contraries".

p. 11), just as

calls

terms which are negatives of one another

The Development of Symbolic Logic

41

Thus the
of terms

rule

is

that two propositions having the same terms contradict


is

one another when one


is

affirmative, the other negative,

and the distribution

exactly opposite in the


for transforming

two

cases.

The

rule

propositions which has been stated and

exemplified, together with the observation that

may

any symbolized proposition be read or written backwards, provided the distribution of the terms
from universals, particulars from particulars.
"Each

be preserved, gives us the principles for the immediate inference of universals

For the

rest,

we have
67

the rule,

universal affirms the particulars of the

same

quality".

For

syllogistic reasoning, the test of validity


6i

and

rule of inference are

as follows:

"There is inference: 1. When both the premises are universal; 2. When, one premise only being particular, the middle term has different quantities

in the

two premises.
conclusion
is

"The

[parenthetic

curves]."

found by erasing the middle term and its quantities This rule of inference is stated for the special

arrangement of the syllogism in which the minor premise is put first, and the minor term first in the premise, the major term being the last in the
second premise.
Since

any proposition may be written backward,


According to the rule,
"No

this

arrangement can always be made.


is
F",

X))Y,

"All

X
is

is

F",

and F) (Z, "No Y is and F(-(Z, "Some F


things are neither
A"

Z",

give X)-(Z,

X
-

is

Z".

A>(F,

"No

is

not

Z",

give

A")

(Z, or

A ) (Z, which

"Some

nor

Z."

The

reader may, by inventing other


is

examples, satisfy himself that the rule given


reasoning, with

sufficient for all syllogistic

any

of

De Morgan

eight forms of propositions.

De Morgan

also invents certain


in a fashion

pound syllogisms
"1.

compound propositions which give com somewhat analogous to the preceding: 69

X)0)Forboth A))Fand

AF

All

Xs

and some things be

sides are
2.

s.

For

both

A ))Fand A ((F
X((Y and X(-(Y

All

3.

X(0 (F
)

or both

X s are F and X Among X s are all the


s,

all

are

s.

and

some things
4.

besides.

A O (For

both X)-(Y and

X)(Y

Nothing both some things

and

F and

neither.

67

Sytt., p. 16.

**Sytt.,p. 19. 69 SyU., p. 22.

42

A
5.

Survey of Symbolic Logic

Z|- For both


|

A>(Fand

X(-)Y

and F and Nothing both everything one or the other.


Everything either

6.

X(O)Y

or both

X(-)Y and

XQY

or

Y and

some things

both."

Each
sign

of these propositions

may, with due regard

for the

meaning

of the

O, be read or written backward,

just as the simple propositions.


is

The

rule of transformation into other equivalent forms

slightly different:

Change the quantity,


by
its

or distribution, of

any term and replace that term

negative.

We

are not required, as with the simple propositions, to

change at the same time the quality of the proposition. This difference is due to the manner in which the propositions are compounded.

The

rules for mediate, or


70

"

syllogistic",

inference for these

compound
or 6,

propositions are as follows:


"If

any two be joined, each of which is [of the form of] with the middle term of different quantities, these premises
clusion of the

1, 3, 4,

yield a con

same kind, obtained by erasing the symbols of the middle term and one of the symbols [O]. Thus X)O(Y(O)Z gives X)O)Z: or
if

nothing be both

X and

7 and some
things both,

things neither, and


it

if

everything be
tivo lots

either

other
"

Z and some things are Z s.


Y
or
in

follows that

all

and

of

In any one of these syllogisms,


or

it

follows that

may

be written for

one place, without any alteration of the conclusion, except )O( )O) reducing the two lots to one. But if this be done in both places, the con clusion is reduced to or and both lots disappear. Let the reader
|

examine

for himself the cases in

which one

of the premises

is

cut

down

to a

simple universal.
"The Letters written following exercises will exemplify what precedes. under one another are names of the same object. Here is a universe of 12

instances of which 3 are

and the remainder


the remainder

P s;

5 are

and the

remainder Q

s;

7 are

Z s and

R s.

XXX PP PP PPPPP
YYY YY QQ QQQQQ Z Z Z Z Z Z Z RRRRR
We
can thus verify the eight complex syllogisms

X)0)Y)0)Z P(0)Y)Q(R
70

P(0)Y)0)Z

P(0(Q(O)Z

X)0)Y)0(R

X)Q(Q(0(R

P(O(Q(O(R X)O(Q(O)Z

Sytt., p. 23.

The Development of Symbolic Logic


In every case
it

45f

be seen that the two lots in the middle form the quantity of the particular proposition of the conclusion."
will

In so
too

much much tied

of his

work

as

we have thus

far reviewed,

to his starting point in Aristotelian

De Morgan is still He somewhat logic.

simplifies traditional

methods and makes new generalizations which include


distinctly the old logic.

old rules, but

it is still

He

does not question the

inference from universals to particulars nor observe the problems there involved. 71 He does not seek a method by which any number of terms

may

be dealt with but accepts the limitation to the traditional two. And his symbolism has several defects. The dot introduced between the
parenthetic curves
is

read ()

as,

"It

is

not the sign of negation, so as to false that The negative of ()


()".

make
is )

it

possible to

(,

so that this

simplest of all relations of propositions

is

represented by a complex trans


in the

formation applicable only when no more than two terms are involved
prepositional
relation.

Also, there are two distinct senses in which a

term

in a proposition

may

be distributed or

"mentioned universally",

and

De Morgan,

following the scholastic tradition,

fails

to distinguish them and

symbolizes both the same way.

This
is

is

the secret of the difficulty in reading

X)(Y, which looks


neither
A"

like

"All

A"

all

7",

and

really

is

"Some

things are

Mathematical symbols are introduced but without any mathematical operations. The sign of equality is used both corresponding for the symmetrical relation of equivalent propositions and for the un".

nor

72

73 symmetrical relation of premises to their conclusion. His investigation of the logic of relations, however,

is

more

successful,

and he
is

This topic suggested to him by consideration of the formal and material elements
laid the
field.

foundation for later researches in that

in logic.
71

He

says:

7l

But he does make the assumption upon which


is

particular from a universal

simple term has members. ideas, is tacitly claimed for the terms of every syllogism". 72 A universal affirmative distributes its subject in the sense that it indicates the class to which every member of the subject belongs, i. e., the class denoted by the predicate.
Similarly, the universal negative,

all inference (in extension) of a necessarily based: the assumption that a class denoted by a He says (F. L., pp. 110), "Existence as objects, or existence as

not-A

is F, indicates that every is not- Y, every Y is No The particular nega particular proposition distributes a term in that sense. tive tells us only that the predicate is excluded from some unspecified portion of the class and Y in this sense only. Comparison with denoted by the subject. A)(F distributes
.

No

its

equivalents shows us that

it

can

tell us,

of

A",

only that

it is

specified portion of not-F; and of Y, only that it is excluded from of not-A. is wholly included in Y, or cannot infer that

We

excluded from some un some unspecified portion Y in X, or get any other

relation of inclusion out of


73

it.

In one passage (Camb. Phil. Trans., x, 183) he suggests that the relation of two C. premises to their conclusion should be symbolized by A B
<

74

Camb.

Phil. Trans., x, 177, footnote.

44
there

A
"Is

Survey of Symbolic Logic

any consequence without form?


Is
is

of the

machinery?

Is not consequence an action not logic the science of the action of the machinery?
act of the

Consequence
to ask,
on,

always an

mind
the

What

kind of act?

What

is

act,

on every consequence logic ought as distinguished from the acted


?

and from any

inessential concomitants of the action

For these are

of

the form, as distinguished from the matter.


".

The copula performs


. . .

certain functions;

it is

competent to those

functions
validate

because
. . .

it

has certain properties, which are sufficient to


is,

its use.
it

The word

which

identifies,

does not do
is

its

work

because

identifies,

except insofar as identification


is

a transitive and
;

convertible

motion:
is

that which is

B means A
is
,

is

and

A
is

is

B
fit

means

Hence every

transitive

and convertible relation

as

and by the same proof in each case. Some forms are valid when the relation is only transitive and not Thus if Y represent convertible; as in give and 7 connected
.

to validate the syllogism as the copula

by

a transitive copula, Camestres in the second figure

is

valid, as in

EveryZ
... In the

7,

No

X
:

Y,

therefore

No
is

X~Z.

following chain of propositions, there

exclusion of matter,

form being preserved at every step

Hypothesis
(Positively true)

Every man Every man Every


is

is is

animal

Y
a
/?

has existence.

X 7 Every X 7 a of X 7
(Probability
(3)

X has existence.
is

a transitive relation.
<

is is

a fraction

or or

1.

a of .Y

a fraction

<

1.

The

last is nearly the


it,

purely formal judgment, with not a single material

point about
"...

except the transit! veness of the copula. 75 I hold the supreme form of the syllogism of one middle term to
is

be as follows: There
there
is

the probability a that

is

in the relation

the probability

the probability a /3
"...

M to Z; that X in the relation L of M to Z.


that
is

to 7;
is

is

in the relation

whence there

76

The copula
of the

post hoc

is

and effect, of motive and action, of all which form and propter hoc (perhaps) of the matter, will one day
of cause
77
logic."

be carefully considered in a more complete system of


75 76

Ibid., pp. 177-78.


Ibid., p. 339.

77

Ibid., pp. 179-80.

The Development of Symbolic Logic

45

De Morgan
general,

is

and to consideration

thus led to a study of the categories of exact thinking in of the types and properties of relations.

His division of categories into logico-mathematical, logico-physical, logico78 metaphysical, and logico-contraphysical,

is

inauspicious,

and nothing
"The

much comes

of

it.

But
it,

in

connection with

this,

and an attempt to rebuild


con
is

logic in the light of

he propounds the well-known theorem:


[logical

trary [negative] of an aggregate

sum]

the

compound

[logical

product] of the contraries of the aggregants: the contrary of a is the aggregate of the contraries of the components." 79

compound

For the
L, of

logic of relations,

X, Y, and

will

represent the class names;


is

M, N,
.

relations.

LY will
is

signify that

A"

some one

of the objects

thought which stand to Y

in the relation L, or is

one of the
. .

of 7. 80

X LY will signify that X not any one of the L s of Y. X (LM) Y or X LM Y will express the fact that AT is one of the L s of one of the M s of y, or that X has the relation L to some Z which has the relation M to Y. X LM Y will mean that X is not an L of any M of Y.
.
.

It should be

noted that the union of the two relations


"

L and
.

M
. .

is
.

what
.

we should
If

"

call

today
.
.

their

relative product
.

that
. .

is,

A"

together give A^

LM Z, but A
"brother
of"

LY and A

is

the relation
"

and

MY do not give
"aunt

is

the relation
Y".

will

mean

is

a brother of an aunt of
of

(Do not

MZ LM Y. A LM Y of say hastily, X
LY and Y
A"

",

"

is

uncle of

7".

"Brother

an

aunt"

is

not equivalent to
written by

"uncle"

since

some uncles have no


that

sisters.)

L, or

M,

itself, will

represent

which has the relation L, or

M,

that

is,

a brother, or an aunt, and


to F, that
is,

LY

stands for any

X which

has the relation

a brother of T. 81

In order to reduce ordinary syllogisms to the form in which the copula

has that abstractness which he seeks, that

is,

to the form in
it is

which the
necessary

copula

may

be any relation, or any relation of a certain type,

to introduce symbols of quantity.

L of Accordingly every M, that is, something which has the relation L to every member of L*M is to indicate an L of none the class (say, a lover of every man). but The mark of quantity, * or *, always s (a lover of none but men).
is

LM*

to signify an

M
78 79

See

ibid., p. 190.

Ibid., p. 208.
"On

See also

Sytt., p. 41.
3,

Pp. 39-60 of Syll. present in


in
General."

summary

the ideas

of the paper,
80
81

the Syllogism, No.

and on Logic

Camb.

Phil. Trans., x, 341.

We

1 tried at first to

make De Morgan

follow the order of the paper from this point on. s symbolism more readily intelligible by intro

ducing the current equivalents of his characters. But his systematic ambiguities, such as the use of the same letter for the relation and for that which has the relation, made this impossible. For typographical reasons, I use the asterisk where he has a small accent.

46

Survey of Symbolic Logic

goes with the letter which precedes


the letter which follows.
that
it is

it,

but

L*M
L
of

is

read as

if

*]

modified
suggests

To

obviate this difficulty,

De Morgan
"An

L*M
an Z,

be read,
"

"An

every-!, of If;

an

M in every way in which


L
of

but we

shall stick to the simpler reading,

none but

s".

LM*X means an L of every M of X: L*MX, an L of none but M s of X:


L*M*, an L
of every

M and
1

of

none but 3f

s:

LMX*,
of

an

of

an

M of
.

every X, and so on.

If

Two more L is
".

symbols are needed.


of",

The converse
of";

L is

l symbolized by L~

"lover

L-

is "beloved

if

is

"aunt",

L-

is

"niece

or

nephew
ized

The contrary

(or as

we should

say, the negative) of

L is symbol
:

by m. by 1; the contrary of In terms of these relations, the following theorems can be stated
Contraries of converses are themselves contraries.

(1) (2) (3)

Converses of contraries are contraries.

The contrary
If

of the converse

is

the converse of the contrary.


(a)
(6)

(4)

the relation

L
l
,

be contained
is

converse of L,

L~

in, or imply, the relation M, then contained in the converse of M, M~ l and


\

the
the

contrary of

M, m,

is

contained in the contrary of L,


of"

I.

For example,

if "parent

is

contained in

"ancestor

of",

(a)

"child

of"

is contained in "descendent
"not

of",

and

(b)

"not

ancestor

of"

is

contained in

parent

of".

compound relation is accomplished by converting both components and inverting their order; thus, (LM) 1 = M~ l L~ l
(5)
.

The conversion

of a

If

X be teacher of the

child of Y,
is

When

a sign of quantity

parent of the pupil of X. involved in the conversion of a compound


is

relation, the sign of quantity changes

its

place on the letter; thus,

If

X be

teacher of every child of F,


in a

is

parent of none but pupils of X.


is

(6)

When,

compound

relation, there
its

a sign of quantity,

if

each

component be changed into


from one component to
the resulting relation
is

contrary, and the sign of quantity be shifted the other and its position on the letter changed,

equivalent to the original;

thus

LM* =

l*m and

L*M =
A

lm*.

lover of every

man

is

a non-lover of none but non-men; and a lover

of none but

men

is

a non-lover of everv non-man.

The Development of Symbolic Logic


(7)
is

47

When

compound

relation involves

no sign of quantity,

its

contrary

found by taking the contrary of either component and giving quantity


"

to the other.
"Not

The contrary of LM is IM* or L*m. is "non-lover of every man" (lover of a man) but non-men"; and there are two equivalents, by
((>).

or

"lover

of

none

But

if

there be a sign of quantity in one component, the contrary

is

taken by dropping that sign and taking the contrary of the other component.

The contrary
"Not

of

LM*
"

is

IM;

of

L*M

is

Lm.
of a
man";

(lover of every

man)" is "non-lover
".

and

"not

(lover

of

none but men) is "lover of a non-man So far as they do not involve quantifications, these theorems are familiar though
seems not generally known that they are due to The following table contains all of them:
it

to us today,

De Morgan.
Combination

Converse

Contrary
l

Converse of Contrary Contrary of Converse

LM

M- L~
l

/J/* or
1

L*M

J/*-

/- 1

or
l

m~ Ll

LM*mhm
L*Morlm*
The

M*~ L~
l

or
or

M^L-

m~Hm*- l~
}

*
l

IM

M~
l

l~ l
l

Lm
is

m~ L~
"contained
in"

sense in which one relation

said to be

or to

"imply"

another should be noted:

is

contained in

which has the relation L to any Y has

also the relation

M in case every X M to that Y. This


between two
rela
is
is

must not be confused with the


tive terms.

relation of class inclusion


is

Every grandfather

also a father, the class of grandfathers


"grandfather
of"

contained in the class of fathers, but


"father
of",

not contained in

because the grandfather of


of"

is

not also the father of Y.


of",

The
uses
"All

relation "grandfather

is

contained in

"ancestor

since the grand

father of

is

also the ancestor of Y.


"

But De Morgan appropriately


M"

the same symbol for the relation

contained in
is,

that he uses for

is

M",

where

L and

M are class terms, that


of

L))M.

In terms of this relation of relations, the following theorems can be


stated
(8)
:

If

L))M, then the contrary

is

contained in the contrary of L,

that

is,

L))M

gives m))l.

Applying
(8
-)

this

theorem to compound

relations,

we have:

LM))N
If

gives w))/J/* and n))L*m.


l 1

(8")

LM))N, then L~ n))m and wJ/-

))/.
l
l
.

Proof: If

LM))N, then n))lM*.

Whence nM- ))lM*M~

But an

of

48

Survey of Symbolic Logic


be an
l

every

M of an

M~

of

Z must

of Z.
l

Hence

nM~ ))l.
l
.

Again;

if

LM))N,
l

then ri))L*m.

relation converse-of-Z to

Whence L~ n))L- L*m. But whatever has the an L of none but m s must be itself an m. Hence
K"

L~ n))m. De Morgan
(9)
If

calls this "theorem

from
1

its

use in Baroko and Bokardo.

LM

= N,

then

Proof: If
for

any X,

LM MM~ X
1 1

L))NM~ and M))L~ N. = N, then LMM~ = NM~ and


l l
l

L^LM = Z~W. Now


X; hence
the

and L~

LX
l

are classes which contain

theorem.

We

do not have

L =
==

NM~

and
l

M
--=

==

L~ N, because
1
:

it is

not generally
of the of the

MM~ X X and Lr LX X. of X may not be X but brother parent


true that
A"

For example, the child but the class children


"

parent of
then

X"

will contain
l

X.
1

The

relation

MM~

or

cancel out.

MM~
1 1

and

M~ M are always symmetrical relations


1

M~ M will not always


1
;

if

XMM~

YMM~ X. If X is child of a parent of Y, then Y is child of a parent of X. But MM- and M~ M are nctf exclusively reflexive. XMM^X does

If we know that a child of the parent of X is a celebrated we may not hastily assume that X is the linguist in question. linguist With reference to transitive relations, we may quote 82

not always hold.

"A

relation

is

transitive

when a

relative of a relative

is

a relative of
so on.

the same kind; as symbolized in ZZ))Z, whence


"A

ZZZ))ZZ))Z; and

transitive relation has a transitive converse, but not necessarily a transitive contrary: for L~ l L~ l is the converse of LL, so that ZZ))Z gives

L L
1

its

From these, by contraposition, and also by theorem contrapositions, we obtain the following results:
*))L
l
.

and

L
I

is

contained in

LL-

*,

Z*H, HZ*, L*~

ZZ"

1
,

Z-!Z ...... Z- 1

omit demonstration, but to prevent any doubt about correctness of

printing, I subjoin instances in words:


82

Z signifies ancestor and L~

descendent.

Camb.

Phil. Trans., x, 346.

For

treats all reciprocal relations, such as

XLL^Y,

this discussion of transitive relations, De Morgan as also reflexive, though not necessarily

exclusively reflexive.

The Development of Symbolic Logic


"An

49

ancestor

is

always an ancestor of

all

descendents, a non-ancestor

none but non-descendents, a non-descendent of all non-ancestors, and a descendent of none but ancestors. A descendent is always an ancestor of
of

none but descendents, a non-ancestor of


ancestor

all non-descendents, a non-descend ent of none but non-ancestors, and a descendent of all ancestors. A nonis always a non-ancestor of all ancestors, and an ancestor of none but non-ancestors. A non-descendent is a descendent of none but non-

descendents,

and a non-descendent
all

of

all

descendents.

Among
all

nonnonall

ancestors are contained


ancestors
of

descendents of non-ancestors, and

descendents.

Among
all

non-descendents

are

contained

ancestors of non-descendents, and

non-descendents of

ancestors."

In terms of the general relation, L, or


syllogisms of traditional logic

M,

representing any relation, the


83

may

be tabulated as follows:

The Roman numerals here


clusion, there are

indicate the traditional figures.

All the con

clusions are given in the affirmative form;

but for each affirmative con


relation
.

two negative conclusions, got by negating the


its
1

and
l

replacing
.

it

by one or the other of


.

contraries.

A IH*Z and A L*mZ;


and so on
63

1M~ Z

gives

LMZ gives Thus LM~ *Z and A hm- Z,


A"

for

each of the others.

Ibid., p. 350.

50

A
When
the copula of
all

Survey of Symbolic Logic


three propositions
is

limited to the same transitive


84

relation, L, or its converse, the table of syllogisms will be:

X..LY
I

X.LY
Y..L-1Z

X..LY
Y.L~ Z X L~ Z
1

Y..LZ

X ..LZ

X .LZ

Y ..LX
IV

Y .LX

Y ..LX
1

Z.L- Y
1

X.LZ
Here, again, in the logic of relations,

Z..L~ Y X.L~ Z
1

Z..LY X..L~

1
Z"

De Morgan would

very likely have


itself.

done better

if

he had

left

the traditional syllogism to shift for

The

introduction of quantifications and the systematic ambiguity of L, M, etc., which are used to indicate both the relation and that which has the

hurry him into complications before the simple analysis of rela tions, and types of relations, is ready for them. This logic of relations was
relation,

destined to find

its

importance

in the logistic of
of,

mathematics, not

in

any

applications to, or modifications


cations of

Aristotelian logic.

And

these compli

De Morgan

s,

due largely to

his following the clues of formal logic,

had to be discarded
Boole
s

later, after Peirce

discovered the connection between


logic of relative

algebra and

relation theory.

The

terms has been

reintroduced by the work of Frege and Peano, and more especially of Whitehead and Russell, in the logistic development of mathematics. But
it is

there separated

and has to be separated


Nevertheless,
it

of the relations themselves.

from the simpler analysis should always be remembered

was De Morgan who laid the foundation; and if some part of his work had to be discarded, still his contribution was indispensable and of
that
it

permanent value.
84

In concluding his paper on relations, he justly remarks

85
:

Ibid., p. 354. Ibid., p. 358.

85

The Development of Symbolic Logic


"And

51

here the general idea of relation emerges, and for the first time in the history of knowledge, the notions of relation and relation of relation
are symbolized.

And

here again
is

is

seen the scale of graduation of forms,


is

the

difference of form at one step of the ascent difference of matter at the next. But the relation of algebra to the

manner

in

which what

higher

developments of logic
It will hereafter
it

is

a subject of far too great extent to be treated here.

be acknowledged that, though the geometer did not think


et

necessary to throw his ever-recurring principium

tion of

Omnis homo

est

animal, Sortes

est

homo,

etc.,

exemplum into imita yet the algebraist was

living in the higher


relation, before it

atmosphere of syllogism, the unceasing composition of was admitted that such an atmosphere existed." M
V.

BOOLE

The beginning from which symbolic logic has had a continuous develop 87 is that made by His significant and vital contribution George Boole. was the introduction, in a fashion more general and systematic than before,
ment
of

Indeed Boole allows operations which have no direct logical interpretation, and is obviously more at home in mathe matics than in logic. It is probably the great advantage of Boole s work
that he either neglected or was ignorant of those refinements of logical theory which hampered his predecessors. The precise mathematical

mathematical operations.

development of logic needed to make its own conventions and interpreta tions; and this could not be done without sweeping aside the accumulated
traditions of the non-symbolic Aristotelian logic. As we shall see, all the nice problems of intension and extension, of the existential import of universals

and

particulars, of
It
is

empty

classes,

and

so on, return later

and demand

consideration.

enough

well that, with Boole, they are given a vacation long to get the subject started in terms of a simple and general procedure.
s first

book, The Mathematical Analysis of Logic, being an Essay toward a Calculm of Deductive Reasoning, was published in 1847, on the
86

Boole

omit, with

some misgivings, any account

of

De Morgan

contributions to prob

and judgment. (See SylL, pp. 67-72; 384-87, and 393-405.) His work on this topic is less closely connected with symbolic logic than was Boole s. The allied subject of the "numerically definite syllogism" (see Syll., pp. 27-30; F. L., Chap, vm; and Camb.
Phil. Trans., x, *355-*358)
87

ability theory as applied to questions of authority F. L., Chap, ix, x; and Camb. Phil. Trans., vm,

is

also omitted.

George Boole (1815-1864) appointed Professor of Mathematics in Queen s College, Cork, 1849; LL.D. (Dublin, 1852), F.R.S. (1857), D.C.L. (Oxford, 1859). For a biographi cal sketch, by Harley, see Brit. Quart. See also Proc. Roy. Rev., XLIV (1866), 141-81.
Soc.,

xv

(1867), vi-xi.

52
o

Survey of Symbolic Logic

~*same day as
Calculus of
article

De Morgan s Formal Logic. 88


and

The next

year, his article,

"The

Logic,"

appeared in the Cambridge Mathematical Journal.

This

summarizes very

posed by Boole.
in

briefly clearly the important innovations pro But the authoritative statement of his system is found

An Investigation of the Laws of Thought, on which are founded the matical Theories of Logic and Probability, published in 1854. 89
Boole
s

Mathe

algebra, unlike the systems of his predecessors,

is

based squarely

upon the relations of extension.


his

The

three fundamental ideas


"

upon which

(1) symbols"; (2) the Jaws of thought expressed as rules for operations upon these symbols; (3) the observation that these rules of operation are the same which would hold for an algebra of the numbers and I. 90

method depends

are:

the conception of

elective

For reasons which


objects" is

will

appear shortly, the


1.

"

universe of conceivable

represented by

All other classes or aggregates are supposed

to be formed from this


in
1, all

by

selection or limitation.
l-.r

This operation of

electing,
all

the

A"s,

is

represented by

or x;

the operation of electing

the

s is

similarly represented

by

1-y or y,

and

so on.

Since Boole does

not distinguish between this operation of election represented by x, and the result of performing that operation an ambiguity common in mathe matics
.r

becomes, in practice, the symbol for the class of


representing ambiguously

all

the A^

s.

Thus

x, y, z, etc.,

operations of election or classes,


"elective symbols"

are the variables of the algebra.


to distinguish

Boole speaks of them as

them from

coefficients.

This operation of election suggests arithmetical multiplication: the suggestion becomes stronger when we note that it is not confined to 1.
1
"

-x -y or

xy

will represent the

universe",

operation of electing, first, and from this class by a second operation,

all all

the

in the

the Fs.

The

result of these
A"s

and F

s.

two operations Thus xy is the

will

be the class whose members are both

class of the

common members

of x

and
to

y;
z,

xyz, the class of those things which belong at once to and so on. And for any x, l-x = x.

x, to y,

and

The operation of "aggregating parts into a whole" is represented by + + y symbolizes the class formed by x combining the two distinct classes,
.

x and

y.

It is a distinctive feature of

Boole

algebra that x and y in x + y

must have no common members.


88

The
"On

relation

may

be read,

"that

which
Camb.

See

De Morgan s note to

the article

Propositions Numerically

Definite",

Phil Trans., xi (1871), 396. 89 London, Walton and Maberly.


90

Work

This principle appears for the hereafter cited as L. of T.

first

time in the Laws of Thought.

See pp. 37-3g.

The Development of Symbolic Logic


is

53

either x or y but not both". Although Boole does not remark it, x + y cannot be as completely analogous to the corresponding operation of ordinary algebra as xy is to the ordinary algebraic product. In numerical

number may be added to itself: but since Boole conceives the terms of any logical sum to be quite distinct 91 mutually exclusive classes, x + x cannot have a meaning in his system. As we shall see, this is very awkward, because such expressions still occur in his algebra and have to be
algebras a
"

",

dealt with by troublesome devices.

But making the relation x + y completely disjunctive has one advantage it makes possible the inverse relation of "subtraction". The "separa
tion of a part, x, from a whole,
?/",

is

represented by y

x.

If

x+

then since x and

z
]

have nothing

in

common, y

and y

= =

y,

x.

Hence
x +
of x or

and

are strict inverses.

y,

then, symbolizes the class of those things which are either


of y,

members

members

but not of both,

x-y or xy symbolizes the class of

those things which are both


sents the class of the

members of x and members of y. x y repre members of x which are not members of y the x s
two
classes

except the y

s.
i.

=
]

represents the relation of

which have the

same members,

e.,

have the same extension.

These are the fundamental

relations of the algebra.

The

entity

(1

.r)
.r

is

of especial
all

universe except the

s,

or

importance. This represents the It is, then, the things which are not x s.

supplement or negative of x.

With the use


classes, x and
y,

of this

symbolism for the negative of a class, the sum of two which have members in common, can be represented by
xy + x(l

-#)

(!

x)y.
s

The

first

term of

this
s

sum

is

the class which are both x


s;

and y

s;

the second,

those which are x

but not y

the third, those which are y

but not x

s.

Thus the three terms represent classes which are all mutually exclusive, and the sum satisfies the meaning of + In a similar fashion, x + y may
.

be expanded to
x(l

y)

(l

x)y.

Consideration of the laws of thought and of the meaning of these bols will show us that the following principles hold
:

sym
x.

(1)

xy

yx

(2)
91

x +y

y+x

What What

is is

both x and y is both y and either x or y is either y or .r.

See L. of

T., pp. 32-33.

54

A
(3) z(x

Survey of Symbolic Logic

+ y)

zx + zy

That which
is

is

both
z

and and

(either
z

.r

or y)
y.

either both
is

and x or both
z
(a;

and

(4) z(.r

y)

zx

zy

That which
is

both

but not
2

y)
y.

both

and x but not both

and

(5)

If

.T

y,

then

= z + x = x-z =
Z.T
.r

zy
z

+y

(6)

.r

= -

y +

This

last is

an arbitrary convention: the

first

half of the expression gives

the meaning of the last half.


It is a peculiarity of
"logical

symbols"

that

if

the operation

.r,

upon

1,

be repeated, the result


l-x

is

not altered by the repetition:


.
.

=
(7)

l-x-x
2

=
X

x x x

Hence we have

.T

Boole

calls this

the

"index law".

92

All these

laws, except
logic,
"If

(7),

hold for numerical algebra.

It

may
At

be

noted that, in
glance, this

.r

y,

then zx

=
zy"

is

not reversible.

first

may

seem to be another difference between numerical algebra

and the system

in question.

But

"If

zx

zy,

then x

=
y"

does not hold

Law (7) is, then, the distinguishing in numerical algebra when z = 0. The only finite numbers for which it holds are principle of this algebra. and 1. With 1. All the above laws hold for an algebra of the numbers and
this observation, Boole
2 modified by the law x

adopts the entire procedure of ordinary algebra, x, introduces numerical coefficients other than

and

1,

and makes

use,

properties of functions,
to serve his purpose. 93

and

on occasion, of the operation of division, of the of any algebraic transformations which happen

This borrowing of algebraic operations which often have no logical interpretation is at first confusing to the student of logic; and commen
tators have seemed to smile indulgently

upon

it.

An example
it,

will help:

the derivation of the


of
duality",

"law

of contradiction" or, as Boole calls


94

the

"law

from the

"index law".

92 In Mathematical Analysis of Logic he gives he avoids this, probably because the factors of x n

it

also in the

form x n
x) are

(e. g.,

z3

= x, but in L. of T. not always logically

iriterpretable.
93 and 1, and the fractions which can This procedure characterizes L. of T. Only be formed from them appear in Math. An. of Logic, and the use of division and of fractional coefficients is not successfully explained in that book.
94

L. of T., p. 49.

The Development of Symbolic Logic


Since x 2

55

.r,

a-

0.

Hence, factoring,

.r(l

x)

0.

Whoever says
is

This transformation hardly represents any process of logical deduction. 2 "What is both x and a-, a* is equivalent to x; therefore what
,

both x and not-z

is nothing"

may well
if
.r 2

be asked for the steps of his reason

ing.

Nor should we be

satisfied
.r

he reply by interpreting in logical terms

the intermediate expression,

0.

Nevertheless, this apparently arbitrary


algebraic processes
is

way

of using uninterpretable

thoroughly sound. Boole s algebra may be viewed as an abstract mathematical system, generated by the laws we have noted, which has two interpretations. On the one hand, the "logical" or "elec
tive"

cal

symbols may be interpreted as variables whose value is either numeri or numerical 1, although numerical coefficients other than and 1 are
be remembered that such coefficients do not obey which holds for "elective" symbols. All the usual alge
it

admissible, provided

the

"index law"

braic transformations will have an interpretation in these terms.

On

the

other hand, the

symbols may be interpreted as classes. For this interpretation, some of the algebraical processes logical of the system and some resultant expressions will not be expressible in terms
"logical"

or

"elective"

of logic.

But whenever they are interpretable, they will be valid conse quences of the premises, and even when they are not interpretable, any

further results, derived from them, which are interpretable, will also be
valid consequences of the premises.
It

must be admitted that Boole himself does not observe the

proprieties

His consistent course would have been to develop this al without reference to logical meanings, and then to discuss in a thorough gebra
of his procedure.

fashion the interpretation, and the limits of that interpretation, for logical

such a method, he would have avoided, for example, the + .r. We should have + x = 2.r, the interpretation of difficulty about
classes.

By

.1-

.1-

which
classes

for the

numbers

and

1 is

obvious, and

its

interpretation for logical

which

will

would depend upon certain conventions which Boole made and be explained shortly. The point is that the two interpretations

should be kept separate, although the processes of the system need not be limited by the narrower interpretation that for logical classes. Instead
of

making this separation of the abstract algebra and its two interpretations, Boole takes himself to be developing a calculus of logic; he observes that
"axioms"
95

its

are identical with those of an algebra of the

numbers

and

95

1;

L. of T., pp. 37-38.

56

Survey of Symbolic Logic

hence he applies the whole machinery of that algebra, yet arbitrarily rejects from it any expressions which are not finally interpretable in terms of logical The retaining of non-interpretable expressions which can be relations.
transformed into interpretable expressions he compares to

ment

of the uninterpretable
"

of trigonometry.

96

It

"the employ symbol V 1 in the intermediate processes would be a pretty piece of research to take Boole s

algebra, find independent postulates for

it (his

laws are entirely insufficient

as a basis for the operations he uses), complete

it,

and systematically

investi

gate

its

interpretations.
of

But neglecting these problems


logical relations in
will etc.

method, the expression of the simple


will

Boole

symbolism
will

now be

entirely clear.

Classes
y),

be represented by

x, y, z, etc.;

their negatives,

by

(1
is

x), (1

That which
.r(l

is

both x and y

be
;r(l
i.

y), etc.

That which

is

x but not y will x or y but not both, will be x + y, or x or y or both will be x +


or
y)
(1

be xy\ that which

y) + (1

x}y.
is

That which

is

x)y

e.,

that which

x or not x but y

xy + x(l
that which
"universe"

(1

x)y
x.
1

is

both x and y or x but not y or y but not


"everything".

represents the
is

or

The

logical significance of

determined

Oy = 0: the only class which remains unaltered of electing from it whatever is the class "nothing". by any operation Since Boole s algebra is the basis of the classic algebra of which logic

by the

fact that, for

any

y,

the topic of the next chapter it will be unnecessary to comment upon those parts of Boole s procedure which were taken over into the classic
is

algebra.

These

will

be clear to any

who understand

the algebra of logic

in its current
II.

form or who acquaint themselves with the content of Chapter

turn our attention chiefly to those parts of his method which are peculiar to him.
shall, then,

We

Boole does not symbolize the relation

"x

is

included in

?/".

Conse

quently the only copula by which the relation of terms in a proposition can be represented is the relation =. And since all relations are taken in
extension, x

membership.

y symbolizes the fact that x and y are classes with identical Propositions must be represented by equations in which

or == 1, or else the predicate must be quantified. something is put = Boole uses both methods, but mainly relies upon quantification of the This involves an awkward procedure, though one which still predicate.

survives
96

the introduction of a symbol v or w, to represent an indefinite

L. of T., p. 69.

The Development of Symbolic Logic


class

57

and symbolize
vy:
"Some
class"

"Some".

Thus
#",

"All

is

(some)
If
v,

i/"

is

represented by
"the

is

(some)

by wx =
this

vy.

or w. were here

indefinite

or
v,

"any

class",

method would be
that

less objectionable.
it

But
class

in

such cases

or w,

must be very
but
this,

definitely specified:
it
97

must be a
of

"indefinite

in all respects

contains some

members

the class to whose expression

it is prefixed".

The

universal affirmative

can also be expressed, without this symbol for the indeterminate, as .r(l y) = 0; "All xisy" means "That which is x but not y is nothing". Negative
cate.

propositions are treated as affirmative propositions with a negative predi So the four typical propositions of traditional logic are expressed as
98

follows:
All x

is is

y:
y:
is

x
.r

= =

vy,
v(l

or,

x(l

y)

0. 0.

Xo

y), y),
y),

or
or, or,

Some x

y:

v.r

= =

w(\

xy = v =
v

xy.

Some
Each

.r

is

not y:

vx

w(l

x(l

y).

of these has various other equivalents

which

may

be readily deter

mined by the laws

of the algebra.

To

reason

by the

aid of this symbolism, one has only to express his

proper manner and then operate upon the resultant equation according to the laws of the algebra. Or, as Boole more explicitly
premises explicitly in the

puts
"

it,

valid reasoning requires:

"

1st,

That a

fixed interpretation be assigned to the

in the expression of the data;

and that the laws

of the

symbols employed combination of these

symbols be correctly determined from that interpretation. "2nd, That the formal processes of solution or demonstration be con
ducted throughout in obedience to all the laws determined as above, with out regard to the question of the interpretation of the particular results
obtained.
"3rd,

That the

final result

be interpretable in form, and that

it

be

actually interpreted in accordance with that system of interpretation which

has been employed in the expression of the

data."

As we
97

shall

see,

Boole

methods

of solution
is

sometimes involve an

uninterpretable stage, sometimes not, but there


L. of T., p. 63.
is is

provided a machinery by

the above statement

This translation of the arbitrary v by "Some" is unwarranted, and inconsistent with Boole s later treatment of the arbitrary coefficient.
null.

There
98 99

no reason why such an arbitrary coefficient may not be See Math. An. of Logic, pp. 21-22; L. of T., Chap. iv.

L. of T., p. 68.

58

A
may

Survey of Symbolic Logic


be reduced to a form which
is

which any equation


this

interpretable.

To

we must first understand the process known as the develop comprehend ment of a function. With regard to this, we can be brief, because Boole s method of development belongs also to the classic algebra and is essentially
the process explained in the next chapter. 100

Any
the form

expression in the algebra which involves x or

(1

x)

may
it

be

called a function of x.

A
.r).

function of x

is

said to be developed

when

has

Ax + B(l

It is here required that

x be a

"logical symbol",

susceptible only of the values

and
be

1.

But the
"logical

coefficients,
symbol"

not so limited: A, or B,
"law

may

such a

and B, are which obeys the


or
1,

of

duality",

or
If

it

may

be some number other than

or involve

such a number.

the function, as given, does not have the form

Ax

+ B(1

.r), it

may

be put into that form by observing certain interesting

laws which govern coefficients.

Let

/(.r)

Then

/(I) /(O)

And
Hence

= Ax + B(l - x) = A-1 + B(1 ~1) = A - A-Q + B(l - 0) = B =


/(I) -x +/(0)
(1

f(x)

x)

Thus

if

f(x)

~,
2x +
;

Hence

f(x)

(1

x)

developed function of two variables, x and y

will

have the form:


x)(l

y)

And

for

any function,

/(.r, y),

the coefficients are determined by the law:

f(*,y)

/(I, !).?# +/(!,()) x(l

-y)

+/((),

!)(! -x)y

See Math. An. of Logic, pp. 60-69; L. of T., pp. 71-79; "The Calculus of Logic," Cambridge and Dublin Math. Jour., m, 188-89. That this same method of development should belong both to Boole s algebra and to the remodeled algebra of logic, in which + is not completely But a completely developed function, disjunctive, is at first surprising. either algebra, is always a sum of terms any two of which have nothing in common. This accounts for the identity of form where there is a real and important difference in the meaning of the symbols.
[

The Development of Symbolic Logic

59

Thus

if

f( x y)
,

= = =

ax + 2by,
a-1 + 26-1
a1

/(I, 1) /(I, 0)
/((), 1)

=
=

a + 26

+ 26-0
26-1

a
26

/((),

0)

= a-0 + = tt-0 +

= =

26-0

Hence

/(.r,

y)

a + 21) xy + ay (I

y)

+ 26(1

.r)y

An

exactly similar law governs the expansion and the determination of

coefficients, for functions of

any number

of variables.

In the words of

Boole:

101

"The

general rule of development will


will relate to the

consist of

two

parts, the

formation of the constituents of the expansion, the second to the determination of their respective coefficients. It is as
follows
:

first of

which

"1st.

To expand any function of

the

symbols x,

y, z

Form

a series

of constituents in the following

manner: Let the

first

constituent be the

product of the symbols: change in this product any symbol z into 1 z, for the second constituent. Then in both these change any other symbol Then in the four constituents y into 1 y, for two more constituents.
thus obtained change any other symbol .r into 1 .r, for four new constit uents, and so on until the number of possible changes has been exhausted.
"2ndly.

To find

the coefficient of

any

constituent

If
.r

that constituent
into
1
;

involves x as a factor, change in the original function involves


1

but

if it

x as a factor, change in the original function .r into 0. Apply the same rule with reference to the symbols y, z, etc.: the final calculated
value of the function thus transformed will be the coefficient
sought."

Two
solutions
is 0.

further properties of developed functions, which are useful in

and interpretations,

are:

(1)

The product

of

any two constituents

If one constituent be, for example, xyz, any other constituent will have as a factor one or more of the negatives, 1 1 z. .r, y, 1 Thus the product of the two will have a factor of the form x(l And .r).

where
is

.r

is

a
0.

"logical

symbol

",

susceptible only of the values


of
1.

and

1,

x(l

x)

always

ficient 1, the

And (2) if each constituent sum of all the constituents is


which
it

any expansion have the coef

All information

may

be desired to obtain from a given set of

premises, represented by equations, will be got either (1)

by a

solution, to
.r,

determine the equivalent, in other terms, of some


101

"logical

symbol"

or

L.

o/T

.,

pp. 75-76.

60

A
by an

Survey of Symbolic Logic

(2)

elimination, to discover

what statements

(equations), which are


in

independent of volve
x, or (3)

some term

x, are warranted

by given equations which

by a combination
/,

of these two, to determine the equivalent

of x in terms of

other
result.

from equations which involve x, t, u, v, and some logical" symbol or symbols which must be eliminated in the desired Formal reasoning is accomplished by the elimination of "middle"
u,
v,
" "

terms.

The student
of equations

of symbolic logic in its current

form knows that any

set

may

involving a term x

be combined into a single equation, that any equation = 0, and that may be given the form Ax + B(l x)

the result of eliminating x from such an equation is solution of any such equation, provided the condition
will

AB = AB =

0.

Also, the

be satisfied,
s

be x

= B
1,

+ v(l

A), where

is

undetermined.

Boole

methods

achieve these same results, but the presence of numerical coefficients other

than

and

as well as the inverse operations of subtraction

and

division,

necessarily complicates his procedure.


of solutions in the

And

he does not present the matter,

in which we should expect to find it but in a more complicated fashion which nevertheless gives equivalent results. We have

form

now

Boole obviates the

by which which have been mentioned. algebra The simplest form of equation is that in which a developed function,
difficulties of his

to trace the procedures of interpretation, reduction, etc.

of

any number

of variables,

is

equated to

0, as:

Ax + B(l -x) =
Axy +
It
is

0,

or,
.r)(l

B.r(l

y)

+ C(l

x)y + D(l

y)

0,

etc.

an important property of such equations that, since the product of in a developed function is 0, any such equation gives any one of its constituents, whose coefficient does not vanish in the develop ment, = 0. For example, if we multiply the second of the equations given

any two constituents

by

we

xy, all constituents after the first will vanish, giving shall have xy = 0.

Axy =

0.

Whence

Any equation in which a developed function is equated to reduced to the form in which one member is by the law;
1

may be If V = 1,
1
"

--

V =

0.

The more
symbol",

general form of equation

is

that in which some

logical

w,

is

suppose x
x =

equated to some function of such symbols. For example, yz, and it be desired to interpret z as a function of x and y.

yz gives z

x/y-

but this form

is

not in terpre table.

We

shall, then,

The Development of Symbolic Logic


develop x/y by the law
).(!

61

x)y

+/(0,0)-(1 -x)(l

-y)

By

this law:

If z

3*

-, then

= x+-x(l
These fractional

7)

+ 0(1

coefficients represent the sole necessary difference of Boole s

methods from those at present familiar


of division
in

a difference due to the presence

system. Because any function can always be de and the difference between any two developed functions, of the veloped,
his

same

variables,

is

always confined

to

the

coefficients.

If,

then,

we can

interpret

and successfully deal with such


Boole
s

fractional coefficients, one of the

difficulties of

system

is

removed.

The fraction 0/0 is indeterminate, and this suggests that a proper inter pretation of the coefficient 0/0 would be to regard it as indicating an unde termined portion of the class whose coefficient it is. This interpretation may be corroborated by considering the symbolic interpretation of "All
x
is
?/",

which
y)

is

x(l
0,

y)

0.

If

x(l

then x

xy

and x =

xy.

Whence y

x/x.

Developing x/x, we have y


If
"All

x + -

(1

x).

is

?/",

the class y

is

portion of the class

not-.r.

made up Whence

of the class x plus

an undetermined

0/0

is

equivalent to an arbitrary

parameter
or as
"All,

v,

which should be interpreted as some, or none


of".

"an

undetermined portion

of"

The

coefficient 1/0 belongs to the general class of


"index
law",

symbols which do not


"law

obey the
.r(l

x~ =

.r,

or

its

equivalent, the

of

duality",

x)

0.

At

least Boole says it belongs to this class,


fact,

numerical properties of 1/0 would, in


belong to Boole
s

though the depend upon laws which do not


with the class of

system.

But

in

any

case, 1/0 belongs

such coefficients so far as

its logical

interpretation goes.

Any constituent of a

developed function which does not satisfy the index law must be separately

equated

to 0.

Suppose that

in

any equation

w = At + P

62

A
be a
"

Survey of Symbolic Logic

logical

symbol",

whose
be the
law.

coefficient

be a constituent of a developed function does not satisfy the index law, A 2 = A. And let P
t

and

sum Then

of the remaining constituents

whose

coefficients

do

satisfy this

w =
2

w,

t*

t,

and

P = P
2

Since the product of any two constituents of a development

is 0,

Hence

w = (At+P) = A w = A t+P
2 2 2

2 2
t

Subtracting this from the original equation,

(A

Hence

since

- A )t = = A(l = A(l - A) 4= 0,
2

A)t

Hence any equation

of the

form

w = P+OQ+
is

IR+IS
S =

equivalent to the two equations

w = P + vR
which together represent
It will
its

and

complete solution.

be noted that a fraction, in Boole s algebra, is always an am biguous function. Hence the division operation must never be performed: the value of a fraction is to be determined by the law of development only, except for the numerical coefficients, which are elsewhere discussed.

We

have already remarked that ax = bx does not give a = b, because .r may have the value 0. But we may transform ax = bx into a = bx/x and determine this fraction by the law
f(b,x)

/(!, l).fcr+/(l,0).&(l

-*)+/<)

!!

- bx

We

shall

then have

a= -==
and
this
is

bx

bx *

(1

~^
6.

not, in general, equivalent to

Replacing 0/0 by indeterminate

coefficients, v
If

and w,
then
a

this gives us,

ax

bx,

x)

The Development of Symbolic Logic

63

And

this result

is

always

valid.

Suppose, for example, the logical equation

rational

men = married men


Our equation

and suppose we wish to discover who are the rational beings.


will

not give us
rational

married

but instead
married men + v married non-men + w non-married non-men our hypothesis is satisfied if the class "rational is, beings" consist of the married men together with any portion (which may be null) of the
rational
-

That

class

"married women"

and a similarly undetermined portion of the


s

class

"unmarried women".

If

we

consider Boole

system as an algebra of

and

1,

and the

fact that

for

any

fraction, x/y,

--a*+ 6 *(l-)+
we
shall find,
(1)

(l

-*)(!-)
-

by investigating the cases


and y =
1;

(2)

=
x

1,

y
0,

0;

(3)
0,

0,

1;

and
that
it

(4)

requires these three possible cases:


.

Or, to speak

susceptible of the values

more accurately, it requires that 0/0 be an ambiguous function and 1.

Since there are, in general, only four possible coefficients, 1, 0, 0/0, and such as do not obey the index law, of which 1/0 is a special case, this means

that any equation can be interpreted, and the difficulty due to the presence of an uninterpretable division operation in the has disappeared. system

And any equation can

be solved for any

"logical

symbol"

.r,

by trans

ferring all other terms to the opposite side of the equation,

by subtraction
is

or division or both, and developing that side of the equation. Any equation may be put in the form in which one member

by

64

Survey of Symbolic Logic

bringing all the terms to one side.


fully

When

this is done,

and the equation

expanded, all the coefficients which do not vanish may be changed to Boole calls this a "rule of unity, except such as already have that value.
102

interpretation".

Its validity follows

from two considerations:


0,

(1)

Any
not

constituent of an equation with one

member

whose

coefficient does

vanish in development,

may

the constituents thus separately equated to

be separately equated to 0; (2) the sum of will be an equation with one

member

in

which each
coefficients

coefficient will be unity.

Negative

may

be eliminated by squaring both sides of any

equation in which they appear.


are not altered

The

"

logical symbols" in

any function
x),

by squaring, and any expression


symbol",
is

of the

form

(1

where

is

"logical

not altered, since


is

it

and

1.

Hence no constituent

altered, except that its coefficient

can have only the values may be

altered.

And any

negative coefficient will be

made

positive.

Xo new
effected

terms

will

be introduced by squaring, since the product of any two terms


is

of a developed function

always

null.
is

Hence the only change


is

by squaring any developed function


ficients into positive.

the alteration of any negative coef


of

Their actual numerical value


1

no consequence,

because coefficients other than


above.

can be dealt with by the method described


single equation, Boole
103

For reducing any two or more equations to a


first

proposed the

"method

of indeterminate

each equation, after the first, is equations then added. But these indeterminate multipliers complicate the process of elimination, and the method is, as he afterward recognized, an
inferior one.

by which multiplied by an arbitrary constant and the


multipliers",

be reduced, by the methods just described, to the form in which one member is 0, and each coefficient is 1. They may then be simply added; the resulting equation will combine
simply, such equations

More

may

the logical significance of the equations added.

Any
logic.

"logical

symbol"

which

is is

not wanted in an equation


familiar to
is

may

be

eliminated by the

method which

all

students of symbolic

To

eliminate x, the equation

reduced to the form


a-)

Ax + B(l The
result of elimination will be 104

AB =
102

L. of T., p. 90.

See Math. An. of Logic, pp. 78-81; L. of T., pp. 115-120. See L. of T., p. 101. We do not pause upon this or other matters which will be entirely clear to those who understand current theory.
104

103

The Development of Symbolic Logic

65

By

these methods, the difference between Boole


it is

algebra and the classic

algebra of logic which grew out of


results obtainable

reduced to a minimum.
classic algebra

Any

logical

Boole

by the use of the procedures. The difference is

may

also be got

by

solely

one of ease and mathematical


classic algebra

neatness in the method.

Two
s

important laws of the

which

do not appear among Boole


(1)

principles are:
x,

2xy y). with coefficients other than unity, 2x may be replaced in the equation by x, and 2xy + x(l - y) by xy + x(l - y), which

algebra, the methods of reduction which have been discussed will always give x in place of x + x or of x + xy, in any equation in which these appear. The expansion of x + x gives 2x-, the expansion of x + xy gives + x(l

x + xy These seem to be inconsistent with the Boolean meaning of + the first them does not hold for x =-- 1; the second does not hold for x = 1, y = But although they do not belong to Boole s system as an abstract
(2)

x +x

and

of
1.

By

the

method

for dealing

is

equal to x.

should

The methods of applying the algebra to the relations of logical classes now be sufficiently clear. The application to propositions is made
familiar device of correlating the
"

by the
times

logical

symbol",

x,

with the

when some

proposition,

is

true,

xy

X and
times

will represent the

times
1"

when

are both true; x(l

y), the times

when
of

A"

is

true

and

is false,

and so on.

Congruent with the meaning


A"

x+y

will represent the

the times

when either when A

or

(but not both)

is

true.

or

or both are true,


1,

we

xy + x(l

- y)+
and

(1

In ord-r to symbolize must write x + (1 x)y, or


represent
"all

x)y.

the

"

universe",

will

times"

or

"always";
"A

will

be

"no

time"

or
==

"never",

will represent
is

is

always

true";

or

(i

.r)

1,

"A"

is

never true,

always

false".

Just as there

is,

classes, so there is

with Boole, no symbol for the inclusion relation of no symbol for the implication relation of propositions.
is
T"

For

classes,

"All

A"

or

"A"

is

contained in

Y"

becomes x

vy.

Cor
"If

respondingly,

"All

times

when

then
A"

"

or

"A

implies

F"

when

is

true and the times

when Y is or = vy. x = y will mean, "The times when Y is true are the same" or implies
A"

is

true are times

true"

is

"A

Y and Y
The
follows:
105

implies

\
"secondary
propositions"
is

entire procedure for


105

summarized as

L. of T., p. 178.

66

A
"Rule.

Survey of Symbolic Logic

"Eliminate

Express symbolically the given propositions. separately from each equation in which
.

it

is

found the

indefinite

symbol

"Eliminate

the remaining symbols which

it is

desired to banish from

the final solution: always before elimination, reducing to a single equation


those equations in which the symbol or symbols to be eliminated are found.
Collect the resulting equations into a single equation [one
isO],

member

of

which

V =

0.

"Then

proceed according to the particular form in which

it is

desired

to express the final relation, as


1st.

If in

the form of a denial, or system of denials, develop the


all

function V, and equate to

those constituents whose coefficients do

not vanish.
2ndly.
If in

the form of a disjunctive proposition, equate to

the

sum

of those constituents
If in

whose

coefficients vanish.

3rdly.

the form of a conditional proposition having a simple


1

element, as x or

x,

for its antecedent, determine the algebraic

expression of that element, and develop that expression.

the form of a conditional proposition having a com .r)(l pound expression, as xy, xy + (1 y), etc., for its antecedent,
4thly.
If in

equate that expression to a new symbol

t,

and determine

as a developed
. .
.

function of the symbols which are to appear in the consequent.

5thly. only be desired to ascertain whether a particular elementary proposition x is true or false, we must eliminate all the
If it

...

symbols but
tion
is

x\

then the equation x

true, x

that

it is false,

= 1 =

will indicate that the proposi

that the premises are insuf


false."

ficient to

determine whether

it is

true or

It is

a curious fact that the one obvious law of an algebra of


is

and

which Boole does not assume

exactly the law which would have limited

the logical interpretation of his algebra to propositions.


If
is is

The law
1

4=

1,

and

if

4= 0,

exactly the principle which his successors added to his system


to be considered as a calculus of propositions.
his

when

it

This principle would have

made
and

system completely inapplicable to logical classes. For propositions, this principle means, If x is not true, then x
"

is false,

if

is

not

false, it is

true".

But

careful attention to Boole


in his

interpre

tation for

"propositions"

shows that

system x

should be inter-

The Development of Symbolic Logic


preted
"x

67

is
"x

false at all times (or in all


is

terpreted
"

true at

all times".

1 should be in This reveals that fact that what Boole


cases)",
"

and x =

calls

propositions"
is,

are

what we should now

that

statements which

may

prepositional functions be true under some circumstances and false

call

",

upon what we now call "propositions" namely that they must be absolutely determinate, and hence simply true
limitation put
or false
sitional

under others.

The

does not belong to Boole


symbols"

system.

And

his

treatment of

"prepo

gives

them

in the application of the algebra to probability theory the character of prepositional functions" rather than of our
"

absolutely determinate propositions.

one hundred and seventy-five pages of the Laws of Thought are devoted to an application of the to the solution of problems in algebra 106 This application amounts to the probabilities. of a new
last

The

method
is

inyention involved in the problem logical analysis performed as automatically as the purely mathematical operations. a

method whereby any

We

can make this provisionally clear by a single illustration All the objects belonging to a certain collection are classified in three
:

ways

as

^4 s

or not, as

or not, and as
s

C
s

or not.
(2)

It is

then found

that (1) a fraction

m/n

of the .4

are also

B
s

and

the

consist of the
s.

A
it

which are not

together with the


if

which are not

Required: the probability that will also be a C.

one of the

be taken at random,

By

premise

(2)

C = A(l - B)
Since A, B, and

Hence,

C are = A* (I - B) AC

"logical

+ .1(1
s

- A)B = A and A (I - A) = symbols", A - A)B = A(l - B).


+
(l
2

0.

The
that

which are also C

are identical with the

which are not

s.

Thus the probability that


it is

a given

is

also a

is

exactly the probability

m/n, which is the required solution. In any problem concerning probabilities, there are usually two sorts of difficulties, the purely mathematical ones, and those involved in the logical
;

not a

or

by premise

(1 ), 1

analysis of the situation

The methods
106

of Boole

upon which the probability in question depends. algebra provide a means for expressing the relations
and then transforming these
logical

of classes, or events, given in the data,

See also the Keith Prize essay "On the Application of the Theory of Probabilities to the Question of the Combination of Testimonies or Judgments", Trans. Roy. Soc. Edinburgh, xxi, 597 ff. Also a series of articles in Phil. Mag., 1851-54 (see An article on the related topic "Of Propositions Numerically Definite" appeared Bibl).

Chap. 16 ff.

posthumously;

Carafe. Phil. Trans., xi,

396-411.

68

Survey of Symbolic Logic

equations so as to express the class which the quaesitum concerns as a func It thus affords a method for untangling tion of the other classes involved.
the problem
often the most difficult part of the solution.

The
Boole
is

parallelism between the logical relations of classes as expressed in

algebra and the corresponding probabilities, numerically expressed,

striking.

the event
107
A".

Suppose x represent the class of cases (in a given total) in which the occurrence of occurs or those which "are favorable
to"

The
then

p be the probability, numerically expressed, of the event X. total class of cases will constitute the logical "universe", or 1; the

And

let

null class will be 0.

Thus,

if

1
is

if all

the cases are favorable to


If

X
0.

p =

the probability of

X
=

"certainty".

0,

then p

Further, the class of cases in which


1

X does
1.

not occur,
is

will

be expressed by
1

x] the probability that


(1

X will not occur


p)

the numerical

p.

Also,

x+

x)

and p +

(1

This parallelism extends likewise to the combinations of two or more If x represent the class of cases in which events. occurs, and y the class

of cases in

which

occurs, then xy will be the class of cases in

which

and 1 both occur;


(1

x(l

y), the cases in which

occurs without Y;
(1

x)y, the cases in

which

occurs without
y~)

X;

x)(l

y),

the

cases in

X or
and

which neither occurs; x(l occurs but not both, and so on.
events,

+ y(l

x), the cases in which

Suppose that

X and
of

Y are
X,

"

"

simple

"independent"

and

let

p be the probability

q the

prob

ability of Y.

Then we have:
of events
s

Combination

expressed in Boole

algebra

Corresponding probabilities numerically expressed

xy

pq
p(l
(1 (1

- y) x(l - x)y (1
(1

- q) - q)p
+
(1

-*)(! -y)

-p)(l -q)

x(l

y) +

(1

x)y

p(l

q)

p)q

Etc. etc.

In fact, this parallelism

is

complete, and the following rule can be

formulated:
107

108

asserting
thing.
108

Boole prefers to consider x as representing the times when a certain proposition, an occurrence, will be true. But this interpretation comes to exactly the same

L. ofT., p. 258.

The Development of Symbolic Logic


"If

69

p, q,

r,

events, x, y,

z,

are the respective probabilities of unconditioned simple the probability of any event V will be [V], compound
,

this function [V] being


x, y, z,
. . .

formed by changing,
.
.

in the function F, the

symbols

into p,

q, r,

"According

the event

V occur,

to the well-known law of Pascal, the probability that if the event will occur with it, is expressed by a fraction

whose numerator

the probability of the joint occurrence of V and V, and whose denominator is the probability of the occurrence of V. We can then extend the rule just given to such cases:
is
"The

probability that

if

the event

occur,

any other event

will

also occur, will be

[V
-

V]
-

where [V

V]

denotes the result obtained by

multiplying together the logical functions


result x, y,
z,
. . .

and V, and changing

in the

into p,

q, r,

."

The inverse problem of finding the absolute probability of an event when its probability upon a given condition is known can also be solved. Given The probabilities of simple events x, y, z, are respectively when a certain condition V is satisfied. r, p, q,
:

To determine:

the absolute probabilities

/,

m,

n,

of x, y,

z,

By

the rule just given,

U f]
>

[yV]

[zV]

W\

>

And
I,

the
7i,

m,

number of such equations will be equal to the number of unknowns, ... to be determined. 109 The determination of any logical expres
form

sion of the

xV

is is

peculiarly simple since the product of x into

any

developed function
as a factor.
if

the

sum

of those constituents of

V which
*)(!

contain x

For example:
xyz + x(l

V =

xV = xyz + x(l y)z = xyz+(l -x)y(l - z) yV zV = xyz + x(l - y)z + (1 Thus any equation
of the

y)z + (1

x)y(l

z)

(1

y)z,

.r)(l

y)z

form

109

On

Question in the Theory of

xxm
ibid.,

certain difficulties in this connection, and their solution, see Cayley, "On a Probability" (with discussion by Boole), Phil. Mag., Ser. iv,

(1862), 352-65,

and Boole,

"On

a General Method in the Theory of

Probabilities",

xxv

(1863), 313-17.

70

Survey of Symbolic Logic

is

example
"

readily determined as a numerical equation. in illustration: no

Boole gives the following

Suppose that in the drawings of balls from an urn, attention had only been paid to those cases in which the balls drawn were either of a particular

white/ or of a particular composition, marble/ or were marked by both of these characters, no record having been kept of those cases in which
color,

which was neither white nor of marble had been drawn. Let it then have been found, that whenever the supposed condition was satisfied, there
ball

was a probability p that a white ball would be drawn, and a probability q that a marble ball would be drawn: and from these data alone let it be
required to find the probability

ence at

all

that in the next drawing, without refer to the condition above mentioned, a white ball will be drawn;

also a probability
"Here

n that a marble

ball will

be drawn.

if

ball,

the condition

x represent the drawing of a white ball, y that of a marble F will be represented by the logical function

xy + x(l

y) + (1

x)y

Hence we have

xV = yV =
Whence
[xV]

xy + x(l y) + (1 - x)y xy

= =

or

m,

[yV]

= n

and the

final

equations of the problem are

m
mn + m(l
n) + (1

m)n
m)n
+ =p

= P
=
q

mn + m(l
from which we find

n) + (1

m =

p+

q
q

... To meet a possible objection, I here remark that the above reasoning does not require that the drawings of a white and a marble ball should be independent, in virtue of the physical constitution of the balls.
"In

general, the probabilities of

any system

being given, the probability of


logical equation of the

any event

X may
+

independent events be determined by finding a


of

form
x

= A+OB

^C

ID
by a change
of letters.

110

L. of T., p. 262.

have

slightly altered the illustration

The Development of Symbolic Logic

71

where A, B, C, and D are functions of the symbols of the other events. As has already been shown, this is the general type of the logical equation, and its interpretation is given by

= A D =
x

+ vC,

where

v is arbitrary

and

By

the properties of constituents,

we have

also the equation,

A+B+C+D
and, since

=
1

D =

0,

A+B + C
A+B+ C
abilities

thus gives the

universe

of the events in question,

and the prob

given in the data are to be interpreted as conditioned by A + B + C = 1, since = is the condition of the solution x = A + vC. If the given probability of some event S is p, of T is q, etc., then the supposed absolute

probabilities of S, T, etc.,

may

be determined by the method which has

been described.

Let

V = A+B+C,
[sV]

then

W]
where
[sV],
[tV],
etc.

-=P,

M
[IV]

-?,
probabilities"

are the

"absolute

sought.

These,

being determined,

may

be substituted in the equation


Prob.

w =

[A + vC]

\yr~

which

will furnish the required solution.

"The

term vC

will

appear only

in cases

where the data are


it

insufficient

to determine the probability sought.


this probability

Where
v

does appear, the limits of

may

be determined by giving

the limiting values,

and

1.

Thus

Lower

limit of Prob.

w =

\A] -

[A +

Upper

limit

cr

method, and with the theoretical difficulties of its application and interpretation, we need not here concern ourselves. Suffice it to say that, with certain modifications, it is an entirely workable
detail of this

With the

marked advantages over those more matter of surprise that this immediately useful application of symbolic logic has been so generally overlooked.
method and seems
to possess certain

generally in use.

It is a

72

Survey of Symbolic Logic


VI.

JEVONS
s
"calculus

It

has been shown that Boole

of

logic"

is

not so
of

much

system of logic as an algebra of


or propositions,
reasoning.
If

the numbers

and

1,

some

whose ex

pressions are capable of simple interpretation as relations of logical classes,

and some

of

whose transformations represent processes


Boole himself failed to recognize this

of

the entire algebra can, with sufficient ingenuity, be inter


still

preted as a system of logic,

fact,

and

this indicates the difficulty

and unnaturalness

of

some parts
s

of this

interpretation.

Jevons 111 pointed a

way

to the simplification of Boole

algebra, dis

carding those expressions which have no obvious interpretation in logic,

and laying down a procedure which is just as general and is, in important In his first book on this subject, Jevons says: l12 respects, superior.
"So

long as Professor Boole

system of mathematical logic was capable


it

of giving results

beyond the power of any other system,

had

in this fact

an impregnable stronghold. Those who were not prepared to draw the same inferences in some other manner could not quarrel with the manner
of Professor Boole.

But

if

it

chapters

is

of equal

power with Professor Boole

be true that the system of the foregoing s system, the case is altered.

of

There are now two systems of notation, giving the same formal results, one which gives them with self-evident force and meaning, the other by dark
processes.

and symbolic

The burden

of proof

is

shifted,

and

it

must be
in

for the author or supporters of the

dark system to show that

it is

some

way

superior to the evident

system."

He sums up
follows:
"1.

the advantages of his system, compared with Boole

s,

as

113

Even- process is of self-evident nature and force, and governed by laws as simple and primary as those of Euclid s axioms.
"2.

The The

process

is infallible,

and gives us no uninterpretable or anom


less labor

alous results.
"3.

inferences

may

be drawn with far

than in Professor

Boole

system, which generally requires a separate computation and


for

development
111

each

inference."

William Stanley Jevons (1835-1882), B.A., M.A. (London), logician and economist; professor of logic and mental and moral philosophy and Cobden professor of political economy in Owens College, Manchester, 1866-76; professor of political Uni

economy,

versity College, London, 1876-80.


112 113

Pure Logic, or
Ibid., p. 74.

the Logic of Quality apart

from Quantity,

p. 75.

The Development of Symbolic Logic

73

The
unduly
his

third

of these

observations

is

not entirely warranted.

Jevons

restricts the operations


is

and methods

of Boole in such wise that

own procedure

be expeditious.

often cumbersome and tedious where Boole s would Yet the honor of first pointing out the simplifications
in the algebra of logic belongs to

which have since been generally adopted


Jevons.

He
the

discards Boole
of a

inverse operations, a
"

and

a/6,

and he interprets

sum

and

b as

either a or
shall

6,

where a and

b are not necessarily

exclusive

classes".

(We
114

A+B
is

or

-J^.)

symbolize this relation by a + b: Jevons has Thus, for Jevons, a + a = a, whereas for Boole a + a
if it

not interpretable as any relation of logical classes, and


in the algebra of

be taken as

an expression
so that a + a

and

1, it

obeys the usual arithmetical laws,


is

2a.

As has been
and

indicated, this

a source of

much awk

ward procedure

in Boole s system.

The law a + a = a

eliminates numerical

1, and this is a most important simplification. Jevons supposes that the fundamental difference between himself and Boole is that Boole s system, being mathematical, is a calculus of things taken in their logical extension, while his own system, being "pure

coefficients, other

than

logic",

is

a calculus of terms in intension.

It

is

true that mathematics requires

that classes be taken in extension, but

it is

not true that the calculus of

logic either requires or derives important advantage


of intension.

from the point of view

Since Jevons

system can be interpreted in extension without


supposed difference.
:

the slightest difficulty,

we

shall ignore this

The fundamental
(1)

ideas of the system are as follows

a b denotes that

which

is

both a and

b,

or (in intension) the

sum

of the

meanings of the two terms combined. (2) a + b denotes that which is either a or

b or both, or (in intension) a

term with one of two meanings. 115 = b a is identical with 6, or (3) a


(4)
1

(in intension) a

means the same


Boole
s

as

b.

-b
116

Not-6, the negative of

b,

symbolized

in

system by

6.

(o)
"excluded
114 115

indicates that which is contradictory or According to Jevons, from thought". He prefers it to appear as a factor rather than
\

A +B

in Pure Logic; A Jevons would add "but

it is

in the other papers. not known which".

(See Bibl.) (See Pure Logic, p. 25.)

But

hardly correct; it makes no difference if it is known which, since the meaning of a + 6 does not depend on the state of our knowledge. Perhaps a better qualification would be "but it is not specified which". 113 Jevons uses capital roman letters for positive terms and the corresponding small
this is
italics for their negatives.

Following

De Morgan,

he

calls

and a

"contrary"

terms.

74
117

A
itself.

Survey of Symbolic Logic


is

by

intension.

a proper interpretation of the symbol in Its meaning in extension is the null-class or "nothing", as with

The meaning given

Boole.

Jevons does not use any symbol for the "universe", but writes out the This "logical alphabet", for any number n of ele "logical alphabet".
ments,
a, 6, c,
. . .

n consists of the 2 terms which, in Boole 1. b,

system, form

the constituents of the expansion of


the
"logical

alphabet"

consists of a
z,

terms, x - Zf -x

y, z, it

consists of x y

x y

-z,

two elements, a and 6, Thus, a-b, -ab, and -a -b. For three x -y z, x -y -z, x y -z, -x -y z, and
for

Jevons usually writes these in a column instead of adding them -y and putting the sum == 1. Thus the absence of 1 from his system is simply
a

whim and represents no The fundamental laws


(1) (2) (3) (4) (5) (6)
(7)

real difference

from Boole

procedure.

of the
}

system

of Jevons are as follows:

(8) (9)

If a = b and b = c then a = c. a b = b a. a a = a. a -a = 0. a + b = b + a. a + a = a. = a. This law is made use of but is not stated. a+ = a b + a c and (a + b)(c + d) =ac + ad + bc + bd. a(b + c) = a. This law, since called the of absorption", a+ a b
"law

allows

a direct simplification which


multiplication

is

not possible in Boole.

Its

analogue for

a(a+
follows from
(8), (3),

b)

a
of absorption extends to

and

(9).

The law
also
.

any

number

of terms, so that

we have

a+
(10)

ab + ac + ab d+
This
is

=
a,

a(b + -b)(c + -c) ....

the rule for the expansion of


b, c,

any term,
it

with reference to any other terms,

etc.

For three terms

gives us

a(b + -b)
is

(c

+ -c)

= abc

ab-c + a-bc + a-b-c


in

This expansion

identical with that


.

which appears

Boole

system, except

But the product any two terms of such an expansion will always have a factor of the form a -a, and hence, by (4), Thus the terms of any expansion will always represent classes will be null.
for the different meaning of +

of

117

See Pure Logic, pp. 31-33.

The Development of Symbolic Logic

75

which are mutually exclusive.


the different meaning of +
,

This accounts for the fact that, in spite of developed functions in Boole s system and in

Jevons
(11)

always have the same form.

The

"logical

alphabet" is

made up
this
.

of

any term plus


will

its

negative,

a + -a.

It follows

immediately from

and law
.
.

(10) that the logical

alphabet for any number of terms,

a, b, c,

be

and

will

have the character which we have described.


1

the expansion of
its

in Boole s

system because

it is

corresponds to a developed function and

It

terms are mutually exclusive.

procedure by which Jevons sets great store is the "substitution of similars of a for b or b for a when a = b. Not only is this procedure valid
",

when the

expressions in which a and b occur belong to the system, but it holds good whatever the rational complex in which a and b stand. He
first

considers this the


Aristotle
s

principle of reasoning,
et nullo. 11 *

more fundamental than

and yet there


is

is undoubtedly correct, another principle, which underlies Aristotle s dictum, which equally fundamental the substitution for variables of values of these is

dictum de omni

In this he

variables.

And

this procedure

is

not reducible to any substitution of

equivalents.

The only copulative


All a

relation in the system


is

of simple logical propositions


is b:

hence the expression the same as with Boole: substantially


is
[

= ];

No
is

is

a
c a

Some
"U"

is b:
"

= = =

a b a -b
c

ab

or

Ua = Va

used to suggest
of

Unknown".

The methods
than Boole
s,

working with

this calculus are in

some respects simpler

in

some respects more cumbersome.


s

But, as Jevons claims,

they are obvious while Boole


"intrinsic"

are not.

Eliminations are of two sorts,

and

"

extrinsic".

Intrinsic eliminations

substituting for any part of one


other.

member
get

of

may be performed by an equation the whole of the


=
a,

Thus from a =
a

b c d,

we

a c d

= ab

a c

a d
a b

This rule follows from the principles a a

b a,

and

if

b,

ac =
If

be.

For example

a-a
118

= bcd = bed-bed = bb-cc-dd = bcd-d =

ad.

See Substitution of Similars, passim.

76

Survey of Symbolic Logic


-6), or of the

Also, in cases

where a factor or a term of the form a(b +

form

a -a,

is

involved, eliminations
0.
is

may

be performed by the rules a(b + -b)

and a -a =
which

Extrinsic elimination

that simplification or
or

"solution"

may

occur

when two

more are united.

of equations Jevons does not add or

multiply such equations but uses them as a basis for striking out terms in the same "logical alphabet".

This method

is

equivalent, in terms of current procedures, to

first

forming the expansion of 1 (which contains the terms of the logical alphabet) and then putting any equations given in the form in which one member is and "subtracting" them from the expansion of 1. But Jevons did not
hit

upon the current procedures.


"1.

His own

is

described thus:

119

Any

premises

being given, form a combination containing every

term
into

involved therein.
its

Change successively each simple term


all

of this

contrary [negative], so as to form

the possible combinations of


if a,

the simple terms and their contraries.

[E. g.,

b,

and

are involved,

form the

"logical

alphabet"

of all the

terms in the expansion of

(a
"2.

+ -a)(6 + -&)(c + -c).]


[or

Combine

successively each such combination

term, as a be,]

with both members of a premise.

When

the combination forms a con

tradiction [an expression having a factor of the form (a -a)] with neither side of a premise, call it an included subject of the premise; when it forms a

contradiction with both sides, call

it

an excluded subject
call it

of the premise;

when

it

forms a contradiction with one side only,

a contradictory

com

bination or subject,
"We

and

strike it out.

an included or excluded subject a possible subject as distinguished from a contradictory combination or impossible subject. Perform the same process with each premise. Then a combination

may

call

"3.

is

an included subject of a

series of premises,

of of

any one;
any one;

it is it

a contradictory subject

is

an excluded subject

when it is an included subject when it is a contradictory subject when it is an excluded subject of

every premise.

consists of

any term [as a or b] involved in the premises the included and excluded subjects containing the term, treated as alternatives [in the relation + ].
"4.

The

expression of

all

"5.

Such expressions

may

be simplified by reducing

all

dual terms

[of

119

Pure Logic, pp. 44-46.

The Development of Symbolic Logic


the form a(b + -b)], and by intrinsic elimination of in the expression.
"

77

all

terms not required

6. When it is observed that the expression of a term contains a com bination which would not occur in the expression of any contrary of that term, we may eliminate the part of the combination common to the term

and
in

its
"7.

expression.

Unless each term of the premises and the contrary of each appear one or other of the possible subjects, the premises must be deemed in

consistent or contradictory.
possible subjects.

Hence there must always remain at

least

two

Required by the above process the inferences of the premise a = b c. possible combinations of the terms a, b, c, and their contraries are as given [in the column at the left, which is, for this case, the logical
"

"The

alphabet

].

Each

of these being

combined with both

sides of the premise,

we have the

following results:

abc
ab-c
a-b
c

ale
ab-c
a-b
c

= ab = ab

abc

included subject
contradiction contradiction

c -c

ab

-c

a -b -c

a-b

-c

= ab-bc = = a b -b c -c = =
-a
b c

a -b c a -b -c

contradiction

-abc
-ab-c
-a -be
-a -b -c
"It

= = = =

a -a b c
a -a b -c

-abc
=

contradiction

a -a -b c

= -ab c -c = -ab-b c
= -ab
-b
c -c

-a

b -c
c

excluded subject
excluded subject

=0
=

-a -b

a -a -b -c

-a -b -c excluded subject

appears, then, that the four combinations ab-c to -abc are to be struck out, and only the rest retained as possible subjects. Suppose we now require an expression for the term -b as inferred from the premise

b c.

Select from the included


c

and excluded subjects such as contain


c

-6,

namely -a -b
"Then

and -a -b

-c.

-b

= -a
c

-b

+ -a -b -c, but as -a

occurs only with -b, and


eliminate -b from

not with

6,

its

contrary,

we may, by Rule

0,

-a-b

hence -b

-a

+ -a -b

-c."

This method resembles nothing so much as solution by means of the Venn diagrams (to be explained in Chapter III). The "logical alphabet"
is

compartments in such a diagram those marked "contradiction" are the ones which would be struck out in the diagram by transforming the equations given into the form in which one member is 0.
a
list

of the different

78

Survey of Symbolic Logic

The advantage which Jevons claims


ness,

for his method, apart from its obvious namely, that the solutions for different terms do not require to be
is

separately performed,
all

also

an advantage

of the diagram,

which exhibits

the possibilities at once.


If

any problem be worked out by


it will

of Boole,

alphabet"

this method of Jevons and also that be found that the comparison is as follows: The logical consists of the terms which when added give 1, or the universe.
"

Any term marked


ficient

"

"

contradiction

will,

by Boole
will

method, have the coef


will

or 1/0;

any term marked

"included subject"

have the coef

ficient 1;

any marked
is

"excluded subject"

have the

coefficient 0/0, or v

where

arbitrary.

If,

then,

we remember

that, according to Boole,

and thus eliminated, we two methods give substantially the same results. The single important difference is in Boole s favor: the method of Jevons does not distinguish decisively between the coefficients 1 and v. If, for example,
see that the

terms with the coefficient 1/0 are equated to

the procedure of Jevons gives x or x = v -y z.

x -y

z,

Boole

will give either

-y

One

further, rather obvious, principle

may

be mentioned

12
:

Any

subject of a proposition remains an included, excluded, or con

tradictory subject, after combination with any unrelated terms. This means simply that, in any problem, the value of a term remains its value as a factor when the term is multiplied by any new terms which may be

introduced into the problem. In a problem involving a, 6, and c, let a -be be a "contradictory" term. Then if x be introduced, a-bcx and a -b c -x will be "contradictory".

On

the whole Jevons

methods are

likely to be tedious

and have

little

of mathematical nicety

about them.

Suppose, for example, we have three

equations involving altogether six terms.


consist of sixty-four

The
will

"logical

alphabet"

will

members, each of which

have to be investigated

separately for each equation,


operations.
of rejecting

making one hundred and ninety-two separate


It

Jevons has emphasized his difference from Boole to the extent

much

that would better have been retained.

remained for

others,

notably Mrs. Ladd-Franklin and Schroder, to accept Jevons s amended meaning of addition and its attendant advantages, yet retain
s

Boole

methods

solution.

development and similar methods of elimination and But Jevons should have credit for first noting the main clue to
of

this simplification
120

the laws a + a

a and a + a

a.

Pure Logic,

p. 48.

The Development of Symbolic Logic


VII.

79

PEIRCE

contributions of C. S. Peirce 121 to symbolic logic are more numerous and varied than those of any other writer at least in the nineteenth
century.

The

Boole

He understood how to profit by the work of his predecessors, and De Morgan, and built upon their foundations, and he anticipated
when he did not
himself.

the most important procedures of his successors even

work them out


be

recent developments in

Again and again, one finds the clue to the most the writings of Peirce. These contributions may
(1)

summed up under

three heads:

He improved
more

the algebra of Boole

by distinguishing the relations which are


are

characteristic of logical

classes (such as multiplication in Boole s algebra)

from the relations which

more

closely related to arithmetical operations (such as subtraction

and

division in Boole).

The

resulting algebra has certain advantages over the


it

methods of develop ment, transformation, elimination, and solution, and certain advantages over the algebra of Boole because it distinguishes those operations and
retains the mathematical
relations

system of Jevons because

which are always interpretable


"illative"

for logical classes.


in",

Also Peirce
into

introduced the

relation,

"is

contained

or

"implies",

Following the researches of De Morgan, he marked advance in the treatment of relations and relative terms.

symbolic

logic.

(2)

made The

method

of dealing with these

is

made more

precise

and

"mathematical",

and the laws which govern them are related to those classes. Also the method of treating "some" and

of Boole s algebra of
"all"

propositions as

sums

(2) and products (II) respectively of "propositions" containing This is the historic origin of "formal variables was here first introduced.
implication"

and

all

that has been built upon


(3)

it

in the

more recent develop

ment

of the logic of mathematics.

logic to be the science of mathematical form in general,

Like Leibniz, he conceived symbolic and did much to

revive the sense of logistic proper, as

we have used that term.

He worked

out in detail the derivation of various multiple algebras from the calculus of relatives, and he improved Boole s method of applying symbolic logic to

problems
121

in probability.

Charles Saylnders Peirce (1839-1914), son of Benjamin Peirce, the celebrated mathematician, A.B. (Harvard, 1859), B.S. (Harvard, 1863), lecturer in logic at Johns Hopkins, ^596- ?. For a number of years, Peirce was engaged in statistical researches for the U. S. Coast Survey, and was at one time head of the Department of Weights and Measures. His writings cover a wide variety of topics in the history of science, meta According to William James, his physics, mathematics, astronomy, and chemistry.
articles

on

"Some

Illustrations of the Science of

Logic",

Pop. Sri. Mo., 1877-78, are the

source of pragmatism.

80

Survey of Symbolic Logic


in the order
is

We

shall take

up these contributions

named.
mainly in the 122 and in Logic",

The improvement
brief article,
"On

of the Boolian algebra


s

set forth

an Improvement in Boole
the Algebra
.of Logic".
123

Calculus of

two papers,
It will

"On

be remembered that Boole

calculus has four operations, or rela


of the

tions

a + b indicates the class


6;
[

made up
[

two mutually
if

exclusive classes,

a and

is

the strict inverse of

] t

so that

x+b

a,

then x

b;

denotes the class of those things which are common to a and 6; and division is the strict inverse of multiplication, so that if x b = a, then
b or a b

ax

afb.
[ ]

and

These relations are not homogeneous in type. Boole s [+] have properties which approximate closely those of arithmetical
If [n]x indicate

addition and subtraction.


class x,

the

number

of

members

of the

[n]a+ [n]b

[n](a + b)

is

because a and b are mutually exclusive classes, and every a member of (a + b) and every member of b is a member of

member

of a

relation, then, differs

a and b
larly,

This (a + b). from arithmetical addition only by the fact that are not necessarily to be regarded as numbers or Simi quantities.

[n]a

[n]b
s

[n](a
b or

b,

b)

But

in contrast to this, for

Boole

a x

[n]ax
will

[n]b

[n] (a b)

not hold except for

and

this relation
is

is

not of the type of


its

its

arith

metical counterpart.

And

the same

true of

inverse, a/6.

Thus, in

Boole

calculus, addition

as arithmetical addition

are different in

and subtraction are relations of the same type and subtraction; but multiplication and division from arithmetical multiplication and division. type

Peirce rounds out the calculus of Boole


relations,

by completing both sets of these adding multiplication and division of the arithmetical type, and

124 addition and subtraction of the non-arithmetical type. character of these relations is as follows
:

The

general

122

Proc. Amer. Acad., vn, 250-61.


".

This paper will be referred to hereafter as

"Boole

Calculus
123

will

Amer. Jour. Math., in (1880), 15-57, and vn These two papers (1885), 180-202. be referred to hereafter as Alg. Log. 1880, and Alg. Log. 1885, respectively.
124

"Boole

s Calculus,"

pp. 250-54.

The Development of Symbolic Logic


A.
(1)

81

The

"

Non-Arithmetical" or Logical Relations


s

a+

denotes the class of those things which are either a


inverse of the above, a

or 6

or

both.

125

(2)

The
\-b.

\-b,

is

such that

if

x+b

a,

then

Since x and

6,

in

x+

b,

an ambiguous function.
a
\-

need not be mutually exclusive Suppose x + b = a and all b is x.

classes, a h b is

Then

x,

and

a h b

a
b
i.

Thus a
isx.

h b has
\-

an upper
b coincides

limit, a.

But suppose that x +


b (a

=
e.,

a and no b

Then a

with a
x,

which
a
\-

is

not a

6)

f-

and
6,

b
it)

Thus a
a -b.

f-

has a lower limit, a


in

or (as

we elsewhere symbolize
all

And

any

case, a

(-

b is

not interpretable unless

b is a, the

class b contained in the class a.

We may

summarize
[0]

all

these facts

by

a h b

a -6 + v a b +
[0]

b
it is

where

v is

undetermined, and
null.

indicates that the term to which

prefixed
(3)

must be
Boole
a

a b denotes the class of those things which are both a


s
b.
if

and

s.

This

is

(4)

The
is
126

inverse of the preceding, a/b such that


s a/b.

b x

o,

then x

a/6.

This

is

Boole

a/b

an ambiguous function.

Its

upper limit

is

a + -b;
i.

its

lower

limit, a.

It is uninterpretable unless b is contained in a

e.,

a/b

a b + v -a -b +

[0]

a -b

B.
(5)

The

"Arithmetical"

Relations

b denotes the class of those things 6 are

which are either a


This
is

or 6

s,

where a and

mutually exclusive classes.


a

Boole

a +

6.

a -6 + -a 6 +

[0]

a 6
"a

(6)
is

The
6".

inverse of the preceding,

b signifies the class

which

not
(7)
125

a
i

X
rce

pe

As has been mentioned, it coincides with the lower limit of a [-6. 6 and a -f- 6 are strictly analogous to the corresponding relations indicates the logical relations by putting a comma underneath the sign of
is

the relation: that which


126
"

both a and

b is a, b.
:

Peirce indicates the upper limit by a only in the paper Boole s Calculus".

6,

the lower limit by a

-5-

6.-

These occur

82
of arithmetic.
"

Survey of Symbolic Logic

logical"

They have no such connection with the corresponding b. relations as do a + b and a Peirce does not use them
relations, the following familiar laws are stated:

except in applying this system to probability theory.

For the

"logical"

a+ a

= =

a
b
c

a a

a
b a

a+ b
(a +

+ a

ab =
a+
(b

b)

c)

(a b)c

= =

a(b
(a

c)

(a + b)c

= ac

+ bc

a b+ c

c) (b

c)

The

last

Peirce

two are derived from those which precede. s discussion of transformations and solutions

in this

system

is

inadequate.

Any sufficient account would carry us quite beyond what he has given or suggested, and require our report to be longer than the We shall be content to suggest ways in which the methods original paper.
As has been pointed
out,
if

of Boole s calculus can be extended to functions involving those relations

which do not appear in Boole. be developed by Boole s laws,


/(*)
<f>(*>

any function

=/(!)- x +/(0).-z,
y}

=
<p(l,

l)-xy +

<p(l,

Q)-x-y +

^(0, !)

-xy +

<p(Q,

0)-

-x-y,

Etc., etc.,

the terms on the right-hand side of these equations will always represent mutually exclusive classes. That is to say, the difference between the
"logical"

relation,

+, and the

"arithmetical"

relation,

+, here

vanishes.

Thus any
it

can be interpreted by developing to the above laws, provided that we can interpret these rela according
s

relation in this system of Peirce

tions

when they appear

in the coefficients.

And

the correct interpretation

of these coefficients can always be discovered.

Developing the
x+y

"logical"

sum, x +

y,

we have,
0)-

(l

+ l)-xy+(l +

Q)-x-y+(0+l)- -zy+(0 +
and (0+0) =
\-b,

-x-y
(1

Comparing
=

this
==

with the meaning of x + y given above, we find that


1,

1)

1,

(1

+ 0)

(0+1)

==

1,

0.

Developing the
x
\-y

"logical"

difference, a

we have
(0 f-0)
.

(1

\-l)-xy+(l hO)-z-2/+(0 hi)- -x y +


with the discussion of x
\-y

-x -

y
is

Comparing
127

this

above,
0;

we

see that (1 hi)


(1

equivalent to the
"Boole

undetermined
pp. 250-53.

coefficient

that

hO)

1;

that

s Calculus,"

The Development of Symbolic Logic


(0 hi) is

83
it is

equivalent to
null,

[0],

which indicates that the term to which


(0 |-0)

prefixed

must be

and that

0.

The

interpretation of the

"arithmetical"

relations,

and ^,

in coef

not to be attempted. These are of service only in probability theory, where the related symbols are numerical in their
significance.

ficients of class-symbols is

The reader does not


Logic of

require to be told that this system

is

too complicated

to be entirely satisfactory.
Relatives", all

In the

"

Description of a Notation for the


-f-

these relations except


"

are retained, but in later

papers we find only the

logical"
in"

relations, a + b

and a

b.

The

relation of

"inclusion

or

"being

as small

as"

(which we shall

c) appears for the first time in the "Description of a Notation for the Logic of Relatives". 129 Aside from its treatment of relative terms and the use of the "arithmetical" relations, this monograph
symbolize by
gives the laws of the logic of classes almost identically as they stand in the

128

algebra of logic today.


(1)

The

130 following principles are stated.

If

cy and
b,

y cz, then x cz.


is

(2) If (3)

acb, then there


a c

such a term x that a + x such a term y that b y

=
a.

b.

If

then there
a,
(c

is
b.

(4) If b
(5)

=
6,

then a c
+ a) c
(c

If

a c

b)

(6)
(7) (8)
(9)

If
If

a cb,

a
c

a cb, a

cc b. cb c.

abca.

xc(x

y).

(10) x + y = y + x. (11) (x + y) + z = x + (12) x(y + z) (13) x y = y

(y

z).

= xy +
x.

xz.

(14)

(xy)z

x(yz).

(15) x x

(16) (17)
128

= x. x -x = O. 131 x + -x = 1.
is
<

Peirce s symbol

which he explains as meaning the same as

<

but being sim

pler to write.
129 130

Memoirs of

the

"Description

Amer. Acad., n. s., ix (1867), 317-78. of a Notation for the Logic of Relatives," loc.
is

cit.,

pp. 334-35, 338-39,


"

342.
131

In this paper, not-x

1 symbolized by n

"different

from every

x,

or

by

z
<r~

84
(18) z +
(19)

A
= = = x.
1.

Survey of Symbolic Logic

x+

(20)
(21)

<p(x)

?(1)
[*(!)

*+*>(())

-a.

<K.r)

(22)

If

<p(x)

(23) If

= =

0,
1,

<p(x)

equation of condition and the elimination re Boole had stated (22), which sultant for equations with one member 1. is the corresponding law for equations with one member 0, but not (23).

The

last of these gives the

Most
by

of the

above laws, beyond

(9),

had been stated

either

Jevons.

(1) to (9) are, of course, novel, since the relation

by Boole or c appears

here for the same time since Lambert.

Later papers state further properties of the relation c


If

notably,

x c y, then -y c -x.

And

the methods of elimination and solution are given in terms of this


Also, these papers extend the relation to propositions.
"If

relation. 132

In this
is

interpretation, Peirce reads x cy,

a;

is

true,

is

true,"

but he

well

aware
of
"a;

of the difference

between the meaning of x


says:
133

cy and
is

usual significance

implies
is

y".

He

"It

stated above that this

means

if a: is

true, y

true

But

this

meaning

is

greatly modified by the circumstance that only the actual state


is

of things

referred

to.

... Now

the peculiarity of the hypothetical


if

proposition [ordinarily expressed by

is

true, y

is

true

is

that

it

goes

out beyond the actual state of things and declares what would happen were The utility of this is that it puts things other than they are or may be.
us in possession of a rule, say that if A we afterward learn something of which
is

true,

is

true

such that should

we

are

now

ignorant,

namely that

is

true, then,

by virtue

of this rule,

we

shall find that


.

we know something
.
.

else,

namely, that
true
if

is

true.
if

[In contrast to this]

the proposition,
is false.

ac
.
.

b, is
.

is

false or

b is true,

but

is false if

is

true while b

For example, we shall see that from -(x cy) [the negation of x cy] we can infer z ex. This does not mean that because in the actual state of
things x
is

true and y false, therefore in every state of things either z


;

is

false or x true

but

it

does

mean

that in whatever state of things


is false

we

find x

true and y

false, in
is

that state of things either z


anyway]."

or x

is

true [since,

ex hypothesi, x
132

true

133

2. Alg. Log. 1880, see esp. Alg. Log. 1885, pp. 186-87.

The Development of Symbolic Logic

85

We now call this relation, x


theorems which are true of
of
it

"

cy,

material

implication,"

and the peculiar

are pretty well


if

them.

They

will

be intelligible
is

the reader

"The

actual state of things

not one in

Peirce gives a number remember that x c y means, which x is true and y false".
"A

known.

(1)

arc (yea:).

This

is

the familiar theorem:

true proposition is

implied by any
(2)

proposition".

[(xcy) ex] ex.

If

"x

implies

y"

implies that x

is

true,

then x

is

true.
(3) [(a: cy) ca] ex, where a is used in such a sense that means that from x cy every proposition follows.
(ar

cy) c a

The
meaning
time. 134

difference between
of
"implies"

"material

implication"

and the more usual


this

is

a difficult topic into which

But

it

is

interesting to
its

we need not go at note that Peirce, who introduced


some
of his successors

the

relation, understood

limitations as

have not.

Other theorems in terms of this relation are:


(4) (5) (6)
(7)
ar

c x.

[xc(ycz)] c [yc(xcz)].

xc[(xcy)cy]. (ar cy)c[(ycz)c


of the

(xcz)]..

This

is

a fundamental law, since called

the

"Principle

Syllogism".

Peirce

worked most extensively with the

logic of relatives.

His interest

here reflects a sense of the importance of relative terms in the analyses of

mathematics, and he anticipates to some extent the methods of such later


researches as those of Peano and of Principia Mathematica.
his successive papers

To

follow

on

this topic

would probably
shall

result in

complete con

fusion for the reader.

Instead,
Peirce:

we
(1)

make

three divisions of this entire

subject as

treated

by

the modification and extension of

De

Morgan

calculus of relatives

by the introduction of a more

"mathe

matical"

symbolism for the most part contained in the early paper, "Description of a Notation for the Logic of Relatives"; (2) the calculus

of relations, expressed without the use of exponents

and

in a

form which

makes

it

an extension of the Boolean algebra


its

a later development which

may

be seen at

best in

"The

Logic of

Relatives",

Note B

in the Studies

in Logic by members of Johns Hopkins University; consideration of the theory of relatives, which
papers, but has almost complete continuity.
134

and

(3) the systematic

is

scattered throughout the

But

see below, Chap, iv, Sect,

i,

and Chap,

v, Sect. v.

86

A
The terms

Survey of Symbolic Logic

of the algebra of relatives


"ancestor",

may

usually be regarded as simple


etc.

relative terms, such as


will

"lover,"

Since they are also class

names, they obey may be taken for granted without further discussion. But relative terms have additional properties which do not belong to non-relatives; and it is to these that our attention must be given. If w signifies "woman" and s,
all
"servant,"

the laws of the logic of classes, which

logic

is

"logical

product"

concerned not only with such relations as "servant woman", s + w, the sum" "logical
both)
",

iv,

the

"either

servant or
in the

woman (or class women

and

relations
first

cw, "the class servants which belong to class-terms


s

is

contained

in general of a

but also with the relations


woman," "servant

symbolized
"

by De Morgan,
of
U5

"servant

of every

woman,

and

"servant

none but
This
is

women".

We may

represent

"servant

of a woman"

by s\w.

a kind of

"multiplication"

relation.

It is associative,

s\(l\w)
"Servant

=
is

(s\l)\w
"

of a lover-of-a-woman

"

servant-of-a-lover of a

woman".

Also,

it is

distributive with respect to the non-relative


,

"addition"

symbol

ized

by +

s\(m + w)
"Servant

= s\m + s w
a
"

of either a
".

man

or a

woman" is "servant of
:

man

or servant of
"

a
is

woman

But

it is

not commutative
"lover

s\lis not

l\s,

servant of a lover

not equivalent to

of a

servant".

To

distinguish s

from

s w,

or s x
call s
|

w w

the class of those

who
s

are both servants

and women
w
,

we

shall

the relative product of


"servant

and w.
proposed
s

For

of every woman" Peirce


s

and

for

"servant

of none but

women"

w.

As we

shall see, this notation is suggested

by

certain mathematical analogies.


of the class

as

W W W
ly

We may
and the

represent individual
class of all the
,

members

2,

3,

etc.,

Ws

as

W +W
l

+ ....

Remembering

the interpretation of +

we may

write

w =
and
that
this
is,

W W
1

W, +

means,

"The

class-term, w, denotes

or

or
s.

or

...,"

denotes an unspecified

member

of the class of PF

The servant

of a (some, any)

woman

is,

then, s w.

sw
"A

woman" is either
135

or

W
is s

or

3,

etc.;
s,

"servant

of a woman"

is

either

Peirce s notation for this

w, he uses

for the simple logical product.

The Development of Symbolic Logic


servant of

87

or servant of

or servant of PF3 etc.


,

of every woman" is servant of


etc.
;

Similarly,
z

"servant

and servant of
,

and servant of

z,

or remembering the interpretation of x

n represents the relative product, and x s\ n of the non-relative logical product translated by "and". The represents above can be more briefly symbolized, following the obvious mathematical
"s
,"

where, of course,

analogies,

w = 2 W s\w = 2 w (s W) w = H s w (s\W)
|

Unless

represent a null class,

we

shall

have
or

jc*w
"servants

The

class

"servants

of every woman"

is

contained in the class

of a woman".

This law has numerous consequences, some of which are:

lover of a servant of

all

women

is

a lover of a servant of a
(/]*)

woman.

Kc
A

lover of every servant-of-all-women stands to every

woman

in the rela

tion of lover-of-a-servant of hers (unless the class s w be null). s w c Is w I


|

lover of every servant-of-a-woman stands to a (some)

woman

in the

relation of lover-of-a -servant of hers.

From

the general principle, 136

m\[U x f(x)]cU f (m\f(x)]


136

The

proof of this

theorem
...
. . .

is

as follows

a
or a

= abc = abc

+ab

-c ...

+a

-b c ... + ...,

+ P, where

is

the

sum

of the

remaining terms.
+
,

Whence,

if

represent any

relation distributive with respect to


.

Similarly,

mOa = mQabc mOb = mQa be mOc = mQa b c


Etc., etc.
let a, 6, c, etc.,

+mQP + mOQ
+

rnQR

Now

be respectively
the
left side,

/(.Ci), f(x*),

fM,

etc.,

and multiply together

all

the above equations.

On

we have

88

A
also,

Survey of Symbolic Logic

we have

^c*,
If
/

or

8cs
woman
in the

lover of a (some) servant-of-every-woman stands to every

relation of lover-of-a-servant of hers.

We

have also the general formulae of inclusion, c s,


then
l

w w

c sw c
I
s

and,

If s

c w, then

The

first of

these means:

If all lovers are servants,

then a lover of every


:

woman is also a servant of every woman. The second means are women, then a lover of every woman is also a lover of
These laws
are, of course, general.

If all

servants

We
=
I

have

also:

every servant.

(l\s)\w

(s\w)

The
or a

last of these is read:

lover of every person

who

is

either a servant

woman

is

a lover of every servant and a lover of every

woman.

An

interesting law

which remainds us of Lambert


(I

"Newtonian formula" is,

s) w

+ 2 q (l w

-<*

x s9) + s w

One who

is

either-lover-or-servant of every

woman,
women,
is

is

either lover of every

woman

or, for

some portion

q of the class

lover of every
q, or, finally, is

woman
servant

except members of q and servant of every


of every

member

of

woman. Peirce also gives this law in a form which approximates even more closely the binomial theorem. The corresponding law for the
product
is

simpler,
(I

xs)

xs w

which

is

On

the right side,

we have
(raOP) + (mQQ) +

(mOabc
where

...) +

(mOR)

..,

or

(mQabc...)+K
is

is

sum

of other terms. ...}


is

But

(mOabc

?nQ[f(x

x f(x2 ) x /(z3 )

.],

which

mOH x f(x)
Hence [mOHf(x)]+K = U x [mOf(x)]. Hence m O n /Or) c n x [m Of(x)
]
.

Peirce does not prove this theorem, but illustrates


"Description
137

it

briefly for logical multiplication (see

of a

Notation",

p. 346).

"Description of

a Notation, p. 334.

The Development of Symbolic Logic

89

One who

woman

both-lover-and-servant of every and a servant of every woman.


is

woman,

is

both a lover of every

Peirce introduces a fourth term, and summarizes in a diagram the inclu sion relations obtained by extending the formulae already given. 138 The

number

of such inclusions, for four relatives,

is

somewhat more than one


precise

hundred eighty.

He

challenges

the

reader to accomplish the

formulation of these by means of ordinary language and formal logic. An s of none but members of w, Peirce symbolizes by 8 w. He calls this
operation
"

"backward involution",
relatives",

and

relatives of the type

w he

refers to

as

infinitesimal

on account of an extended and

difficult

mathe

matical analogy which he presents. 139

The laws
l

of this relation are analo

gous to those of

sw

If s
If all

c w, then

c w
l

servants are

women, then a
If
/

lover of none but servants

is

lover of none

but women.

c s, then

wc w
l

If all

lovers are servants, then a servant of none but

women

is

a lover of

none but women.


i(*

w) =

("OH;

The

lovers

of

none but servants-of-none-but-women are the

lovers-of-

servants of none but

women.
i+*w

= w
l

Those who are either-lovers-or-servants of none but women are those who
are lovers of none but

women and

servants of none but

women.

(w xv)

= w
s

The servants
those

of

none but those who are both women and

violinists are

who

are servants of none but

women and
)&

servants of none but vio

linists.
<">MJC

Whoever

is

lover-of-a-servant of

none but

women

is

a lover-of-every-

servant of none but

women.
l\

wc

<*>w

lover of one

who

is

servant to none but

women

is

a lover-of-none-but-

servants to none but

women.
l

wc

(s

w)

138

Ibid., p. 347. Ibid., pp.

139

348 jf.

90

A
to a

Survey of Symbolic Logic


in the relation of lover-of-nothing-but-servants

Whoever stands
of hers
is

woman

a lover of nothing but servants of

women.
by the laws:

The two kinds

of involution are connected

lover of none but those

who

are servants of every

woman

is

the same as

one who stands to every


servants of hers.

woman
i

in the relation of a lover of

none but

-/

- HO
It

Lover
from
s

of

none but servants


x

is

non-lover of every non-servant.


:

appears

this last that x


)

and y are connected through negation


a lover of every servant a lover of
is

= -O) =
-(I

-l\s,
l\-s,

Not Not

non-lover of a servant.
is

none but servants

lover of a non-

servant.
l

-s

~(l\s}

-I s ,

A
,

lover of none

but non-servants

is

one who

is

not lover-of-a-servant, a non-lover of every servant. -i s = = l~ s A non-lover of none but servants -(-/| -s)

is

one

who

is

not a non-lover-of-a-non-servant, a lover of every non-servant. We have the further laws governing negatives: 141

In the early paper,


of
Relatives",

the Description of a Notation for the Logic negatives are treated in a curious fashion. A symbol is
"On

used for

"different
s".

from

"

and the negative

of s

is

represented by

s
,

"

differ

ent from every

Converses are barely mentioned in this study. In the paper of 1880, converses and negatives appear in their usual notation, "relative addition" is brought in to balance "relative multiplication", and
the two kinds of involution are retained.
in the

But

in

"

The Logic

of Relatives

"

Johns Hopkins Studies in Logic, published in 1882, involution has


"relative addition"

disappeared, converses and negatives and

are retained.

This last represents the

final

form of Peirce

calculus of relatives.

We

have here,
(1)

Relative terms,

a, b,

... x, y, z.

(2)
140
141

The negative
ibid., p. 353.

of x, -x.

See

Not-z

is

here symbolized by

(1

z)

Alg. Log. 1880, p. 55.

The Development of Symbolic Logic


(3)

91

The converse
a;

of x, ^x.

If

is

"lover",

^.r is

"beloved";

if

KC

is

"lover",

is

"beloved".

(4)

Non-relative addition, a +

b,

"either

a or

6".

(5)
(6) (7)

Non-relative multiplication, a x6, or a


"a

b, "both

a and

6".

Relative multiplication, a\b, of a Relative addition, a t b, of everything but 6


6".
"a

s,

a of every non-6

".

(8)
(9)

The The

relations

= and c

as before.
1,

universal relation,

"consistent

with,"

which pairs every

term with
(10)

itself

and with every

other.
1.

(11)
itself.

The The

null-relation, 0, the negative of

relation

"identical

with",

7,

which pairs every term with

(12)

The

relation

"different

from",
142

N, which pairs any term with

every other term which

is

distinct.

In terms of these, the fundamental laws of the calculus, in addition to


those which hold for class-terms in general, are as follows:
(1)

(2)
(3) (4) (5)

- a = w(- ) -(wo) = (-6 c (a c b)


.(.a)
If If

acb, then (a\x)c(b\x) and (x\a)c(x\b). acb, then (a t x) c (6 t x) and (x t a) c (x t

6).

(6)
(7)

(x a)|6 x\(a\b) x t (a t b) = (x t a) t b

(8)
(9)

x\(a-tb)c(x (a t b) x c a t

a)

t6

(6 x)

(10) (a\x) + (b\x)c(a+b)\x


(11) (12)

z|(at&)c(3ta)(a:t&)

(13)
(14)

(a+b)xc(a\x) + (b\x) (at*)(6tz)c(a|6) to: -(at 6) - -a -6


|

(15) -(a 6)
1

(16) (17)
(18)

(19)

= -at -6 ^(a+ 6) = ^a+ ^b = a ^6 (a b) t 6) = ^a t ^6 w(a w( |6) = v a |w6


1, 0,
/,

For the relations


given
142
:

and N, the following additional formulae are


may
see

have altered Peirce

notation, as the reader

by comparison.

92

A
(20)

Survey of Symbolic Logic


(21)

Ocx
= x + 1 = 1 ztl = 1 1 t.T = 1 sttf = x Art* = x + -x = 1
x+
.r

xcl
= = z-0 z|0 = 0\x =
x

(22) (24) (26)


(28) (30) (32)

(23) x-l
(25)

(27)
(29)

(31)

z|/-cg

(33) /|z

(34)
(36)

or

(35) (37)

= x x -x =
fcr|

7c[a;1X-aO]
is,

(-*)]

ctf

This calculus
ciples of solution

as Peirce says, highly multiform,


laid

and no general prin

and elimination can be

down. 143

Not only

the variety

of relations, but the lack of


relative addition,
e. g.,

symmetry between

relative multiplication

and

in (10)-(13) above, contributes to this multiformity.

But, as

we now know, the


from
less
its

chief value of

any calculus
s

of relatives

is

not in

any elimination

or solution of the algebraic type, but in deductions to be

made

directly

formulae.

Peirce
is

devices for solution are, there

fore, of

much

importance than

his calculus of relatives is built.

the theoretic foundation upon which It is this which has proved useful in later

research and has been

made

the basis of valuable additions to logistic

development. This theory

is

practically unmodified throughout the papers dealing


"Description
Relatives"

with relatives, as a comparison of of Relatives" with "The Logic of

of a

in the

Notation for the Logic Johns Hopkins studies

and with the paper


"Individual"

of 1884 will indicate.


"elementary"

or

relatives are the pairs (or triads, etc.)

of individual things.
etc.,

If

the objects in the universe of discourse be A, B, C:


will constitute the

then the individual relatives

two-dimensional array,

A:A,A\B,A:e,A:D,
B
:

... ... ...

A,

B,

C,

D,

C:A, C:B, C:C, C:D,


. . .

Etc., etc.
itself is

be noted that any individual thing coupled with relative but that in general A B differs from B A
It will
: :

an individual

individual relatives

are ordered couples.

A
143

general relative
"Logic

is

conceived as an aggregate or logical


in Studies in Logic by

sum

of such

of

Relatives"

members of Johns Hopkins University,

p. 193.

The Development of Symbolic Logic


individual relatives.
If b

93

represent
b

"benefactor",

then

= Z,S

y (6) t.,.(/

,/),

where (&)/
factor of J,

is

a numerical coefficient whose value


0,

is

in case / is a

bene

and otherwise

and where the sums are to be taken

for all the

individuals in the universe.

That

is

to say, b

is

the logical
is

sum

of

all

the

benefactor-benefitted pairs in the universe.


of
"definition

This

the

first

formulation
in

in

extension",

now widely used

in logistic,

though seldom

exactly this form.

By
It is
b

this definition, b is the aggregate of all the individual

relatives in our two-dimensional array

which do not drop out through having


of the form,
r
:

the coefficient 0.

some expression

= (X: y)!+(Z:
logical

Y) 2

+(X
,

) 3

+...

If,
"b

now, we consider the


is

meaning

of +

we

see that this


".

may

be read,

either

(X

F), or

(X

F) 2 or

(X

F) 8 or ...
is,

To say

that 6 repre
b repre
it

sents the class of benefactor-benefitted couples

then, inexact:

sents an unspecified individual relative, any one of this class.

should represent

"some"

in a sense

(That which denotes more than one at once


admits
is

which the meaning of +


fact that
exclusive.)
"real

in the general case

precluded by the

any two

distinct individual relatives are ipso facto


relative, so defined, is

mutually
calls

genera)

what Mr. Russell

variable".

Peirce discusses the idea of such a variable in a most

144 illuminating fashion.

"Demonstration of the sort called

mathematical

is

founded on suppo
the required condi
case, his

sition of particular cases.

The geometrician draws


quantity

a figure; the algebraist

assumes a
tions.
esis
is

letter to signify a certain

fulfilling

But while the mathematician supposes a particular

hypoth

yet perfectly general, because he considers no characters of the

must belong to every such case. The ad vantage of his procedure lies in the fact that the logical laws of individual terms are simpler than those which relate to general terms, because indi
individual case but those which
viduals are either identical or mutually exclusive, and cannot intersect or be subordinated to one another as classes can.
. .
.

"The

old logics distinguish between

indimduum signatum and

indi

mduum vagum. Julius Caesar is an example of the former; man of the latter. The indimduum vagum, in the days when
,

a certain

such con

ceptions were exactly investigated, occasioned great

difficulty

from

its

having a certain generality, being capable, apparently, of logical division.


144

"Description

of a Notation, pp. 342-44.

94
If

A
we
,

Survey of Symbolic Logic


as

include under

indimduum vagum such a term


all

any individual
is

man

these difficulties appear in a strong light, for

what

true of

any

individual

man

is

true of
it

men.

Such a term

is in
it

one sense not an

individual term; for

represents every

man.
is

But

represents each

man

as capable of being denoted by a term which


it is

not
.

itself
. .

terms.

so, though an individual term, it stands for any one of a class of such The letters which the mathematician uses (whether in algebra

individual; and

or in geometry) are such individuals

by second

intention.

All the

formal logical laws relating to individuals will hold good of such individuals

by second

intention,

and at the same time a universal proposition

may

be

substituted for a proposition about such an individual, for nothing can be

predicated of such an individual which cannot be predicated of the whole


class."

The
It
is

relative

b,

fitted pairs in the universe,

denoting ambiguously any one of the benefactor-beneis such an individual by second intention.
of the
:
"

defined

by means

prepositional

function",
"I

"I

benefits
J"

J",

as the logical

sum

of the (7

J) couples for

which

benefits

is

true.

The compound
If

relations of the calculus can be similarly defined.

S,-Sy(a)iy(Z

J),

and

SiSy(&)<y(/

J),

then a +

S,-Sy[(a),-y +

(&),-,](/

J)
:

That
"7

is,

if

"agent"

is

the logical

sum

of all the (7
is

J) couples for which


of all the (7
:

is

agent of

J"

is

true,

and

"benefactor"

the

sum
"

J)

couples for
is

which

"7

benefits

J"

is

true, then

"either

agent or

benefactor"
is

the logical

sum
J"

of all the (7
is

J) couples for which

Either 7
facts

agent of

or 7 benefits

true.

We

might indicate the


+

same
6) t-,-.

more simply

by

defining only the

"prepositional function", (a

145

The

definition of a + b given above, follows

immediately from this simpler


relations are similar:

equation.

The

definitions of the other

compound

(ax6)
or
"Both

(a) t-yx(6) iy
:

a xb

SiSy[(a) tV x (6) iy ](7

J)
:

agent and
"7

benefactor" is

the logical

sum
J"

of the (7
is

J) couples for

which

is

agent of

J and
(a
I

is

benefactor of

true.

&).-/

S*{(a)*x(6) Ay
<A

or
145

a\b
",

S,-Sy[S*{(a)
188.

(b) hi ]](I

J)

See

"Logic

of Relatives

loc. cit., p.

The Development of Symbolic Logic


"Agent

95
J) couples such
is

of a benefactor"

is

the logical

sum
//
is

of all the (7

that, for

some

//,

"I

is

agent of 77

and

benefactor of

J"

true.

There are two

difficulties in
"agent

the comprehension of this


of a benefactor".

last.

The

first

concerns the meaning of


treats

Peirce, like

De Morgan,
itself

his relatives as denoting

ambiguously either the relation


or the class

or

the things which have the relation

either relations or relative terms.

is

either the relation

"agent

of"

name

"agent".

Now

note

that the class


relation.
clear.

name denotes the first term in With this in mind, the compound
of a benefactor"
"

the pairs which have the


relation,
s

a\b, will
:

become
which

"Agent

names the 7

in the /
".

pairs

make up the field of the relation, agent of a benefactor of Any reference J s at the other end of the relation is gone, just as omits The second any reference to the J s in the field of the relation "agent concerns the operator, SA, which we have read, "For some difficulty
to the
"agent"
of".

H".

is

Consider any statement involving a "prepositional the variable representing the individual of which

function",
<p

<pz,

where

is

asserted.

That

is,

2z
3

<pz

symbolizes
",

"Either
<p

is

true of Zi or

<p

is

true of
"For

or
z

(p

is

true of

or ...

and

this

is

most simply expressed by


<pz

some

(some
"/

2 or other),

<pz".

In the particular case in hand,


is

is

(o)t*x(6) A y,

is

agent of // and //
the relation
"7

benefactor of

J".

The

terms, I and J, which stand in


are those for which there is
is

is

agent of a benefactor of
is

J",

some

H or other such that 7

agent of
"

H and H

benefactor of J.
<pz

Suppose we
ator n.

consider any

propositional

function",

with the oper

Tl 2 (pz

=
<f>Zi

(pZ/z

<pZs

x
<p

...
is

That
of
3

is,

"

<pz

symbolizes
",

is
<p

true of Z\ and
z".

true of

and

<p

is

true

Z and

or
b.

"

is
<p

true for every

This operator

is

needed

in the

definition of a t

(ot&),/- n A {(a)
"

+(6) A
is

,-}

agent of everyone but benefactors of either 7 is agent of 77 or is benefactor of


is

"

J"

equivalent to

For every

77,

H ot6-

J".

2&[Il h {(a) ik +(b) h j}](I:J)


is is

"Agent

of all non-benefactors"

the logical

sum

of all the (7
is

J) couples

such that, for every H, either 7

agent of

H or 77

benefactor of J.

The

same considerations about the ambiguity

of relatives

denoting either the

96

Survey of Symbolic Logic

relation itself or those things


in this case also.

which are

first

terms of the relation


still

applies

We

need not, for the relations

to be discussed, con
"prepositional

sider the step


tions",

from the definition of the compound


is

func

(a t b)a in the above, to the definition of the corresponding relation,

a t

b.

This step

always taken in exactly the same way.

The

converse, converse of the negative, and negative of the converse,

are very simply defined.

That the negative

of the converse
-(6).-/

is

the converse of the negative follows


.

from the obvious fact that

(-6)/ f

All the formulae of the calculus of relatives,


also to the calculus of non-relative terms, 146
definitions.

beyond those which belong may be proved from such

For example:

To

prove, v(a + b)
(a

^a + ^b
(a

&),-/

&),-,-

(a) /t-+
(6) yf
t

(&)/<

But

(a)ji

= (a)

t -/,

and
w o) (

(w6) tV

Hence

(a + &)/
:

-/+

(W>)v

of

Hence S<S,Ka + 6),-,-} (7 J) S;Z y {(-a) t7 + (&)/}(/ J) Q.E.D. For the complete development of this theory, there must be a discussion the laws which govern such expressions as (a) or in general, expressions
:

,-,-,

of the
is

form

(px,

where

<px

is

a statement which involves a variable, x, and

(px

either true or false

whenever any individual value


"

of the variable is
functions".
147

specified.
(a)*,

or in

Such expressions are now called the more convenient notation,


or
<p(x,

propositional
is

<px,

a propositional function of

one variable;

(a),-/, y), may be regarded as a propositional function of two variables, or as a function of the single variable, the individual rela

tive (7

J), or

(X

Y).
is

This theory of propositional functions

stated in the paper of 1885,

On

the Algebra of

Logic".

It

is

assumed, as also in earlier papers, that

4 he laws

of the algebra of classes hold for propositions as well. 148


is

The
time.

Additional law which propositions obey

stated here for the

first

46 The formulae of the calculus of classes can also be derived from these, considered as themselves laws of the calculus of propositions (see below, Chap, vi, Sec. iv). 147 Peirce has no name for such expressions, though he discusses their properties acutely

(see Alg. Log. 1880,


148

2).
first

This assumption

appears in Alg. Log. 1880.

The Development of Symbolic Logic

97

The current form


"If

of

this

law

is

"If

4= 0,

then x
is

1",

which gives
is

& * 1, then x = immediately and if x is not true, then x is false".


"false",

0"

"If

not

false,

then x

true,

Peirce uses v
is

instead of

and

0,

and the law


(x

stated in

and / for the form

"true"

and

-f)(v-x)

=0
though derived from the
"x

But the calculus


true nor false.
in

of

propositional
is

functions,

algebra for propositions,

not identical with

it.

is

man" is

neither

A
"x

propositional function
a:

may
a

be true in some cases, false


true in
all cases,

some

cases.

"If

is

a man, then x

is

mortal" is

or

true of any x;
of x.

is

man" is

true in some cases, or true for

some values

For reasons already suggested,

2x

"

<px

represents
<px

<?x

is
2

true for
is

some value
<px

of the variable, x
. .
."

that

is,

either

is

true or

<px

true or

3 is

true or

Similarly,

TL x TL t
tf
<px

=
<px

(pxi

<px

<px

...

represents
true and
(a) xy ,
,
<px

tpx
2

is

true for

all
is

values of the variable, x


true and
..."

that

is,

<pxi

is

is

true and #r 3

If

or

more conveniently,
\f/(x,

<p(x,

y), represent

"x

is

agent of
y",

y",

and

(b) xy or

more conveniently,
IL x

y),

mean

"x

is

benefactor of

then

2 y [<p(x,y) *t(x,

y)]
y,
"x

will

mean
is

that for

all

values of x and some values of


y"

is

agent of y

and x

benefactor of

is

true

that

is,

it

represents the proposition

"Everyone is
if

both agent and benefactor of


[<f>(x,

someone".

This

will

appear

we expand n x 2 y

y) x^(.r, y)]:

2/0
2/j)

xiKzi,
,

2/i)]

+
+

[<?Oi,

2/2)

x x

{[<p(x

2,

x^(.r 2 yi)]

[<f>(x

z,

y z ) x^(z z y 2 )] +
,

..)
.

U</>(>3,

2/i)

x^(.r 3

2/i)]

[^(.r 3 , 2/2)

x^(.r 3 2/2)]+
,

... Etc., etc.


"

This expression reads directly


factor of yi] or [xi
[x z is
is

{Either [xi
is

agent of y 2 and Xi
a: 2

agent of y\ and x\ is bene benefactor of y z or .[ and {either


is
]
. .

agent of

2/1

and
. .

is

benefactor of
{either [x s is
.r 3

y\] or [x 2

is

agent of y 2
is
.

and x 2

is

bene

factor of y 2 ] or
2/i]

.}

and

agent of y\ and x 3
]
.
.

benefactor of
.

or [# 3 is agent of y 2
8

and

is

benefactor of y z or

and

Etc.,

etc".

98

A
Z,

Survey of Symbolic Logic


"Boolian"

The operator
it,

which

is

nearer the argument, or

as Peirce calls
II,

indicates the operation,

within the

lines.
i.

The
e.,

outside operator,
in the

indicates the operation,

the subscript of

columns; and the operator nearer the Boolian indicates the letter which
x, between the lines
line to line.
little

varies within the lines, the subscript of the outside operator, the letter

which varies from


sional array.

Three operators would give a three-dimen

With a

patience, the reader

may

learn to interpret

any such expression directly from the meaning of simple logical sums and For example, with the same meanings of logical products. y) and
<p(x,

t(x, y),

lUZj^Cr,
will

y) xifr(y, x)]
(y)

mean

"Everyone

(x)

is

agent of some

benefactor of

himself".

(Note the order of the variables in the Boolian.)

And

"

will

symbolize

There
is

is

is

agent of z or z

benefactor of
is

some x and some y such that, for every z, either or, more simply, "There is some
y";

pair,

x and

y,

such that x

agent of

all

non-benefactors of

y".

The laws

for the
:

manipulation of such Boolians with


149

n and 2

operators

are given as follows


"1st.

The

different premises

having been written with distinct indices

(the same index not being used in two propositions) are written together, and all the n s and Z s are to be brought to the left. This can evidently be

done, for

[Or in the

more convenient, and probably more


Il x
<f>x

familiar, notation,

Xll y

<py

= U xU y

(<px

x
<py)

x
<py)]

"2d.

Without deranging the order


and Z
s

of the indices of

any one premise,

the

n
149

belonging to different premises

may

be

moved

relatively
left

to one another,

and as
pp.-

far as possible the

should be carried to the

Alg. Log. 1885,

196-98.

The Development of Symbolic Logic


of the
II s.

99

We

have

n,-ny.T t-y

S.-SyZiy

= n itey = SyZtf.-y
;

[Or, [Or,

mix*,
S.S^fo

y)
y)

= nji^or, y)] = S y S z ^(a-, y)]


x^y)J

and

also
this

Silly&.-yy

= n/Z^y

[Or, SJI y (^r x


i

^) = n y S x (^r

But

formula does not hold when

and

j are not separated.

We do

have, however,
SillyZiy
It will, therefore,
-<

IlyZtf.v

[Or,

ZJI^ar,

y)

c n,2 x ^(.r,

y)]

be well to begin by putting the 2 s to the left as far as because at a later stage of the work they can be carried to the possible,
right but not [always] to the left.

premises are
orders

II

For example, if the operators of two and 2 JI 2 z we can unite them in either of the two ^II*
tf

and

shall

often be
. . .

room
"oth.

usually obtain different conclusions accordingly. There will for skill in choosing the most suitable arrangement.

The next

step consists in multiplying the whole Boolian


itself

part, by the modification of


of

produced by substituting for the index


to the left of
it

any n any other index standing

in the Quantifier.

Thus,

for
Silly/,-,-

[Or, for

2JI 2JI

tf *>(.r,

y),

we can write
"6th.

2 JI//^

tf (

<p(x,

y) x ^(.r, x)}]

The next

step consists in the re-manipulation of the Boolian


in

adding to any part any term we like; 2d, from any part any factor we like, and 3d, in observing that
part, consisting, 1st, in

dropping

xx
so that
"7th.

= =

/,

x+x
(x

v,

xxy +

+ x + y)z

and S

in the Quantifier

whose indices no longer appear

in

the Boolian are dropped.


"The

fifth

step will, in practice, be combined with part of the sixth

and seventh.

Thus, from 2

!! ,-/,-/

we

shall at

once proceed to 2

/ lt if

we

like."

We may
<pxi

say, in general, that the procedures

which are valid


x

in this

calculus are those which can be performed

by treating 2 x v?.r
tpx\
<^.r

as a sum,
<^.r

<.r

<px

and

Il x

<px

as

product,
x,

2 JI,,^(,r,

y) as a

sum, for the various values of

of products, each for

100

Survey of Symbolic Logic

the various values of y, and so on.

Thus

this calculus

may

be derived from

the calculus of propositions.

of the principles of the system,

But Peirce does not carry out any proofs and he notes that this method of proof
"It

is to be remarked that SjX,- and would be theoretically unsound. 150 IliXi are only similar to a sum and a product; they are not strictly of that

nature, because the individuals of the universe

may

be

innumerable."

Another way

of saying the

calculus of propositions

same thing would be this: The laws of the extend to 2iX and Ito, because the extension cannot
-

of these laws to aggregates in general,


cal analogies of

sum and product


is

by the method which the mathemati suggest, would require the principle of

mathematical induction, which


gate
is

not sufficient for proof in case the aggre

infinite.

The whole
that

of the calculus of relatives

may

be derived from this calculus

of propositional functions
is,

by the methods which have been exemplified


relation, 6, as 2*2/(&){/(7
of",
:

by representing any
II

J),

relations, such as

"converse

"relative-product,"

and defining the etc., which dis

tinguish the calculus, as

We
of

of the elementary relatives. need not enter into the detail of this matter, since Sections II and III
will

and 2 functions

Chapter IV

modification of Peirce

develop the calculus of propositional functions by a s method, while Section IV of that chapter will show
will indicate the

how

the calculus of classes can be derived from this calculus of propositional

functions, Section
tions

manner
s

in

which the calculus


will suggest

of rela

may

be similarly derived, and Section

VI

how, by a

further important modification of Peirce


logic of
It

method, a theoretically adequate

mathematics

may

be obtained.

remains to consider briefly Peirce s studies toward the derivation of other mathematical relations, operations, and systems from symbolic logic.

The most important paper,


Mathematics".
151

in this connection,

is

"Upon

the Logic of

Certain portions of the paper,


Logic",

Boole

Calculus of

and

of the

an Improvement in monograph, "Description of a Nota


"On

tion for the Logic of

Relatives",

are also of interest.


is

The
+
,

first-mentioned of these
etc., of arithmetic

concerned to show

how

the relations

cal

can be defined in terms of the corresponding logi =, relations, and the properties of arithmetical relations deduced from
logical analogues.
152

theorems concerning their


"Imagine
150
151

...

a particular case under Boole

calculus, in

which the

Alg. Log. 1885, p. 195. Proc. Amer. Acad., vn, 402-12. Loc.
tit.,

152

pp. 410-11.

The Development of Symbolic Logic


letters are

101

no longer terms of
first

first

intention, but terms of second intention,

and

that of a special kind.

... Let

the letters

relate exclusively to

the extension of
things

intensions.

Let the differences of the characters of


let

the letters signify only the differ In other words, the only logical comprehension which the letters considered as terms will have is the greater
ences of classes as wider or narrower.
or less divisibility of the class.

and events be disregarded, and

Thus, n in another case of Boole


States
;

calculus

might, for

example, denote
all
it

New England
make
what

but in the case now

supposed,
neglected,

the characters which

these states

what they

are, being

would

signify only

essentially belongs to a class

which

has the same relation to higher and lower classes which the class of

New

England States has,


11

that

is,

a collection of

six.

if

In this case, the sign of identity will receive a special meaning. For, denotes what essentially belongs to a class of the rank of sides of a
,

cube
State

then [the
is

logical]

m=

will

imply, not that every

New England

the side of a cube, and conversely, but that whatever essentially belongs to a class of the numerical rank of New England States essentially

belongs to a class of the rank of sides of a cube


of this particular sort
If a, b,
c,

and conversely.

Identity

may

be termed equality.

..."

etc. represent

thus the number of the classes,

a,

b,

c,

etc.,

then the arithmetical relations can be defined as logical relations. The logical relation a + 6, already defined, will represent arithmetical addition:

And from
will follow
is

the fact that the logical that the arithmetical

+
is

is

commutative and

associative,

it

so also.

Arithmetical multiplication
153

more
a

difficult to deal

with but

may

be defined as follows:
b are events

b represents

an event when a and

only a
is

if

these events
is

are independent of each other, in


logical product].

which case a

X
3,

[where a b

the
is

By
two

the events being independent

meant that
B\,

it

possible to take

series of terms, A\,

A A
2,
,

etc.,

and

B B
2,

3,

etc.,

such that the following conditions are


individual or class, not nothing;
,

satisfied.
, ,

(Here x denotes any

A m A n Bm B n
(.4

two

series of terms,

and 2 A, 2 B, 2

B)

logical

any members sums of some

of the
of the

A n s,

the

B n s,

and the (A n

BJ

respectively.)

Condition

1.

2. 3.

4.
153

No A m is A n No m is n x = 2 (.4 B) a = 2 A

Loc.

cit.,

p. 403.

102

Survey of Symbolic Logic


5.

Condition

= 2B
is

6.

Some A m

Bn
of the

This definition
a b

is

somewhat involved: the crux


with a

matter

is

that

will, in the case described,

have as many members as there are combina

tions of a

member
will

of a

member

of

b.

Where the members

of a are

distinct (condition 1)

and the members

of 6 are distinct (condition 2), these

combinations
It is

be of the same multitude as the arithmetical a

b.

worthy

of

remark

that, in respect

both to addition and to multi

plication, Peirce has here hit

of

upon the same fundamental ideas by means which arithmetical relations are defined in Principia Mathematical
"second

The

intention"

of a class

term

is,

in Principia,

Nc

b,

in

Peirce
sum"

discussion, corresponds to

what

is

there called the


is

"arithmetical

of

two

logical classes,
s

and a

b to

what

called the "arithmetical


all

product".

But Peirce

discussion does not

meet

the difficulties
it

that

could hardly be expected in a short paper. In particular, define the arithmetical sum in case the classes summed have

does not
in

members

common, and
class,

it
it

though

does not indicate the manner of defining the number of a does suggest exactly the mode of attack adopted in Prin

cipia,

namely, that number be considered as a property of cardinally similar


for the derivation of the laws of various numerical

classes taken in extension.

The method suggested

algebras from those of the logic of relatives is more comprehensive, though here it is only the order of the systems which is derived from the order of

the logic of relatives; there

is

no attempt to define the number or multitude

of a class in terms of logical relations. 155

We

individual in which

are here to take a closed system of elementary relatives, every is either a T or a P and none is both.

Let

c
s

= (T T) = (P:P)
:

p = (P:T)
t

= (T :P)
:

Suppose T here represent an Individual teacher, and P an individual pupil the system will then be comparable to a school in which every person is
either teacher or pupil,
pupil.
354 155

and none

is

The

relative

term, c, will

both and every teacher teaches every then be defined as the relation of one

See Vol. n, Section A. "Description of a Notation, pp. 359 jf.

The Development of Symbolic Logic


teacher to another, that
"

103
P), the rela

is,

colleague".
is,

Similarly, s

is

(P
"

tion of one pupil to another, that


is

"schoolmate".

The

relative

term, p,

any pupil to any teacher, that is, pupil". And the relative term, t, is (T P), the relation of any teacher to any pupil, that is, "teacher". Thus from the two non-relative terms, T and P, are generated the four elementary relatives, c, s, t, and p.
(P
:
:

T), the relation of

The

properties of this system will be clearer

if

we venture upon

certain

explanations of the properties of elementary relatives which Peirce does not give and to the form of which he might object. For any such relative
(7
:

J),
(1)

where the /

and the J

are distinct,

we

shall

have three laws:

(I:J)\J =

1
(7
:

Whatever has the

J) relation to a

J must

be an 7: whoever has the

teacher-pupil relation to a pupil


(2)

must be a

teacher.

(7:./)l/

=
relation to a teacher (where teachers

Whatever has the teacher-pupil

and

pupils are distinct) does not exist.


(3)

(I

:J)\(II:K) =

[(I

J)\H]

K
(7
:

The

relation of those
:

which have the

J) relation to those which have


:

the (77

K)

relation

is

the relation of those-which-have-the-(7

J)-relation-

to-an-77 to a

K.
law which
is

It is this third

the source of the important properties of

the system.
t\p

For example:

= (T :P)|(P
of
fit

T)
s

[(T

P) |P]

T = (T
s

1
:

The teachers
or his

any person

pupils are that person

colleagues.

(Our

illustration, to

the system, requires that one

may

be his

own

colleague

own

schoolmate.)

c\c

= (T
= (T

T)\(T
s

T)

[(T

T)\ T]
s

T = (T

T)

The

colleagues of one
t
t
:

colleagues are one


:
|

colleagues.
T]
:

P) (T

P)

[(T

P)

P = P =

(0

P)

There are no teachers of teachers in the system.

= (P

T)

(P

P)

[(P

T) \P]

(0

P)

There are no pupils of anyone s schoolmates in the system. The results may be summarized in the following multiplication table,
in

which the multipliers are

in the

column at the

right

and the multiplicands

104

Survey of Symbolic Logic


156
:

at the top (relative multiplication not being commutative)


t

The symmetry

of the table should be noted.


it

The reader may

easily in

terpret the sixteen propositions which

gives.

To

the algebra thus constituted


letters.

may

be added modifiers of the terms,


"French",

symbolized by small roman


of the system in case

If f is

will

be a modifier

French teachers have only French pupils, and vice versa. Such modifiers are "scalars" of the system, and any expression of the form
a c +

bt+

p+d

s
a, b, c,

where

c,

t,

p,

and

s are

the relatives, as above, and

d are scalars,

Peirce calls a
is

"logical

quaternion".

The product

of a scalar with a

term

commutative,

bt = tb
since this relation
is

that of the non-relative logical product.

Inasmuch as

any (dyadic,
triads, etc.)

triadic, etc.) relative is resolvable into a logical


it is

sum

of (pairs,

elementary relatives,

plain that any general relative what

ever

is

If

sum of logical quaternions. we consider a system of relatives, each of which


resolvable into a

is

of the

form

az +
where
i,

bj + ck + dl+

...

j, k,

I,

etc. are

each of the form

mu + nv + oiv+
we

...

where m, n, o, etc. are scalars, and u, v, w, etc. are elementary relatives, shall have a more complex algebra. By such processes of complication, multiple algebras of various types can be generated. In fact, Peirce says: 157 can assert, upon reasonable inductive evidence, that all such
"I

[linear

on the principles of the present notation in the same way as those given above. In other words, all such algebras are complications and modifications of the algebra of (156) [for which the multiplication table has been It is very likely that this given].
155

associative] algebras can be interpreted

Ibid., p. 361. Ibid., pp.

157

363-64.

The Development of Symbolic Logic


is

105

true of

all

algebras whatever.

The algebra

of (156),

which

is

of such a

pure algebra and our logical nota tion, has been shown by Professor [Benjamin] Peirce to be the algebra of
in reference to

fundamental character

Hamilton

s quaternions."

Peirce gives the form of the four fundamental factors of quaternions

and

of scalars, tensors, vectors, etc., with their logical interpretations as relative

terms with modifiers such as were described above.

One more item


of probabilities.

of importance
is

is

Peirce

modification of Boole

calculus

This
in

set forth
s

an Improvement
identity of

Boole

with extreme brevity in the paper, "On Calculus of Logic". 158 For the expression of

the relations involved,

we

shall

two

classes in extension

need to distinguish the logical relation of from the relation of numerical equality.

We
of

may, then, express the


b,

fact that the class a has the


all
fe

as the class

or

all

are

s,

by a

b,

and the
of

fact that the


of
b,

same membership number


by a =
b.

members of a is the same Also we must remember the


pressed

as the

number

members

distinction between the logical relations ex


relations

by a +

b,

ab, a
b,

\-b,

and the corresponding arithmetical


and a

expressed by a
"Let

b,

b.

Peirce says:

159

every expression for a class have a second meaning, which is its meaning in a [numerical] equation. Namely, let it denote the proportion
of individuals of that class to be
in the long run.
"Then

found among

all

the individuals examined

we have
If

"

as a class,

Let b a denote the frequency of the 6 s among the a s. Then considered if a and b are events b a denotes the fact that if a happens b happens.

X
"

ba

ab

It will
ties of

be convenient to set
.

down some obvious and fundamental proper


a

the function b a

Xb =
a,

X
1

ab
C) a

<p(b

Ca ) b) a

(1

= =

<p(b,

ba

158 159

Proc.

Amer. Acad., vn, 255 ff.


255-56.

Ibid., pp.

106

Survey of Symbolic Logic


I a ---7- X

&

0(l_a)

The
abilities
(1)

chief points of difference

between

this modified calculus of

prob

and the

original calculus of Boole are as follows:


p, q, etc. for the "probability of a, of b,
etc.",

Where Boole puts


from the

in passing

logical to the arithmetical interpretation of his

equa

simply changes the relations involved from logical relations to the corresponding arithmetical relations, in accordance with the foregoing,
tions, Peirce

and

lets

the terms

a, b, etc.

stand for the frequency of the a

s,

s,

etc.

system under discussion. Boole has no symbol for the frequency of the a s amongst the 6 s, which Peirce represents by a b As a result, Boole is led to treat the
in the
(2)
.

probabilities of all unconditioned simple events as independent

a pro

cedure which involved him in


(3)

many

difficulties

and some

errors.

Peirce has a complete set of

four

logical

operations,

and four

analogous operations of arithmetic.

This greatly facilitates the passage

from the purely

logical

expression of relations of classes or events to the

arithmetical expression of their relative frequencies or probabilities.

Probably there is no one piece of work which would so immediately reward an investigator in symbolic logic as would the development of this calculus of probabilities in such shape as to make it simple and practicable.

Except

for a

monograph by Poretsky and the

studies of H. MacColl, 160

the subject has lain almost untouched since Peirce wrote the above in 1867.
Peirce s contribution to our subject is the most considerable of any up to his time, with the doubtful exception of Boole s. His papers, however, are brief to the point of obscurity: results are given summarily with little
or no explanation

the most valuable of

and only infrequent demonstrations. As a consequence, them make tremendously tough reading, and they

serves. 161

have never received one-tenth the attention which their importance de If Peirce had been given to the pleasantly discursive style of
clearly accurate

De Morgan, or the detailed and work on symbolic logic would fill


160

manner

of Schroder, his

several volumes.

an unusually
logic.
161

Since the above was written, a paper by Couturat, posthumously published, gives clear presentation of the fundamental laws of probability in terms of symbolic

See Bibl.

find our report of Peirce requested to consult the original papers.

Any who

work unduly

difficult or

obscure are earnestly

The Development of Symbolic Logic

107

VIII.

DEVELOPMENTS SINCE PEIRCE

of Peirce

Contributions to symbolic logic which have been made since the time need be mentioned only briefly. These are all accessible and in a
sufficiently close to current notation to be readily intelligible.

form

Also,

they have not been superseded, as have most of the papers so far discussed; consequently they are worth studying quite apart from any relation to later work. And finally, much of the content and method of the most *

important of them
later chapters, or
will

is
is

substantially the

same with what

will
is

be set forth in
there set forth

such that

its

connection with what

be pointed out.

But

for the sake of continuity

and perspective, a

summary account may be given of these recent developments. We should first mention three important pieces of work contemporary
with Peirce
s later treatises. 162

Robert Grassmann had included

in his encyclopedic Wissenschaftslehre


3

a book entitled Die Be griffslehre oder Logik,


Begriffen,
(2)

containing
(3)

(1)

Lehre von den

Lehre von den Urtheilen, and


is

Lehre von den Schlussen.

The

Begriffslehre
this

the second book of Die Formerdehre oder Mathematik,

and as

would

indicate, the

development of
s

logic

is

entirely mathematical.

An important

character of Grassmann

procedure

is

the derivation of the

laws of classes, or Begriffe, as he insists upon calling them, from the laws governing individuals. For example, the laws a + a = a and a- a = a,

where a
where
it,
e,

is

a class, are derived from the laws e + e

e,

e-e

e,

e\-e<i

0,

e\, e 2

represent individuals.

This method has

much

to

commend

but

it

has one serious defect

the supposition that a class can be treated

as

an aggregate of individuals and the laws of such aggregates proved As Peirce has observed, this method generally by mathematical induction.
breaks

down when

the

number

of individuals
is

may

be

infinite.

Another

difference

between Grassmann and others

the use throughout of the


of extension,

language of intension.

But the method and the laws are those


diagrammatic
areas.

and

in the later treatise, there are

illustrations in

which
in

"concepts"
162

are represented

by

Although somewhat incomplete,

Alexander MacFarlane, Principles of the Algebra of Logic, 1879, gives a masterly presentation of the Boolean algebra. There are some notable extensions of Boole s methods and one or two emendations, but in general it is the calculus of Boole unchanged. MacFarlane s paper "On a Calculus of Relationship" (Proc. Roy. Soc. Edin., x, 224-32) re
sembles somewhat, in
its method, Peirce s treatment of "elementary relatives". But the development of it seems never to have been continued. 163 There are two editions, 1872 and 1890. The later is much expanded, but the plan and general character is the same.

108

Survey of Symbolic Logic


s

other respects Grassmann

calculus

is

not notably different from others

which follow the Boolean tradition.

Hugh MacColl
164

first

two papers on
paper
"On

"The

Calculus

of

Equivalent
165

in

printed 1878-80, present a calculus of propositions which has essentially the In others words, it is properties of Peirce s, without II and 2 operators.
Statements",

and

his first

Symbolical

Reasoning",

it

a calculus of propositions, like the Two-Valued Algebra of Logic as we know today. And the date of these papers indicates that their content was
s

arrived at independently of Peirce

studies

which deal with

this touic.

In fact, MacColl writes, in 1878, that he has not seen Boole. 166

The

calculus
167

set

forth

in

MacColl

book, Symbolic

Logic

and

its

Application*,

is

of

an entirely different character.


traditional truth values,
"certain",

Here the funda

mental symbols represent propositional functions rather than propositions;

and instead
have
tain
77,

of the

two

"true"

and

"false",

we

"true",

"false",

"impossible"

and

"variable"

(not cer

and not impossible). and procedures


of

8 respectively.

These are indicated by the exponents r, i, e, The result is a highly complex system, the fundamental

ideas

which suggest somewhat the system

of Strict

Implication to be set forth in Chapter V.

The

calculus of Mrs. Ladd-Franklin, set forth in the paper


Logic"

"On

the

Algebra of

in the

Johns Hopkins studies,

168

differs
.

from the other


b are

systems based on Boole by the use of the copula v or "Some a is classes, a v b represents is-partly
"a

Where a and
6",

6",

av

"

b,

represents

a is-wholly-not-6
b to

"

",

or
0.

No

is

".

and its negative, Thus a v 6 is equiva


a

lent to a b =h 0,

and a v

a b

These two relations can, between

them, express any assertable relation in the algebra,

cb

will

be a v-6,

and a = b is represented by the pair, (a v-b)(-a v b). For propositions, a v b denotes that a and b are consistent a does not imply that b is false and b does not imply that a is false. And a v6 symbolizes and b are
"a

inconsistent"

if

is

true, b

is false;

if

6 is true, a is false.

The
is

use of the

terms

"consistent"

and

"inconsistent"

in this connection

possibly mis

leading:
164

any two true propositions or any two


ix,

false propositions are con177-86.

166

(1) Proc. London Math. Soc., Mind, v (1880), 45-60. Proc. London Math. Soc., ix,

9-20;

(2) ibid., ix,

178.

Longmans, 1906. 38 The same volume contains an interesting and somewhat complicated system by O. H. Mitchell. Peirce acknowledged this paper as having shown us how to express uni versal and particular propositions as n and S functions. B. I. Oilman s study of relative number, also in that volume, belongs to the number of those papers which are important
in connecting symbolic logic with the theory of probabilities.

167

The Development of Symbolic Logic


sistent in this sense,

109

and any two propositions one


This
is
"it is

of

which

is

true

and

the other false are inconsistent.


"consistent"

not quite the usual meaning of

and

"

inconsistent
"

related to

what

is

usually

these terms exactly as the


is

material implication a c b
a".

is

related to

meant by what
x

usually

meant by

"b

can be inferred from

That a given

class, x, is

empty, or a given proposition, x,

is false,

0,

may

be expressed by x v
1.

oo,

where

co

is

"everything"

represented by

That a

class, y,

has members,

is

most systems symbolized by y v oo.


in

where y is a proposition, since Mrs. system does not contain the assumption which is true for propositions but not for classes, usually x =(= 0, then expressed, x = 1, and if x =|= 1, then x = x v oo may be abbreviated to xv,
is

This last

of doubtful interpretation
s

Ladd-Franklin

"If

0".

a b v

v and T/VGO to yv c dv co to c d v understood that if one term of a relation v or v


co

to a b

etc., since it is

always

is

missing, the missing

term
tion:

is

This convention leads to a very pretty and convenient opera v or v may be moved past its terms in either direction. Thus,
oo
.

(a

v 6) =

(a

by) =

va
vx

6)

and

(xvy) = (xyv) =

y)

But the forms (va6) and (vxy) are never used, being redundant both logically and psychologically.
Mrs. Ladd-Franklin
logic particularly well
:

system symbolizes the relations of the traditional

All a

is b.

-b,

or or
or or

a -6 a b a b

No

is b.

avb, avb,
6.

Some Some
Thus v

a a

is b.

is

not

a v -6,

a -6

characterizes a universal,

v a particular proposition.
"

And any
then
all

pair of contradictories will differ from one another simply

by the difference

between v and v
a
"

The

syllogism,

If all

is

and

all b is c,

is c,

will be represented

by v -b) (b v -c) v
(a

(a

v c)

where v, or v, within the parentheses


pression

is interpreted for classes, and v between the parentheses takes the propositional interpretation. This ex

may

also be read,

"

All a

is

and
is

all b is c

is

inconsistent with

the negative (contradictory) of


(a

Some

not

".

It is

equivalent to

-b) (b

v -c) (a v -c) v

110

A
three propositions,

Survey of Symbolic Logic


All a
is

"The

All b

is c,

and

Some

is

not

are inconsistent
three syllogisms:

they cannot

all

three be

true".

This expresses at once

(1)
"If

(a

v-6)(6 v-c) v
all

(a

v-c)

all

is

and

all 6 is c,

then
(a

is

c";

(2)
"If

v -b) (a v -c) v
c,

(b

v -c)
not
c";

all

is

and some a
(3)

is

not
(b

then some

b is

v -c) (a v -c) v
c,

(a

v -6)
is

"

If all b is c

and some a

is

not

then some a

not b

".

Also, this

method

gives a perfectly general formula for the syllogism


(a

v -b) (b v c) (a v c) v

where the order of the parentheses, and their position relative to the sign v which stands outside the parentheses, may be altered at will. This single
rule covers all the

modes and
169

figures of the syllogism, except the

illicit

particular conclusion drawn from universal premises.


this

We

shall revert to

Chapter III. The copulas v and v- have several advantages over their equivalents, = and =j= 0, or c and its negative: (1) v and v are symmetrical rela tions whose terms can always be interchanged; (2) the operation, mentioned above, of moving v and v with respect to their terms, accomplishes trans
formations which are
the copula;
(3) for

matter

in

less

simply performed w ith other modes of expressing


r

various reasons,

it is

psychologically simpler and

more
terms

natural to think of logical relations in terms of


of

v and v than

in

=
,

and

=|=

0.

But v and v have one disadvantage


For better or
.

as against

4=

and c

they do not so readily suggest their mathematical analogues in


for

other algebras.

worse,

symbolic logicians have not


that of Ernst Schroder.

generally adopted v and v

Of the major contributions since Peirce, the

first is

In his Operationskreis des Logikkalkuls (1877), Schroder pointed out that the logical relations expressed in Boole s calculus by subtraction and divi
sion were all otherwise expressible,
of

meaning
has, first

as Peirce had already noted. The + given by Boole is abandoned in favor of that which it now introduced by Jevons. And the "law of duality", which con
,

nects theorems which involve the relation +

or +

and

1,

with corresponding
0, is

theorems
169

in

terms of the logical product


ff.

x, or

x and

emphasized.

See below, pp. 188

The Development of Symbolic Logic


(This parallelism of formulae had been noted by Peirce, in his but not emphasized or made use of.)
first

111

paper,

we know it today. This perfected and elaborated in Vorlesungen uber die Algebra der Logik (1890-95). Volume I of this work covers the algebra of classes;
resulting system
is

The

the algebra of logic as

system

is

Volume

II the algebra of propositions;

and Volume III

is

devoted to the

calculus of relations.

The algebra
is

of classes, or as

we

shall call

it,

the Boole-Schroder algebra,

170 the system developed in the next chapter. rated the theory of functions, but in all essential

We

have somewhat elabo

respects,

we

give the algebra

There are two differences of some importance between Schroder s procedure and the one we have adopted. Schroder s assumptions are in terms of the relation of subsumption, c instead of the
it

as

appears

in Schroder.

relations of logical product

and =; which appear in our postulates. And, second, Schroder gives and discusses the various methods of his predecessors,

as well as those characteristically his own.

The calculus of propositions (Aussagenkalkid) is the extension of the Boole-Schroder algebra to propositions by a method which differs little from that adopted in Chapter IV, Section I, of this book. The
Peirce
s

discussion of relations

is

based upon the work of Peirce.

But

methods are much more


is

the scope of the calculus

precisely formulated by Schroder, and much extended. We summarize the funda

mental propositions which Schroder gives for the sake of comparison both with Peirce and with the procedure we shall adopt in Sections II and III
of

Chapter IV.
1)

A, B, C, D,

symbolize

"elements"

or individuals. 171

These

are distinct from one another

and from

0.

symbolizes the universe of individuals or the universe of discourse of


first

the

order.
j, k,
/,

3)

2,

m,

n, p, q

represent any one of the elements A, B, C, D, ...

of

4)
70

Sri

For an excellent summary by Schroder, see Abriss der Algebra der Logik ; ed. Dr. Eugen Miillor, 1909-10. Parts i and n, covering Vols. i and n of Schroder s Vorlesungen, have so far appeared.
171

The

in, 3-42.

Many

propositions here noted will be found in Vorksungen uber die Algebra der Logik, others, and much discussion of theory, have been omitted.

112

A
5) i 6)
(i
:

Survey of Symbolic Logic


i

j represents

any two elements,

and j,

of I 1 in a determined order.

j)

(i

i),

(i 4= j)

*j

for every
7)

and

j.

i:j*0
I
1

Pairs of elements of

may

be arranged in a

"block":

A:A, A:B, A:C, A:D,


B:A, B:B, B:C, B:D,
C:A, C:B, C
:

...

...
...

C,

D,

:A,

D:B, D:C, D:D,

These are the

"individual

binary
:

relatives".

(A
+ (B + (C

A) + (A A) + (B A) + (C
:

B) + (A
B) + (B
:

C) + C) + C) +
.
.

B) + (C

represents the universe of binary relatives.


10)
I
2

S,-S,- (i

j)

= S -S,t

(i

:j)

= S

iy

(i

j)

9)

and

10)

may
1

be summarized in a simpler notation:

Ziji :j =

A :A+A :B + A :C+
+

...

B :A
C :A

+
+

B :B + B C+
:

11)

C :B

C+

...

+
12)
13)
i

........
"individual

:j

h will symbolize an
S^SyS,(i
:

ternary

relative".

A)

= 2 A t:j:A

Various types of ternary relatives are


14)

A A
:

A,

B A
:

A,

A B
:

A,

A A
:

B,

A B C
:
:

It is

obvious that we

may

similarly define individual relatives of the


order.

fourth, fifth,

...

or

any thinkable

The Development of Symbolic Logic

113

The

general form of a binary relative,

a, is

= Sy an
is 1

(i

j) j) pairs in

where a i;

is

a coefficient whose value


j,

for those

(i

which

has

the relation a to
1

and

is

otherwise

0.

==

Ziii-.j

the null class of individual binary relatives.

/=
JV
(a b)ij

S -/(i =j)(i:j) = 2i(i:i)


t

172

S,- y

(i=t=

j)(i:j)
(a

aij bij
-a.-/

6) i?

= a, =

&,-/

(-a),-,-

-(a,y)

(a &);y
1

= 2A

a,-*

6 A/

(a t 6),v

n* (a* +

fe*,-)

general laws which govern propositional functions, or Aussagenschemata, such as (ab) iit 2 h a ih b hi U h (a ih + b hj ), II a a i} S a a,-y, etc., are as
, ,

The

follows

Au
case,

symbolizes any statement about u\


in case,

and only
is

Au =

for every

there
"A

at least one
u"

for every

u such that A u == and 2 U A U means A u for some


"

U U A U will have the value 1 in u; 2 U A U will have the value 1 if 1. That is to say, U U A U means
u".

a)
j8)

U u A u cA c2 u A u K U A U = A U V A U1
v V

-[2uA u ]c-A v c-[U u A u

2UA U
0,

= A

+ 2UA U

(The subscript
7)
d)

u, in

a and
-/i u
,

represents

any value
]

of the variable u.)

-[n u ,4 u

= 2U

If ^4 M is

independent of
)

u,

then

-[s^ = n u -A* U U A U = A, and S u .4 =


tt

A.

e)

r)

U u (AcB u = (Acn u B u ), n, or n M n,(^ M c5 =


r
i;

U u (A u cB) =
(2 u ^l M

(2 u A u

cB)

cn 5
y

v)

77)

2 u (A u cB) = (n u A u cB), S
I
f

2 u (AcB u ) =
(H u A u
i),

(Ac2 u B u
= =

0)

or 2 u 2 v (A u
(.-i

cB

v)

= =

c2,B

9)

nw

=
=

i)

(n u .4

n u (,i u =
S B (.4
==

0)
1)
;

= =
(a
|

(z M ,i M
(S
tt

o)
1)

172

\2u(A u 0) We write I_ where


-a for a;
<-a

= (n u /l u =
Schroder has

0),
1
;

.4

where he has

6) for (a; 6); (a

6)

for (a j 6);

for a.

114

A
H u (A u cB u )c]
\

Survey of Symbolic Logic

(U U A U c

UUBU

K)

(S.4

tt

c ZU

(The reader should note that


Principia Mathematica,
X)
(x).<px

U U (A U ^B U
o\f/x.)

is

"formal

implication",

in

2 UBU = SM

AB
= S

U)

A + U U B U = U U (A + B u
Au B
V)

//)

(2 U A U )(2 V B V )
yl

ttf

II U

A U + U B = n,
V
V

(A U +
)

v)

nu

= nu A Bu

A
,

+ 2UBU
,

= ? U (A

Bu

(U U A U )(U V B V )

= UU

A U B = UU A U BU
V

o)

2 u n^4 M

Ii v

Su

Au

From

these fundamental propositions, the whole theory of relations

is

developed.

Though Schroder carries this


s

much

further than Peirce, the

general outlines are those of Peirce


esting of the
in the

new items

of

Schroder

calculus. Perhaps the most inter treatment are the use of "matrices"
relatives,

form of the two-dimensional array of individual binary


of the calculus of relatives to

and the application


"chains
",

Dedekind

theory of

as contained in

Was

sind und was sollen die Zahlen.

Notable contributions to the Boole-Schroder algebra were made by

Anton Poretsky
logiques

in his three papers, Sept lois fondamentales de la theorie

des egalites logiques (1899), Quelques lois ulterieures de la theorie des egalites
(1901),

and Theorie des

non-egalites logiques

(1904).

(With
will

his

earlier works, published in Russian, 1881-87,

we

are not familiar.)


of

Poret-

sky

Law

of

Forms,

Law

of Consequences,

and Law

Causes

be

II. As Couturat notes, Schroder had been influenced overmuch by the analogies of the algebra of logic to other algebras, and these papers by Poretsky outline an entirely different procedure which, though based on the same fundamental principles, is somewhat more

given in Chapter

"natural" to logic. Poretsky s method is the perfection of that type of procedure adopted by Jevons and characteristic of the use of the Venn

diagrams.

The work

of Frege,

though
its

intrinsically important, has its historical

interest largely

through

influence

upon Mr. Bertrand

Russell.

Although

the Begrifsschrift (1879) and the Grundlagen der Arithmetik (1884) both

The Development of Symbolic Logic


precede Schroder
there; and
laire is
s

115

hardly more than mentioned his influence upon Peano and other contributors to the Formusurprisingly small when one considers how closely their task is re
Vorlesungen, Frege
is

lated to his.

Frege

is

but, in thorough

German

concerned explicitly with the logic of mathematics fashion, he pursues his analyses more and more

a development of arithmetic of unprecedented but a more or less complete treatise of the logico-metaphysical problems rigor

deeply until

we have not only

concerning the nature of number, the objectivity of concepts, the relations


of concepts, symbols,
his

and

objects,

and many other

subtleties.

In a sense,

fundamental problem is the Kantian one of the nature of the judgments involved in mathematical demonstration. Judgments are analytic, de pending solely upon logical principles and definitions, or they are synthetic.
His
thesis, that

mathematics can be developed wholly by analytic judg


logical,
is

ments from premises which are purely


Russell
s

likewise the thesis of

Principles of Mathematics.

metik, like

Grundgesetee der ArithFrege Principia Mathematica, undertakes to establish this thesis for
s

And

arithmetic

by producing the required development.


s

Besides the precision of notation and analysis, Frege

work

is

important
is

as being the first in which the nature of rigorous demonstration


ficiently understood.

suf

His proofs proceed almost exclusively by substitu

tion for variables of values of those variables,

and the substitution


is

of defined

it is notation, Frege almost diagrammatic, occupying unnecessary space and carrying the eye here and there in a way which militates against easy understanding. It is probably this forbidding character of his medium, combined with the

equivalents.

it

must be admitted

against him:

unprecedented demands upon the reader s logical subtlety, which accounts for the neglect which his writings so long suffered. But for this, the revival
of logistic proper

Frege

might have taken place ten years earlier, and dated from Grundlagen rather than Peano s Formulaire.
publication, beginning in 1894, of

The

Peano

Formulaire de Mathelogic.

matiques marks a

new epoch

in the history of

symbolic

Heretofore,

the investigation had generally been carried on from an interest in exact


logic

and

its possibilities, until,


it

as Schroder remarks,

we had an elaborated
his collaborators, the

instrument and nothing for


situation
of
:

to do.

With Peano and

is reversed symbolic logic is investigated only as the instrument mathematical proof. As Peano puts it: 173 The laws of logic contained in what follows have generally been found
"

173

Formulaire,

(1901), 9.

116
in the

Survey of Symbolic Logic


rules, the

by formulating,
in

form of

deductions which one comes upon

mathematical

demonstrations."

The immediate
rate

result of this altered point of

view

is

new

logic,

no

less elaborate than the old

destined, in fact, to

become much more elabo

but with

its

elaboration determined not from abstract logical con

siderations or
of
in

by any mathematical prettiness, but solely by the criterion De Morgan had said that algebraists and geometers live application.
"a

seems to have required the mathe matical intent to complete the rescue of logic from its traditional inanities. The outstanding differences of the logic of Peano from that of Peirce
higher realm of
syllogism":

it

and Schroder are somewhat as


(1)

follows:

Careful enunciation of definitions and postulates, and of possible


postulates,

alternative

marking an increased emphasis upon rigorous


of a

deductive procedure in the development of the system.


(2)

The prominence

new

relation,

e,

the relation of a

member

of a

class to the class.


(3)

The prominence
implication"

of the idea of a prepositional function


"

and

of

"formal

and

formal

equivalence",

as against

"material

implication"

and

"material equivalence".
"existence"

(4)

Recognition of the importance of

and

of the properties

of classes,
(5)

members of classes, and so on, with reference to their "existence". The properties of relations in general are not studied, and "relative

addition"

in

does not appear at all, but various special relations, prominent mathematics, are treated of.

The disappearance
gain.
(6)
its

of the idea of relation in general is a real loss, not a

The

increasing use of substitution (for a variable of

some value

in

range) as the operation which gives proof.

We
The

here recognize those characteristics of symbolic logic which have

since been increasingly emphasized.

publication of Principia Mathematica would seem to have deter


of further investigation to follow that general direction

mined the direction


indicated by the

The Principia is con cerned with the same topics and from the same point of view. But we see here a recognition of difficulties not suggested in the Formulaire, a deeper
of Frege

work

and the Formulaire.

and more lengthy analysis of concepts and a corresponding complexity of procedure. There is also more attention to the details of a rigorous

method
174

of proof.

All these belong also to the Logica

Mathematica of C. Burali Forti (Milan, 1894).

The Development of Symbolic Logic

117

is

The method by which the mathematical logic of Principia Mathematica developed will be discussed, so far as we can discuss it, in the concluding

We shall be especially concerned to point out the sometimes lost sight of, between it and the older logic of Peirce connection, and Schroder. And the use of this logic as an instrument of mathematical
section of Chapter IV.

analysis will be a topic in the concluding chapter.

CHAPTER
THE
I.

II

CLASSIC,

OR BOOLE-SCHRODER, ALGEBRA OF LOGIC


THE POSTULATES AND

GENERAL CHARACTER OF THE ALGEBRA.


THEIR INTERPRETATION

The algebra of logic, in its generally accepted form, is hardly old enough warrant the epithet "classic". It was founded by Boole and given its to present form by Schroder, who incorporated into it certain emendations
which Jevons had proposed and certain additions particularly the relation or implies" which Peirce had made to Boole s system. contained
"

"is

in"

It is

sound judgment that the result is still an algebra, simpler yet more powerful than Boole s calculus. Jevons, in simplifying Boole s system, destroyed its mathematical form; Peirce, retaining the

due to Schroder

mathematical form, complicated instead of simplifying the original calculus. Since the publication of Schroder s Vorlesungen uber die Algebra der Logik
certain additions
of

and improved methods have been offered, the most notable which are contained in the studies of Poretsky and in Whitehead s Uni

versal Algebra. 1

But
this

if

the term
it

"classic"

is

inappropriate at present,

still

we may
out,
all

venture to use

by way

of prophecy.

As Whitehead has pointed


"algebra",

system

is

a distinct species of the genus


its

differing

from

other algebras so far discovered by

non-numerical character.

It is

certainly the simplest mathematical system with any wide range of useful
applications,

and there are indications that


of

it

will serve as the

parent stem

from which other calculuses


eral

such have appeared.

an important type will grow. Already sev The term "classic" will also serve to distinguish

the Boole-Schroder Algebra from various other calculuses of logic.

Some

of these, like the system of Mrs. Ladd-Franklin, differ through the use of other relations than +
1
,

and =

and are otherwise equivalent

versal Algebra,

For Poretsky s Bk. n.

See Whitehead s Uni studies, see Bibliography; also p. 114 above. Whitehead introduced a theory of "discriminants" and a treatment

of existential propositions by means of umbral letters. This last, though most ingenious interesting, seems to me rather too complicated for use; and I have not made use of "discriminants preferring to accomplish similar results by a somewhat extended study of

and

",

the coefficients in functions.

118

The Classic, or Boole-Schroder, Algebra of Logic


that
is

119

to say, with a

"

of these systems

may

dictionary" of equivalent expressions, any theorem be translated into a theorem of the Boole-Schroder

Algebra, and vice versa.

they go, but partial.


Mathematica,
are
finally, there are

Others are mathematically equivalent as far as some, like the calculus of classes in Principia logically but not mathematically equivalent. And,

And

systems such as that of Mr. MacColl s Symbolic Logic which are neither mathematically nor logically equivalent. Postulates for the classic algebra have been given by Huntington,
the Abriss}, by Del Re, by Sheffer and by Bernstein. 2 set here adopted represents a modification of Huntington s third
(in

by Schroder

The
set.*

It

has been chosen not so

much

for

economy of assumption

as for

"

natural

ness"

and obviousness.
Postulated:

A
1-1

class
If

K of elements
b are
b.

a, b, c, etc.,

and a

relation
b is

x such that:
in

a and

elements in K, then a x

an element

K, uniquely

determined by a and
1

For any element

a,

a xa

=
6,

a.

1 1
1

For any elements a and

a x
c,

x
x

a.

4
5

For any elements

a, b,

and

a x

(b

c)

(a x 6)

x c.

There
a.

is

a unique element,

0, in

such that a x

for every ele

ment
1

For every element


1-61
If

a,

there

is

an element,

-a, such that


x,

x x-a y xa

0,

then x xa

and 1-62

If

y and y x -a

y,

then y

0.

The element
These
1-7 1-8

and the

relations +
:

and c do not appear

in the above.

may 1 = -0

be defined as follows
Def.
Def.

o+ 6 a cb
It

= -(-ax-6)
is

1-9

equivalent to a x6

a
is

Def.

remains to be proved that -a


it

which

will follow that


b.

is

uniquely determined by a, from unique and that a + b is uniquely determined

by a and
2 3

See Bibl. See "Sets of Independent Postulates for the Algebra of Logic", Trans. Amer. Math. Our set is got by replacing + in Huntington s set by x and Soc., v (1904), 288-309. Thus 1 can be replacing the second half of G, which involves 1, by its analogue with 0. omitted. Postulate J is not strictly necessary. defined, and postulates E and
,

120

A
The

Survey of Symbolic Logic

sign of equality in the

above has
if

its

usual mathematical meaning;


<p(x)

j.

e .,

=
j j

is

a relation such that


<p(x)

y and

function in the system, then


interchangeable.
in the system,
It follows

and

<p(y)

an unambiguous are equivalent expressions and


is

from this that

if

$(x)

is

an ambiguous function

terms of
that -a,

x,

is

every determined value of ifr(x), expressible in Suppose, for example, similarly expressible in terms of y.
y,

and x

"negative

of

a",

is

an ambiguous function
"negative

of a.

Still

we may

write

-a to mean, not the function

of

a"

itself,

but to mean some

(any) determined value of that function

any one

of the negatives of a

and

if

-a

b,

then
is

<p(-a)

and

<p(b)

will

be equivalent and interchangeable.

This principle

We
terms,

shall

important in the early theorems which involve negatives. develop the algebra as an abstract mathematical system: the

a, b, c, etc.,
,

and x

and c

may be any entities which have the postulated properties, may be any relations consistent with the postulates.
we
give

But

for the reader s convenience,


all,

two

possible applications:

(1) to

the system of

continuous and discontinuous, regions in a plane, the

null-region included,

and

(2) to

the logic of classes. 4

(1)

For the

first

interpretation, a x b will denote the region

common

to a

and
a

b (their
is

overlapping portion or portions),

and a +

b will

denote that
"Region

which
is

either a or 6 or both,

a c b will represent the proposition,


b

completely contained in region


+

(with or without

remainder)".
1

will
itself,

represent the null-region, contained in every region, and


or the
will
"sum"
{

the plane

of all the regions in the plane.


a, all

For any region

a,

-a

be the plane except

that

is

not-a.

The

postulates will then hold

as follows:
1

1
b,

If
is

a and b are regions in the plane, the region


in the plane.
If

common

to a
is

and

b,

a x

a and 6 do not overlap, then a x b

the null-

region, 0.
1

2
3

For any region

a,

the region
to a

common

to a

and

a,

a x

a, is

itself.

1 1

The
The
c
is

region region

common

and

b is the region
is

common

to b

and

a.

and
1-5
1

the

common region common


common
a,

to a and b x c

the region

common

to a x b

to all three.

The

region

to

any region a and the


is its

null-region, 0,

is 0.

6
4

For every region


Both

there

negative, -a, the region outside or

of these interpretations are

more

fully discussed in the next chapter.

The Classic, or Boole-Schroder, Algebra of Logic


not contained in
1-61
a,

121

and
If

this region

is

such that

common, then the region


and 1-62
in a,
If

-a and any region x have only the null-region in common to x and a is x itself, or x is contained in a;

the region

common

to y

and a

is y,

or y

is

contained
-a, then

and the region common to y and -a is y, or y is contained in must be the null-region which is contained in every region.

That the

definitions, 1-7, 1-8,

and

1-9, hold, will be evident.

(2)

For the second interpretation,


in extension

a, b, c, etc., will

be logical classes, taken

that

is,

b will

mean
and

that a and b are classes composed of

identically the

same members,

a x b will represent the class of those


of b both;

things which are


are either
tion

members

of a

a+

b,

those things which

members of a or members of b or both, that all members of a are also members of b,


is is

a c b will be the proposi or that a is contained in b

(with or without remainder).

the null-class or class of no members;


is
r

and the convention


1 is

required that this class


discourse"

contained in every

class.

the

"universe

of

or the class

w hich contains every

entity

in the system.

of all

For any class a, -a represents the negative of a, or the class The postulates will hold as fol things which are not members of a.

lows
1-1

If

a and b are logical classes, taken in extension, the

members com

mon
in

to a

and

b constitute a logical class.


is

In case a and b have no members

common,

this class

the null-class,
to a

0.

1-2
1

The members common The members common


a.

and a constitute the and


b are the

class a itself.

to a

same as those common to

b
1

and
-4

The members common to a, b, and c, all three, are the same, whether we first find the members common to b and c and then those common to a and this class, or w hether we first find the common members of a and b
r

and then those common to


1

this class

and

c.

The members common


For every
"

to

any

class a

and the

null-class are none, or

the null-class.
1

class a, there

is its

negative, -a, constituted by


a,

all

members

of the

universe of

discourse"

not contained in

and such that:


in

1-61

If

-a and any class x have no members

common,

122
then

A
all

Survey of Symbolic Logic

members
and 1-62

of

x are

common

to x
of

and

a,

or x

is

contained in a;
a,

If all

members

and common
1-7

also to

y and -a,
of

any class y are common to y and then y must be null.


"everything",

The

"universe

discourse",

is

the negative of the

null-class,
1

"nothing".

That which
is
"a

is

either a or 6 or both

is

identical with the negative of

that which
1-9

both not-a and not-6.


"

That

is

contained in b
of a

is

equivalent to
b
".

"The

class a is identical

with the

common members

and

proved by these interpretations. In the form given, they are not independent, but they may easily be made so by certain alternations in the form of statement. 5
is

That the postulates are consistent

The

following abbreviations and conventions will be used in the state


of theorems:

ment and proof


1.

ax6

will generally

be abbreviated to a b or a -b,

ax(bxc)

to a (be),

ax-(6x-c)
2.

to a-(fc-c) or a--(&-c), etc.

In proofs,

we

shall

sometimes mark a lemma which has been established


in that proof refer to the
"Q.E.D."

as (1), or (2), etc.,

and thereafter
shall

lemma by

this

number.

Also,

we

sometimes write

instead of repeating

the theorem to be proved.


3.

The

principles (postulates, definitions, or previous theorems)

by which

any

step in proof is
If

taken
x

will usually

brackets, thus:

0,

be noted by a reference in square then [1-5] a x = 0. Reference to principles

whose use

is

more

Theorems
insertion

will

proceed. be numbered decimally, for greater convenience in the

or less obvious will gradually be omitted as

we

of

theorems without alteration of other numbers.

The non-

decimal part of the number will indicate some major division of theorems, as 1- indicates a postulate or definition. Theorems which have this digit

and the one immediately following the decimal point in common different forms of the same principle or otherwise closely related.
II.

will

be

ELEMENTARY THEOREMS
b c

2-1

If a

6,

then a

and

c b.

This follows immediately from the meaning of


v

= and

1.

2 2

b is equivalent to the pair, a


If

c b and

a.
b.

6,

then [1-2]

ab =

a and b a
s set.

On

this point,

compare with Huntington

The

Classic, or Boole-Schroder, Algebra of Logic

123

And
But

if

a b

=
a b

a and b a

b,

then

[1 -3]

a b

[1 -9]
is,

is

equivalent to a c b and b

= b a = b. a = b to 6 c a.

Equality
2-3

then, a reciprocal inclusion relation.

aca.
a

a,

hence [2-2] Q.E.D.


is "contained
in"

Every element
2-4
a -a

itself.

=
[1-2]

= -a a. a a = a. = (a -a) a. = (a -a) -a. (-a -a) (a -a) a = (a -a) -a = a -a, then a -a = 0. a = a -a. Hence also, -a a = 0.
-a
(a a)

Hence

[2-1, 1-4, 1 -3] a

-a

a (a -a)

Also [1-2] -a -a

-a.

Hence a -a = a

But

[1-62]
[1-3]

if

And

-a

Thus the product of any element modulus of the operation x


.

into its negative

is

0,

and

is

the

2-5

a -b

=
If

is

equivalent to a b

a and to a c

b.

=
if

a,

then

[1-4-5,. 2-1-4]

a -b

(a b) -6

(b -b)

And

[1-61]

a -6

=
is

0,

then

By (1) and And [1 9] a

(2),

a -b a

=
a c

ab = a and ab = a
c6

(2)

are equivalent.

6.

We

shall derive other equivalents of a

later.

The above
"1-4-5"

is

required
"2-1-4"

immediately.
instead of

In this proof,
1-5"

we have
"2-1,

written

and

"1-4,

and

2-4".

This kind of abbreviation in

references will be continued.

2-6

If

acO, then a =
If

0.

a cO, then [1-9] o-O


b,

=
c

a.

But
c b.

[1-5]

a-0

0.

2-7

If

a c

then a

c,

and

a c

If acb, then But [1-2-3-4] (a

[1-9] a b
b) c

a and [2-1]
(ac)

b) c

=
Hence, by

(ac)(cb)

= (b a) c = b = (ac)(bc)
a c
c b
6,

= (ac)
c)

= ac = [a (c c)

(1)
b]

= [(ac)
a
c

c]

(2)

(1)

and

(2), if

then (a c)(b
b c.

= ac and
also c a

[1 -9]
b.

b c.

And
2-8
-(-a)

[1-3] c a

= ac and
-

Hence

cc

a.

[2-4] -(-a) --a

0.

Hence
Hence

[2-5] -(-a)

ca
ca-a.

(1)

By

(1),

-[-(-)] c-a.

[2-7] a- -[-(-a)]

124

A
But
[2-4] a

Survey of Symbolic Logic

-a

0.

Hence a --[-(-a)] cO.

Hence

[2-G] a --[-(-a)]

and

[2-5] a

c-(-a)

(2)

[2-2] (1)

and

(2) are

equivalent to -(-a)

a.

3-1

a cb

is

equivalent to -b c -a.

[2-5]

ac6

is

equivalent to a -6

0.

And And
If

[2-8]

a-b = -ba = -b-(-a).

-6 -(-a)
of

is

equivalent to -b c-a.

be transposed by negating both. then the portion of the plane not in b is contained in the portion of the plane not in a: if all a s are 6 s, all non-6 s are non-a s. This theorem gives immediately, by 2-8, the two corollaries:

The terms
is

any

relation

may

region a

contained in region

b,

3-12 3-13
3-2
a

a c -b

is

equivalent to b c-a; and equivalent to -6 c


a.

-a c

is

b is

equivalent to -a

-6.

[2-2] a [3-1]

=
is

b is

equivalent to (a

cb and bed).

cb

equivalent to -b c-a, and b

ca

to -a c-6.
is

Hence a =
lent to

b is

equivalent to (-a
-6.

c-b and -6 c-a), which

equiva

-a

The negatives
3-22
a

of equals are equals.

By

2-8,

we have

also

-b

is

equivalent to -a

b.

Postulate 1-6 does not require that the function

"negative

of"

be

There might be more than one element in the system having the properties postulated of -a. Hence in the preceding theorems, -a
unambiguous.

must be read
proved of -a,
in

"any

negative of a

",

-(-6)
6,

the negatives of any given negative of


etc.,

and

must be regarded as any one of so on. Thus what has been


But we can now demonstrate

has been proved to hold for every element related to a


postulate.
is

the

manner required by the

that for every element a there


properties postulated of -a.

one and only one element having the

3-3

-a

is

uniquely determined by

a.

By
such.

1-6, there
is

is

at least one element -a for every element a.


let

Suppose there

more than one:

-a x and -a 2 represent any two


[3-2] -a,

Then
Since
all

[2-8] -(- Gl )

-(-a 2 ).

Hence

-a 2

the relation x

functions in the algebra are expressible in terms of the negative, and 0, while is unique and a x b
,

a, b, c, etc.,
is

uniquely

The Classic, or Boole-Schroder, Algebra of Logic

125

determined by a and b, it follows from 3-3 that all functions in the algebra are unambiguously determined when the elements involved are specified.
(This would not be true
"division"

if

the inverse operations of

"subtraction"

and

were admitted.)
1 is

3-33

The element
[1-5]
is

unique.
is

unique, hence [3-3] -0

unique, and [1-7]

-0.

3-34

-1

0.

[1-7] 1

-0.

Hence

[3-2]

Q.E.D.
is

3-35

If

a and b are elements in K, a + 6


b.

an element

in

uniquely deter

mined by a and

The theorem
3-37
If

follows from 3.3, 1-1,


c

and
c

8.

6,

then a +

and

+a

b.

The theorem
,

follows from 3-35

and the meaning

of

=.

3-4

-(a +

6)

-a

-b. b

[1-8] a +

Hence
.3-41

[3-3, 2-8] -(a

-(-a -b). + b)

-[-(-a-b)}

-a -b.

-(a

6)

-a +

-6.

[1-8, 2-8]

-a + -6

-[-(-a) --(-6)]

-(a

6).

3-4 and 3-41 together state

De Morgan

sum

is

the product of the negatives of the

Theorem: The negative of a summands; and the negative of a


s

product is the sum of the negatives of its factors. The definition 1-8 is a form of this theorem. Still other forms follow at once from 3-4 and 3-41,

by 2-8:
3-42
3-43
-(-a + -b)
-(a + -6)

3-44
3-45

-(-a +

b)

= ab. = -ab. = a-b.


-a +
b.

-(a-b)
-(-ab)

= =

3-46

a + -b.
s

From De Morgan
3-34, -1

Theorem, together with the


equals",

principle, 3-2,
1

"The

negatives of equals are

the definition 1-7,

-0,

and theorem
x there
is

0, it

follows that for every theorem in terms of


in

corresponding theorem
replaced by
its

terms of +

If in

any theorem, each element be


result
is

negative,

and x and + be interchanged, the


can, of course, be replaced

valid theorem.

The negative terms

by

positive,

126
since

A
we can suppose x =
system there
is

Survey of Symbolic Logic


-a, y

-b, etc.

Thus

for every valid

theorem

in the

another got by interchanging the negatives and 1 and the symbols x and + This principle is called the Law of Duality. This law is to be illustrated immediately by deriving from the postulates
. .

their correlates in terms of +

The

correlate of

1 is

3 35, already proved.

4-2

a+a

a.

[1-2]

-a-a =

-a.

Hence

[1-8,

3-2,

2-8] a + a

-(-a -a)

-(-a)

a.

4-3

a + b

a.

[1-3]

-a -b

-b -a.

Hence

[3-2]

-(-a -6)

-(-6 -a).

Hence
4-4
a+
(b
\1

[1-8]
c)
4.1

Q.E.D.
(a
(

=
n

+
A

b)
^\

c.

h\

Hence

[3-2] -[-a (-b -c)}

-[(-a -6)

-c].

But

[3-46, 1-8] -[-a(-fc-c)] [3-45, 1-8] -[(-a-6)-c]


1.

And
4-5

= a + -(-6-c) = a+(6 + c). = -(-a-b)+c = (a + b)+c.

a+1 =
Hence

[1-5]

-a-0

0.

Hence

[3-2] -(-a-0)

[3-46] a +

-0

-0,

and
x.

[1-7]

= a+1 = = =

-0.
1.

4-61

If

-x + a
If

1,

then x a
1,

-x + a
[2-5]

then [3-2-34-44] x -a
is

-(-z + a)
x.

=-1=0.

And
4-612

x -a

equivalent to x a

If -a;

+a

1,

then x + a

= =
y,

a.

[4-61] If

-a + x
if

1,

then a x
1,

a,

and
a.
1.

[3-2]

-a + -x

-a

(1)

By
4-62
If

(1)

and 2-8,

-x + a

x+a

= =
a)

y+a
If

y and y + -a
y, [3-2]

= =

then y

And
But
4-8
a + -a

y+a = if y + -a
if

-y -a

y,

-y a

-(y + -(y + -a)

=
-y. -y,

-y.

[1-62]

-y a

-y and -y -a

= =

-y

and y = -0 =

1.

l=-a + a.
-a a

(Correlate of 2-4)

[2-4]

0.

Hence

[3-2] a

+ -a
is 1.

-(-a a)

=-0 =
all

1.

Thus the modulus


4-9 -a +
6

of the operation +

1,

a+ b

=
0,

b,

a -6

=
a,

0,

a,

and a c b are

equivalent.

[2 5]

a -b
6

a 6

and a cb are equivalent.

[3-2]

-a +

1 is

equivalent to a -b

-(-a +

b)

=-1=0.

-M,V)

The Classic, or Boole-Schroder, Algebra of Logic


[4-612] If -a + b

127

=
b,

1,

a+ b

b.

And

if

a+b

= =

[3

37] -a + b

-a +

(a

b)

(-a + a) + 6

= 1+6

1.

Hence a +

b is equivalent to

-a + b

1.

We
5-1
If

turn next to further principles which concern the relation c


a c 6 and b cc, then
[1-9] a
If

ace.

cb
a
is

is

equivalent to a b
c

a,

and
a

cc

to b c

=
a.

b.

a b

=
c

a and b

b,

(a b) c
c.

(b c)

a b

But a

equivalent to a c

the Syllogism.

This law of the transitivity of the relation c is called the Principle of It is usually included in any set of postulates for the algebra which are expressed in terms of the relation c
.

5-2

a b

ca and ab cb. = a (a b) = (a a) b = a b. (a b) a But (a b) a = a b is equivalent to ab ca. = a (b b) = a b, and ab Similarly, (a b) b


a ca + b
[5 2]

cb.

5-21

and

ca +

b.

-a -b c -a and -a -b c -b.

Hence

[3-12]

ac-(-a-b) and 6c-(-a-6).

But -(-a

-b)

(a

b).

Note that 5-2 and 5-21 are


general,

correlates

by the Law

of Duality.

In
,

having now deduced

the fundamental properties of both x and +


in

we

shall give further

theorems

such pairs.

A
5-22

corollary of 5-21 is:


a b

ca + b.
[5-

1-2- 21]
c

5-3

If

a c b and
[1-9] If

d,

then a
c

cb

d.

a c b and

cd, then a

b
c,

a and
c

d
d.

c.

Hence
5-31
If a

(a c) (b d)
b

=
d,

(a b) (c d)

=
c
b

and a

cb

and

then a +

d.

If

a c 6 and
[5-3] [1-8]

cd, [3-1] -b c -a and -d c -c.


[3-1]

Hence Hence

-b-dc-a-c, and
Q.E.D.
a and a + a

-(-a -c) c-(-fc-d).

By
5-32

the laws,
If

aa =

a,

5-3 and 5-31 give the corollaries:

a c

and

b cc,

then

ab

cc.

128

A
If

Survey of Symbolic Logic


c.

5-33

a c c and b c

c,

then a + 6 c
c.

5-34 5-35 5-37

If a If If

and ace, then a cb

a c 6 and ace, then a c 6 + a c6, then a +


c

c.

c6

c.

(Correlate of 2-7)

[2-3]cce.
5-4
a + ab =
a.

Hence

[5-31]

Q.E.D.

[5-21]
[2-3]

aca + ab
a,

(1)

a c

and

[5-2]
(2),

ab ca.

Hence

[5

33] a + a

ca

(2)

[2-2] If (1)

and
a.

then Q.E.D.

5-41

a (a +

b)

[5-4]

-a + -a-b

-a.

But

[3-4] -(-a +

--&) =

-(-a + -a-b) a --(-a -6) =0(0 + 6).


[3-2]

Hence

-(-a)

a.

Law of Absorption. We have next to prove the Distributive Law, which requires several lemmas.
5-4 and 5-41 are the two forms of the
5-5
a
(b

c)

= ab +

ac.

Lemma

1:

[5-2] [5-2] a b

ab + acca(b + c). a6ca and acca.


c
b

Hence

[5-33] a b + a c

ca
cb+
c.

(1)

and a
c
b

c
c

c.

But
6

[5-21] b
c.

and

Hence Hence

[5-1] a b
[5

and ac cb +
c
+
c

33] a b + a c

(2)

[5-34] If (1)

and
q

(2),

then a

+a

ca
is

(b

c).

Lemma

If

pc
-q.

is false,

then there

an element

x, 4= 0,

such that

x c p and x c

p-q
[4-9]
if

is

such an element, for [5-2]


if

p-qcp
q

and p-qc-q- and


then p -q
4= 0.

(This

p-q = 0, then pcq, hence lemma is introduced in order to


3
:

pc

is false,

simplify the proof of

Lemma

3.)

Lemma
4=

(6

c)

+a

c.

Suppose this 0, such that

false.

Then, by lemma

2,

there

is

an element

x,

xca(b + c)
.TC-(6 + ac) But [3-12] if xc-(b + ac), then b + acc-x [5-1] If (1), then since [5-2] o (6 + oc) co, z

(1)

and

(2)

co

(3) (4)
(5)

and
Also

also, since a(6

+ a

c)

c, .r

[5-1] If (2),
[5
1] if

then since [5-21] b cb + a c, b c -x and [3-12] xc-b (2), then since [5 21] a c c b + a c, a c c -x and [3-12]

zc-(oc)

(6)

The

Classic, or Boole-Schroder, Algebra of Logic

129

From

(6)

and
x c

(3), it
a,

follows that
[5

xcc must
But
if

be
c

false;

for

if

.TCC
c),

and

(3)

then

34]

x ca

c.

x ca

and

(6)

x c -(a

then [1-62] x

0,

But

if

x c

be

false,

which contradicts the hypothesis x 4= 0. then by lemma 2, there is an element y, 4=

0,

such that

ycx
and
[5-1] If (7)
If (8)
If (7)

(7)

?/c-c,
(5),

or [3-12]

and
(9),

and

and

(4),

then y c-b and [3-12] then [5-33] b + cc-y and [3-12] yc-(b + then [5-1] ycb + c

cc-y bc-y

(8) (9)
c)

(10)

(11)

[1-9] If (11),

But
y

if

(12),

then y(b + c) then [1-62] y

=y, and if = 0, which

(10),

y~(b +

c)

(12)

contradicts the condition,

*0.
false
is

Hence the supposition that a(b + c)cb + ac be supposition, and the lemma is established.

a false

Lemma

a(b + c)cab + ac. By lemma 3, a (b + c) c Hence [2-7] a [a (b + c)] c a


4
:

+ a

c.

(b

+ a

c)

But a [a (b + c)] = (a a)(b + c) = a (b + c). And a (b + a c) = a (a c + b). Hence a (b + c) c a

(a c

b).

But by lemma

3,

a (a c + b) c a c + a
c.

b.

And
Proof of

a c + a b

a b + a

Hence a
1

(b

c)

a b + a

c.

the theorem:

[2-2]

Lemma
c.

and lemma 4 are together equiva

lent to a (b + c)

a b + a

This method of proving the Distributive


"Sets

Law

is

taken from Huntington,


".

the long and

The proof of Algebra of Logic due to Peirce, who worked it out for his paper of 1880 but mislaid the sheets, and it was printed for the first time in
of

Independent Postulates
difficult

for the

lemma

is

Huntington
5-51
(a

paper.

+ b)(c + d) = (a c + [5 5] (a + b)(c + d)

b c)

(a

d+

b d).

(a +

b) c

(a

b)

(a c

+b

c)

(a

d + b d).

5-52

a + b c

(a + b)(a

c).

(Correlate of 5 5)

[5-51] (a +

b) (a +

c)

(a a + b a) + (a c + b c)

=
But
[5-4] (a + a
b)

[(a

+ a

b)

+ a

c]

c.

+ a

a + a c

a.

Hence Q.E.D.
follows:
300, footnote.

Further theorems which are often useful in working the algebra and

which follow readily from the preceding are as


6

See

"Sets

of

Independent Postulates,

etc.",

loc. rit., p.

10

130

A
a-1

Survey of Symbolic Logic

5-6

1 -a.

[1-5]

a-0
if

0.

Hence a--l =

0.

But
5-61

[1-01]

a--l

= =

0,

then a-1

a.

acl.
[1-9] Since a-1
a,

acl.

-a--0
5-63

-a-1

-a.

Hence

[3 2]

a+

-(-a--0)

-(-a)

a.

Oca.
0-a

a-0

0.

Hence

[1-9]

Q.E.D.

5-64

ca

is

equivalent to a

1.

[2-2] a

1 is 1

equivalent to the pair,

acl and

ca.

But
5-65
a c

[5-61] a c
is

holds always.

Hence Q.E.D.

equivalent to a

0.

[2 2]

is

equivalent to the pair, a c

and

a.

But
5-7
If

[5-63]
a;

ca

holds always.

Hence Q.E.D.
x.

a + 6
If

= and a = 0, then 6 = a = 0, a + 6 = + 6 = 6. =
x and a

5-71

If a 6

= =

1,

then 6

x.

If a

= =

1,

a 6

1-6

6.

5-72

a + 6
If

=
a

is

equivalent to the two equations, a

and

0.

and

=
=

0,

then a + 6

And
But

if
if

a+ 6

-a -6
if

= =

0,
1,

-a -6
a

-(a +

6)

= =

= =
1.

0.

-0

a-1

a(-a -6)

(a -a)

-6

0--6

0.

And
5-73
a 6

[5-7]
1 is

a+ 6

and a

0,

then 6

0.

=
If

equivalent to the two equations, a


1
-

and

1.

and 6 = 1, then a 6 = 1 1 = 1. = 1, -a + -6 = -(a 6) =-1=0. Hence and -6 = 0. But [3-2] if -a = 0, a == 1, and if -6 = 0, 6 = 1.


a
==

And

if

a 6

[5-72]

-a

5-7 and 5-72 are important theorems of the 5-7, "Any null algebra. term of a sum may be dropped", would hold in almost any system; but a sum is null, each of its summands is 5-72, is a special law null",
"If

characteristic of this algebra.

It is

tains no inverses with respect to

due to the fact that the system con and 0. a and -a are inverses with

The Classic, or Boole-Schroder, Algebra of Logic


respect to

131

x and

and with respect to + and


less useful.
a.

1.

5-71 and 5-73, the

correlates of 5-7

and 5-72, are


a b + a -b
(b

5-8

(b

+ -b)
[5-5]

=
a

+ -6)

=
1.

a b + a -b.

And
5-85
a+
b

[4-8] b

+ -b
b.

Hence a

(b

+ -b)

a-l

a.

a + -a

= ab + -ab. Hence a + b = a + (a b + -a b) = (a + a But [5-4] a + a b = a. Hence Q.E.D.


[5-8] b
It will

b)

+a

flb.

be convenient to have certain principles, already proved for two terms or three, in the more general form which they can be given by the
use of mathematical induction.

Where the method

of such extension is

obvious, proof will be omitted or indicated only.


are associative,

Since both

x and

we can

dispense with parentheses by the definitions:


Def.

5-901

a+ b+

c=(a + b)+c
Def.

5-902
5-91

abc = (ab) c a = a (b + -b) (c


[5-8]
1

+ -c) (d + -d)

5-92

=
a

(a

+ -a)(6 + -6)(c + -c)...


[4-8]

5-93

a+

ab + ac + ad+...
(a + d)

[5-4]

5 93 1

=
+

a (a + b) (a +
[5-41]

c)

5-94

(b

+ d + ...)
[5-5]

= ab + ac

ad+
.

...

5-941

a + bcd.,..
[5-52]

--=

(a

+ b)(a + c)(a + d)

5-95

-(a + b +
If

= -a

-b -c ...

the theorem hold for n terms, so that


-(ai + a 2 +
.

+ a n)

-a-i

-a z

-a n

then

it will

hold for n

terms, for by 3-4,

And

[3-4] the

theorem holds

for

two terms.

Hence

it

holds for

any

number

of terms.

132

A
-(a

Survey of Symbolic Logic


-c +

5-951

bed...) = -a + -b +
a+ b + c +

-d+

Similar proof, using 3-41.

5-96

+-a-b

-c

[4-8, 5-951]

5-97

a + b+

is

equivalent to the

set,

0, 6

0, c

0,

[5-72]

5-971

abed...
[5-73]

--=

1 is

equivalent to the

set,

==

1,

1, c

I,

5-98

a-b

[1-2] a a a a

=ab-ac-ad... = a.
.

5-981

a +

(b

+d+

)
.

=
. .

(a

b)

(a

c)

+ (a + d) +

[4-2] a + a + a +

a.
s

The

extension of

De Morgan

Theorem by 5-95 and 5-951

is

especially

important.

5-91, 5-92, and 5-93 are

different forms of the principle

by

which any function


whatever

may

be expanded into a

originally involved in the function

sum and any elements not introduced into it. Thus any expression

be regarded as a function of any given elements, even though in the expression, a peculiarity of the algebra. 5-92, the expression of the universe of discourse in any desired terms, or expansion

may

they do not appear


the basis of

of

1, is

many important

procedures.
if

The theorems 5-91-5-981


involved be
induction.
III.
finite, since

are valid only

the

number

of elements

proof depends upon the principle of mathematical

GENERAL PROPERTIES OF FUNCTIONS


etc., to

We may
only

use/Or), $(x, y),


of the class

members

denote any expression which involves and the relations x and + The further
.

requirement that the expression represented by /(.r) should involve x or its negative, -x, that $0, y) should involve x or -.r and y or -y, is unnecessary, for if .r and -x do not appear in a given expression, there is an equivalent
expression in which they do appear.
a

By

5-91,
-?/)

a (x +

-.r)

a x + a

-x

(a

x + a -x) (y +

= axy + ax-y + a-xy +


a x + a -x
to x.

a -x -y, etc.
reference
reference

may be called the expansion, or development, of a with And any or all terms of a function may be expanded with
and
so
of elements.

to x, the result expanded with reference to y,

on for any elements


ele-

and any number

Hence any expression involving only

The Classic, or Boole-Schroder, Algebra of Logic

133

ments

in

and the

relations

x and +

may

be treated as a function of

any elements whatever. If we speak of any a such that x = a


of x being given, the value of

as the

"value

of x

",

then a value

as in

any

other.

any function of x is determined, in this algebra But functions of .r in this system are of two types: (1)

those whose value remains constant, however the value of x

may vary, and those such that any value of the function being assigned, the value of x (2)
thereby determined, within limits or completely. Any function which symmetrical with respect to x and -x will belong to the first of these
in general, a function

is
is

classes;

which

is

not completely symmetrical with

respect to x

and -x

will

belong to the second.

But

it

must be remembered,

in this connection, that a

unless

it

symmetrical function may not look symmetrical be completely expanded with reference to each of the elements

involved.

For example,
a + -a b + -b

is

symmetrical with respect to a and -a and with respect to panding the first and last terms, we have
a
(b

and

-b.

Ex

+ -b) + -

(a

+ -a) -b
b.
it

a b + a -b + -a b + -a -b

whatever the value of a or of

Any
is

function in which an element,

.r,

does not appear, but into which

introduced by expanding, will be

symmetrical with respect to x and -x. The decision what elements a given expression shall be considered a
function of
is,

in this algebra, quite arbitrary except so far as

it

is

deter

mined by the form


and
"variables"

of result desired.
"unknowns"

The

distinction between coefficients

or

is

not fundamental.

In fact, we shall

frequently find

it

say

of x

and

y,

convenient to treat a given expression first as a functionthen as a function of z, or of x alone. In general, coef

ficients will

be designated by capital letters.


of

The Normal Form


can be given the form

a Function.

Any

function of one variable,

/(.r),

A
where

x+

-x
is

and

are independent of x.

This

the normal form of functions

of one variable.

Any function

of

one variable, /(.r),


.r,

is

such that, for some

and some

which are independent of

/(.r)

= A

x +

-x

134

A
Any

Survey of Symbolic Logic

the relations

and expression which involves only elements in the class x and + will consist either of a single term a single
or of a

element, or elements related by x

sum

of such terms.

Only
x,

four kinds of such terms are possible:


(2)

(1)

those which involve

those which involve -x, (3) those which involve both, and (4)

those which involve neither. 7 Since the Distributive Law, 5


of x, of -x
5,

allows us to collect the coefficients


of such

and

of (x -x), the

most general form


(x -x)

an expres

sion

is

px + q -x + r
where
p,
q, r,

and

s are

independent of x and -x.


0.

But

[2-4]

And

[5-9]

r(x-x) = r-0 s = s x + s -x.


+ q -x +
r (x

Hence

px

-x) + s
q +

=
s,

(p + s) x + (q + s) -x.

Therefore,

A = p + s, B =

gives the required reduction.

The normal form

of a function of
C2,

+
Xn
,

variables,
+!

may

be defined as the expansion by the Distributive


f(Xi,

Law

of

X2

...

X n ) -X n +i +/

(.Tl, .T 2 ,

X n ) -X n+ i
x*>,

where/ and/
in the

are each

some function
is

of the n variables, Xi,


step"

...

x n and
,

normal form.

This

"step

by

definition; the normal form

of a function of

two variables

is

defined in terms of the normal form of

functions of one variable; the normal form of a function of three variables


in

terms of the normal form for two, and so on. 8

Thus the normal form

of a function of

two

variables, $(x, y), will be

found by expanding
-y

(A x +
It will be,

-x)

y+(Cx + D -x)

Axy + B -x y + C x -y + D -x -y
of a function of three variables, ty(x, y, z), will be

The normal form

Axyz + B-xyz + Cx-yz + D-x-yz + Exy-z + F-xy-z


+

G x -y

-z + II -x -y -2

And
7

so on.

Any

function in the normal form will be fully developed with


"involves"

By

a term which

is

meant a term which


in

either is x or has
is

factor".

But

"factor"

seems inappropriate

an algebra

in

which h x

x a always contained
"as

in x, h
8

x ex.

This definition alters somewhat the usual order of terms in the normal form of func tions. But it enables us to apply mathematical induction and thus prove theorems of a generality not otherwise to be attained.

The Classic, or Boole-Schroder, Algebra of Logic


reference to each of the variables involved

135

that

is,

each variable, or

its

negative, will

appear

in

every term.
be given the normal form.
of

6-11

Any

function

may
1
,

(a)

By
If

(>

any function

one variable

may

be given the

normal form.
(6)

functions of n variables can be given the normal form,

then functions of n

1
,

variables can be given the normal form, for,

Let $(1,

Xz, ...

x n Xn+i) be any function of n


will

variables.

By

definition, its

normal form

be equivalent to
(Xi,

f(Xi, Xz, ...

X n ) Xn+l +f

XZ
,

...

X n ) -Xn+l

wLere / and /
form.

are functions of x\, x 2

...

x n and in the normal

By

the definition of a function, $(#1, x 2)

x n x n +i)
,

may

be re

garded as a function of x n +i.

Hence, by
i

6-1, for

some

and some

which are independent


$(Xi, Xz,
. . .

of x n ^
,

X n Xn+l)

= A

X n+ i +

-X n +l
7

Also,

by the

definition of a function, for

some/ and some/


x n)
.

A = f(xi,
and
Hence, for some / and /
&(Xi, X 2
,
. . .

x2

...
.
.

B =

(xi, Xz,

xn)
of x n +\

which are independent


f(Xi,
-To,
.
. .

X n Xn+l)
,

X n ) X n +l +/
(.Ti,

X Zt

X n )Xn+l

Therefore,

if

the functions of n variables, / and /


. . .
>

can be given the

x n x n +\) can be given the normal form. normal form, then $(0*1, Xz, be given the normal form, (c) Since functions of one variable can

and

since

if

functions of n variables can be given the normal form,

functions of n

variables can be given the normal form, therefore


of variables

functions of any

number

can be given the normal form.

The second
is

step, (6), in the

above proof

may seem
,
.

arbitrary.

That

it

valid, is

due to the nature of functions


of

in this algebra.
. -

x n ), the normal form will n variables, &(xi, x 2 all the combinations of x it positive or be a sum of 2 terms, representing with .r n positive or negative, with x z positive or negative, with negative, each term having its coefficient.

6-12

For a function
n

136

A
(a)

Survey of Symbolic Logic

by
will

definition of the

normal form function of one variable has two terms, and normal form of functions of n + 1 variables, if

functions of k variables have 2* terms, a function of k

variables

have 2 k
(b)

2 k or 2 A+1
,

terms.

normal form function of one variable has the further character described in the theorem; and if normal form functions
of k variables
will

have

this character, then functions of k

-f-

variables

have
1

it,

since, by definition, the normal form of a function of

variables will -consist of the combinations of the (k

l)st

variable, positive or negative,

with each of the combinations repre

sented in functions of k variables.


Since any coefficient
tain terms
coefficient

may
1.

be

0,

the normal form of a function


coefficient for a

may

con

which are
is,

null.

of course,

Where no The order

of terms in

term appears, the the normal form of a

function will vary as the order of the variables in the argument of the function is varied. For example, the normal form of y) is, by defini
3>(x,

tion,

A
and the normal form

xy+

B -x y + C x -y + D
is

-x -y

of V(y, x)

P yx + Q-yx + Ry-x + S-y -x


Except for the
tive
coefficients, these differ

order of the elements in the terms.

only in the order of the terms and And since + and x are both associa
is

and commutative, such a difference

not material.
differ materially

6-15

Any two

functions of the

same variables can

only

in the coefficients of the terms.

The theorem
that, for

follows immediately from 6-12.

In consequence of 6-15,

we

can, without loss of generality,


of the

assume
in the

any two normal form functions


of

same variables with which


in

we may be concerned, the order


arguments of the functions
n
is
.

terms and the order of variables

the same.
,

And
is

also,

any function

of

variables,

3>(

Xl

x2
. .

xn

.r

n+1 ),

which
(X,,
1
.To,

equated to
.

/Ol,
x n+l
it

.1-2,

Xn)

X n+1 +f

X n ) -X n+l

may

be any chosen one of the n


last is consistent

variables.

The convention that


normal form

is

always the

with complete generality of the proofs.


of a function in the
is

6-17
null.

The product

of

any two terms

The Classic, or Boole-Schroder, Algebra of Logic

137

By

6-12, for any two terms of a function in the normal form,

there will be

some

variable, x n

such that x n

is

positive in one of

them and negative in the other; since otherwise the two terms would represent the same combination of x\ positive or negative,
t

with

0*2,

positive or negative, etc.


will

any two terms

Consequently, the product of involve a factor of the form x n -or n and will
,

therefore be null.

Unless otherwise specified,


tion

it

will

be presumed hereafter that any func

mentioned

is

in the

normal form.

The

Coefficients in a Function.

The
itself.

coefficients in

any function can be

expressed in terms of the function

6-21

If f(x)

= A = A

y.

11 -.r,

then /(I)

= A.
+ B-0

For /(I)
6-22
If /(.r)

= A-1
x+

+ U--1

= A

A.

For/(0)
6-23
/(.r)

B -.r, then /(O) = B. = A-O + B-Q = + 7M


+/(0)~z.

==

B.

=/(!)

a;

The theorem
These laws,
first

follows immediately from 6-1, 0-21

and 0-22.

stated

by Boole,

are very useful in reducing compli


if

cated expressions to normal form.


^(.r)

For example,
x + -d -x) +
(c

c (d

.r)

reduction by any other

method would be

tedious.

But we have
d

*(1)

and

*(0)

= ac (VM+-d-0) = a c (d-Q + -rf-1)


of ty(x)
is

+ (c+l) d +
(c

= acd+cd+d =

+ 0) d

a c -d + c d

Hence the normal form

given by
c

\F(.r)

d x + (a

-d +

c d)

-x
for functions

Laws analogous to 6-23, also stated by Boole, may be given of more than one variable. For example,
/(.r,

y)

/(I, l)-x

y +/(0,
?/

1)

-a y +/(!,
1, 1)
-.)*

0)

-.r

-y

+/((), 0)
0,

-.*

-//

and

$(.r, y, z)

$(1,

1, l)-.r

+ $(0,

+
<I>(1,

l)-x-ijz
--.r

+ $(0,0, l)--.r-z/2+ $(1, l,0)-.r z/-s + $(0, 1,0)


+ $(1,
0,

y-z

0) x -y -z + $(0,

0, 0)

-.r -//

-z

We

can prove that this method of determining the coefficients extends to

functions of anv

number

of variables.

138

A
If

Surrey of Symbolic Logic

6-24

I -.TI -.i- 2 -.r 3

-x n

f J

be any term of

^(.TI, xz,

x3

...

#)>

then

-j

[-

will

be the coefficient, C.

I 0, 0, 0,

...

(a)
(6)

By
If

6-23, the theorem holds for functions of one variable.


A*

the theorem hold for functions of

variables,

it

will

hold

for functions of k

variables, for,

By
is

6-11,

such that,

any function of k + 1 for some / and some /

variables, $(#1, Xz, ...


,

.T*,

0, 0,

...

Therefore,

if

every term of / be of the form

{1,1, I 0, 0,

...
...

ll
j

f
I

.n

xz...
. . .

xk

\
}

-x, -x 2
is

-x k

then every term of


f

in

which x k+
r

positive will be of the form

fi,i,...ii
I
0, 0, ...
J

x,

x^.. xk
. . .

\
J

-.n -x 2

-xk
\
(^

and the

coefficient of

any such term

will

be /

r U, U,

... U J

which,

And

similarly,

if

every term of /
...
...

be of the form

.,f 1,1,
I 0, 0,

11
j

-.rj -,r,

-x t

then every term of


fl,

*
1,

in
...

which

x/,+i is
f

negative will be of the form

l|

Tl

x,...

.rt l

Mo.o,. .ol

i-.,,-.,,.

.-,,|

The Classic, or Boole-Schroder, Algebra of Logic

139

and the

coefficient of

any such term

will

be /

which,

Hence every term


r
I

of

<

will be of the

form
x2
...
.

1,1,.

..MI
j

0, 0, ... 0,

I -.?! -.r 2

-x k -xk+l

(c)

Since the theorem holds for functions of one variable, and

since
of k

if it

hold for functions of k variables,


variables, therefore
it

it

will hold for functions

holds for functions of any

number

of

variables.

For functions
but
less useful If f(x)

of

one variable, further laws of the same type as 6-23

have been given by Peirce and Schroder.


+ B)

Ax + B-x:
= f(-A
+ -B).

6-25

/(I)
/(())

= f(A

6-26

=f(A-B) =f(-A-B). 6-27 j(A) = A + B =/(-J8) = f(A-B) =f(A+-B)


6-28
f(B)

= A-B =

/(-4)

= f(-A-B) = f(-A

+ B)

The

proofs of these involve no difficulties


it

and may be omitted.


be convenient to denote the coef
,
. .

In theorems to be given later,


ficients in functions of the

will

form

$(.?!,

x2

x n ) by A\,

z,

3,

A^,

or

by

Ci,

C2 C3
,

.,

etc.

This notation

is

perfectly definite, since the


is

order of terms in the normal form of a function


of
,
. .

fixed.

If

the argument
,

x n ), then any one of the variables, x k any function be (a*i, .r 2 be positive in the term of which Cm is the coefficient in case
.

will

p-2

~l
<

m^

(p

l)-2

~l
.r^

where p

= any

in the term.

even integer (including 0). Thus it may be determined,

Otherwise
for

will

be negative
in the
is

each of the variables

function, whether
coefficient,

it is positive or negative in the is

term

of

which Cm

the

and the term


it

thus completely specified.

We make

no use

of this law, except that


it will

validates the proposed notation.

be convenient to distinguish the coefficients of those Occasionally in which some one of the variables, say x k is positive terms in a function from the coefficients of terms in which x k is negative. We shall do this
,

140

Survey of Symbolic Logic


PS,
for coefficients of

by using different letters, as PI, P 2 which Xk is positive, and Q b Q 2 Q 3


,

terms in

for coefficients of

terms

in

which Xk
of terms,

is

negative.

This notation

is

perfectly definite, since the


is
,

number
it is
is

for a function of
is

n variables,

n always 2 the number of those in which Xk

always equal to the number of those in which and the distribution of the terms in which x k is positive, or
positive
is

negative,
is

negative,

determined by the law given above. The sum of the coefficients, Ai + Ao +


cated by
Since the

+
2

.,
3

will frequently
...

be indi
h

^A
is

or

^A
h

the product,

Ai-A A

by JJ/l or YL^h-

number

of coefficients involved will


r

always be fixed by the func


to indicate numerically the

tion which

in question, it

range of the operators

w ill be unnecessary and IJ


.

The Limits
the coefficients.

of a Function.

The lower

limit of

any function
is

is

the prod

uct of the coefficients in the function, and the upper limit

the

sum

of

6-3

A B

cAx + B-xcA + B.
(A B)(A x +

-x)

= AB

x +

AB -x

= A B

Hence

[1-9]

ABcAx + B-x.
B -x)(A
+ B)

And (A
But

x +

Ax + AB -x + ABx + B -x
=

(A

B + A)
and

x +

(AB + B)
A
x+

-x.

A B + A == A, and AB + B = B. Hence (.1 x + B -x) (A + B) = A x + B -x,


[5-4]

[1-9]

-x

cA + B.
6-31

f(B)cf(x)cf(A). [(3-3 and 6-26, 6-27]


If

6-32
then

the coefficients in any function, F(x lf x 2

;i-),

be

C C2 C
lt
,

3,

(a)
(b)

By

3,

the theorem holds for functions of one variable.


,
.
. .

Let $(xi, x z
for
.
. .

x k x k +i) be any function of k


,

variables.

By (Ml,
$(.ri, *2,

some / and some /


x k Xk+i)
,

f(x lt x z

xk )

-iCfc+i
(.TI,

+/
Since this last expression

x2

x k ) -xk+ i

(1)

may
x2

be regarded as a function of Xk+i in

which the
i,

coefficients are the functions

/ and /
$(.ri,

[6-3]
,
. .

x2

x k ) x/

(#,,

xh ) c

x2

x k x k +i)
,

The Classic, or Boole-Schroder, Algebra of Logic Let Ai{$},


Ai[3>\,

141

{3>},

etc.,

be here the coefficients in


<J>;

Ai{f\,

^2 {/I, A*\f},
etc.,

etc.,

the coefficients in/; and A,{f


.

\,

{/

[,

{f

},

the coefficients in /

If

rU{/} c/and IU{/

c/

then [0-3]

and, by

(1),

EU

{/J

JL
,

!/

<.

But
in

since (1) holds,

any
/

coefficient in

will

be either a coefficient

/ or a

coefficient in

and hence
!

lUf/! xIL-i{/
Hence
if

==IU!*i
x2 x k ),

the theorem hold for functions of k variables, so that

II A
then

{/}

c/Cd, * 2

xk )
...

and
,

EU
l,

{/

c/

(x lt

II^l^!
if

c$(a-i, x z

x k x k+} ).

Similarly, since (1) holds, [0-23]

Hence

/c

.!{/}

and / c

$c/+/ then Z^f/


+

[5-31]

*<=Z^{/1

Z^{/

But

since

any
,

coefficient in

is

either a coefficient in
==

ficient in

/ z^f/i Hence $ c ^A{$}. Thus if the theorem hold

z^m
1

/ or a coef

z^j.
it

for functions of k variables,

will

hold for functions of k


(c)

variables.

Since the theorem holds for functions of one variable, and

since

if it

hold for functions of k variables,


variables, therefore
it

it

will hold for functions

of k -f

holds generally.

As we

shall see, these

the basis of the

method by which eliminations


Since
all

theorems concerning the limits of functions are are made.


functions of the same variables

Functions of Functions.

be given the same normal form, the operations of the algebra be performed simply by operating upon the coefficients.

may may frequently

6-4

If f(x)

= A

x+

-x, then

4/0)] -

-.-1

.r

-B

-x.

[3-4]

-(A x +

-x)

= -(A
= =

x)(B -.r)
.r)

(-A + -x) (-B +

-.1

-B

+ -.1

.T

-B -x

(-A -B

+ -A) x + (-A
+

-B

+ -B) -x

But

[5-4] -.1

-B

+ -.1
-x)

Hence -(,1 x +

== -A and -.1 -B = -A x + -B -.r.

-B = -B.

142

A
The negative
of

Survey of Symbolic Logic


in the

6-41

any function,

normal form,
its

is

found by

re

placing each of the coefficients in the function by


(a)
(b)

negative.

By
If

6-4, the theorem

is

true for functions of one variable.


it

the theorem hold for functions of k variables, then

will

hold for functions of k

variables.

Let F(XI, xz, Then by 6-11 and 3


.

Xk, Xk+i)
2, for

be any function of k
,

variables.

some / and some /

+f
But
f(xi, Xz,
.
. .

(xi,

x2

x k )--x k+l

x k )-x k+ i

+f

Oi>

^,

Zk)

-Xk+i

may

be regarded

as a function of a^+i.

Hence, by 6-4,
,

a-o,

x k ) -x k+l +/

(xi,

xz

x k ) -xk+ i]

Hence

if

the theorem be true for functions of k variables, so that

the negative of /
its

is found by replacing each of the coefficients in / by and the negative of / is found by replacing each of the negative

coefficients in

by

its

negative, then the negative of


in

will

be

found by replacing each of the coefficients as has just been shown, any term of

F by

its

negative, for,

-x
in

l}

xz,

x kj
its coefficient is

which x k +i

is

positive

is

such that
X2
,

a coefficient in

-[/(>!,

...

X k )]

and any term

of

-[F(x lt
in

xz,

xk

which x k +\

is

negative

is

such that

its coefficient is

a coefficient in

-[/ (*
(c)

X*

...

.T,; )]

Since (a) and

(b)

hold, therefore the theorem holds generally.


is

Since a difference in the order of terms

not material, 6-41 holds not

only for functions in the


pletely

normal form but

for

any function which

is

com

expanded

so that every

or negative, in each of the terms.

element involved appears, either positive It should be remembered that if any


in the

term

an expanded function is missing, its coefficient is 0, and negative of the function that term will appear with the coefficient 1.
of

The Classic, or Boole-Schroder, Algebra of Logic

143

6-42

The sum
#2,
.
.

of
.

any two functions


is

of the

same

variables, $(0:1, T 2

xn)

and V(xi,

x n ),

another function of these same variables,


F(Xi, To, ... X n ),

such that the coefficient of any term in the corresponding terms in $ and ^.

is

the

sum

of the coefficients of

By

6-15,

$(0:1,0:2,

...

x n ) and V(xi, x 2

...

xn)

cannot

differ

except in the coefficients of the terms.

Let AI, At,

3,

etc.,

be the coefficients in $;

BI,

B
TI

2,

B
x*

3,

etc.,

the

coefficients of the corresponding

terms in ^.
. .

For any two such corf


.

r responding terms,
f

Ak J 1
.!

TI

TO

Tn

L -0*1 -.1 2

r ~.l n J

and B k
Tn
1
f-

xn
.
.

]
}

l
I -.Ti -.r 2
.

-X n

*1

*2

...

Ah
f

nk

f
-j

T!

0: 2

...
.
. .

-Ti -1 2

-Xn

-Xi -X 2

Xn

Ti

T2

Xn
-x n
of the

-TI -T 2

And

since addition
is

two functions

and commutative, the sum equivalent to the sum of the sums of such
is

associative

corre

sponding terms, pair by


6-43

pair.

The product
^f(xi,
0*2,

of

two functions
is

of the

same

variables,

$(1,

o: 2 ,

...

xn)

and

...

x n ),

another function of these same variables,


F(Xi,
.T 2 ,
-

Xn),

such that the coefficient of any term in


cients of the corresponding terms in
T Let Ak J
"j

F
^.
f

is

the product of the

coeffi

$ and

Xi

X2

Xn

-TI -T 2

-x n

and Bk
^.
Tl
0*2

Ti

T2

...
.
. .

Xn

}
r }

-T! -T 2

-x n

be any two

corresponding terms in

$ and
J
I

A k \[
,

TI

T2

n 1

...
.
. .

Xn

-Ti -T^

& H -X n Jfx5

-Ti -T 2

-X n

-(4*Jfc>{ -Ti -To t

-T n

By

6-15,

$ and ^ do not

differ

except in the coefficients, and by

6-17, whatever the coefficients in the normal form of a function, the

product of any two terms is null. Hence all the cross-products of terms in $ and ^ will be null, and the product of the functions will

144

A
be equivalent to the
pair

Survey of Symbolic Logic

sum

of the products of their corresponding terms,

by

pair.

Since in this algebra two functions in which the variables are not the same may be so expanded as to become functions of the same variables,
these theorems concerning functions of functions are very useful.

TV.

FUNDAMENTAL LAWS OF THE THEORY OF EQUATIONS

We
may
of

have now to consider the methods by which any given element be eliminated from an equation, and the methods by which the value
"

an

unknown

"

may

be derived from a given equation or equations.


of equation for eliminations
0.

The
is

most convenient form


equation with one

and solutions

the

member

Equivalent Equations of Different Forms. If an equation be not in the form in which one member is 0, it may be given that form by multiplying

each side into the negative of the other and adding these two products.
7-1 a

b is

equivalent to a -b + -a b

0.

[2-2] a
[4-9] a

b is

equivalent to the pair, a c b and b c a. = equivalent to a -b = 0, and b c a to -a b


6 is

0.

And

[5

72] a -b

and -a

are together equivalent to a -b

+ -a b

0.

The transformation
7-12
a

of

an equation with one member

1 is

obvious:

1 is

equivalent to -a

0.

[3-2]

By
in

6-41,
in

form

xn) = 1 any equation of the form f(xi, ,r 2 which one member is simply by replacing each
,
. . .

is

reduced to the

of the coefficients

/ by

its

negative.
is

Of especial interest

the transformation of equations in which both

members
7-13
If

are functions of the


<J>(a,-i,

same

variables.
,

.T 2 ,

Xn)

and

ty(x 1} x 2

...

x n ) be any two functions of the

same

variables, then

is
if

equivalent to F(XI, x 2

xn)

0,

where
<,

A A A
i,

2,

3,

etc.,

be the coefficients in

and

F is a function such B B B etc., be the


1} 2, s,

that
coef

ficients of the

corresponding terms in ty, then the coefficients of the corre sponding terms in F will be (Ai -Bi + -A l #0, (A 2 -B 2 + -A 2 B 2 ), (A 3 -J5 3 + -.4 3 # 3 ), etc.

The Classic, or Boole-Schroder, Algebra of Logic

145

By
By

7-

1,
-<

$ = ^

is

equivalent to ($ x-\F) +
of the

(-<!>

x ^)

0.
<

6-41,

and -^ are functions

same variables as

and ^.
be a

Hence, by 6-43, $

x-^ and -$
(i-42,

will

each be functions of these


ty)

same

variables,

and by
-^,

(3>

x-^) + (-$ x

will

also

function of these same variables.

Hence

3>,

^,

-<,

<J>

x-^, -$ x ^, and ($ x-^) +

(-<

x ^) are

all

functions of the same variables and, by 6-15, will not differ except in the coefficients of the terms.
If Ate

be any coefficient in

<,

and

Bk

the corresponding coefficient


coefficient in
will

then by 6-41, the corresponding and the corresponding coefficient in -^


in ^,

-$

will

be

-A k
be
.

be

-Bk

Hence, by 6-43, the corresponding coefficient x Ak -B k and the corresponding coefficient in


,
-3>

in

$x-^
be

will

fy will
(<

-A k Bk
(-<

will

Hence, by 0-42, the corresponding be A k -B k + -A k B k


.
>^)

coefficient in

x-^) +

x ^)

Thus ($ x-^) + (-$ x


the theorem holds.

is

the function F, as described above, and

By
in the

7-

1,

for every equation in the algebra there

is

an equivalent equation

be

form in which one member is 0, and by 7 13 the reduction can usually made by inspection. One of the most important additions to the general methods of the
s

algebra which has become current since the publication of Schroder


is

work

Poretsky

Law

of

Forms.

By
t

this law, given

any equation, an equiva

lent equation of

which one member

may

be chosen at will can be derived.


t.

7-15

is

equivalent to

a -t + -a

If a

And
Since

if t

- 0, a-t + -at = Qt+l-t = = a -t + -a then [7-1]


t,

t.

(a -t
t

+ -a

t)

-t

+ (a

+ -a

-t)

a -t + a

may
:

here be any function in the algebra, this proves that every

equation has an unlimited number of equivalents. of the law is

The more

general form

7-16

b is
1]

equivalent to
a

(a b

+ -a -6)

+ (a -b + -a

b) -t.

[7

b is equivalent to a -b

+ -a b

0.

And

[6-4] -(a -b

+ -ab) = ab

+ -a -b.

Hence

[7-15]

Q.E.D.

The number
9

of equations equivalent to a given equation


fondamentales de la theorie des egaliies logiques, Chap.

and expressible
i.

See Sept
11

lois

146
in

A
will

Survey of Symbolic Logic be half the number of distinct functions which


n

terms of n elements

~ can be formed from n elements and their negatives, that is, 2 /2. The sixteen distinct functions expressible in terms of two elements,

a and

b,

are:
b,
f

a, -a,

-b,

(i.

e.,

a -a, b -b, a + -6, -a +

etc.), 1
6,

(i.

e.,

a + -a, b + -b, etc.), a

b, b.

-^

_a i

_ rt

-^

a +

6,

-a +

-b,

ab

+ -a -b, and a -b + -a

In terms of these, the eight equivalent forms of the equation a


a
=

b are:

-=-&;
=

a -6 + -a &

a 6 + -a -6
b.

ab =

a + b;

a -b

-ab; -a -b

-a + -b: and a + -6

-a +

Each

of the sixteen functions here appears

on one or the other

side of

an

equation, and none appears twice. For any equation, there is such a

set of equivalents in

terms of the

elements which appear in the given equation. And every such set has = a -b + -a b) what may be called its "zero member" (in the above,

and and and

its

"whole

member"

(in

the above,

a &+

--&).

If

we observe
member"

the form of 7-10,


"whole
t

we

shall note that the functions in the

"zero

member"

are the functions in terms of which the arbitrarily

chosen
is

is

determined.

Any

the

which contains the function


.

=
{

0}

contained in the function

=
{

1
t,

The

validity of the
i.

law depends
It is

simply upon the fact that, for any

Octfcl,

e.,

1-/ + 0--/.

rather surprising that a principle so simple can yield a law so powerful.

Solution of Equations in
sible

One Unknown.

Every equation which

is

pos

according to the laws of the system has a solution for each of the

un
to

knowns involved.
equations
in

This

is

a peculiarity of the algebra.

We

turn

first

one unknown.

Every equation

in x,
.

if it

be possible in the

algebra, has a solution in terms of the relation c

7-2

x +

-x

is

equivalent to

c x c -A.

[5-72]

x +

-x

is

equivalent to the pair,

Ax

and

B -x = 0. [4-9] B -x = And A x =
7-21

is

equivalent to

ex.

is

equivalent to x -(-A)
A"

0,

hence to x c-A.

solution in the form // c x c


is

is

indeterminate whenever the equa


-.T.

tion which gives the solution


First,
if

symmetrical with respect to x and

the equation be of the form


is,

x+

-x

0.

The But

solution then
if

A
1.

x+

-x

A ex c-A. = 0, then A A
--=

(x

+ -x)

= A

x+

-x

0,

and

-A =

The Classic, or Boole-Schroder, Algebra of Logic

147

Hence the
satisfied

solution

is

equivalent to
x.

Oc.rcl, which [5-61-63]

is

by every value of

In general, any equation symmetrical with respect to x and -x which gives the solution, // c x c K, will give also // c -x c K. c x and // c -x, then [4 9] // x = But if and // -x = H.

Hence

[1-62] II

0.

And =

if

xcK

and -xcK, then

[5-33] x +

-xcK, and
1.

[4-8, 5-63]

I.

Hence

H ex c K will

be equivalent to
if

c a: c

It follows directly

from 7-21 that

neither x nor -x appear in an equa

tion, then although they

may

be introduced by expansion of the functions

involved, the equation remains indeterminate with respect to x.

7-22
ever

An A =

equation of the form

x +

-x

determines x uniquely when

-B,

B = -A. [3-22] A = -B
is

and

-.1

= B

are equivalent;

hence either of

these conditions
[7
21

equivalent to both.
is
it is

x+
if

Hence

B -x = B = -A,
[2-2] to

equivalent to
equivalent

B c x c -A. to B ex cB and
x+

to -.1 c.r c-.4,

and hence

= B = -A.

In general, an equation of the form

-x

=
is

determines x be

tween the
if,

limits

B and

-A.

Obviously, the solution


is

unique

if,

and only

these limits coincide;

and the solution

wholly indeterminate only

when they are respectively and 1, the limiting values of variables generally. 7-221 The condition that an equation of the form A x + B -x = be pos
sible in the algebra,

and hence that

its

solution be possible,
[5-65]
if

is

AB =
0,

0.

By 6-3, A B A B = 0. Hence if A B
=(=

cAx + B-x.
0,

Hence

x +

B -x =

then

then

x+

B -x =

must be

false for all values

of x.

And A x + B -x = A B = is called the


+
is

and the solution B ex


"

equation of

c-A are equivalent. condition" of A x + B -x = 0:


To
call it

it is

a necessary, not a sufficient condition.

the condition that

B -x =

have a solution seems inappropriate:

the solution

Bc.xc.-A
false, or

equivalent to

x +

-x

=
will

0,

whether

x +

-x

be true,

impossible.
solution of

The

sense in which

AB =
made

conditions other forms of the

x+

B -x =

be

clear in

what

follows.

The equation

of condition

is

frequently useful in simplifying the solution.

148
(In this connection,

A
it

Survey of Symbolic Logic

should be borne in mind that


if

AB =

follows

from

x +

-x

0.)

For example,

a b x + (a +

6)

-x

=
of condition
is

then (a +

b) c.r

c-(a

b).

But the equation

a b (a + b)

= ab =

0,
is

or,

-(a

b)

Hence the second


solution

half of the solution

indeterminate, and the complete

may

be written a+ b

ex
is

However,

this simplified

form of the solution

equivalent to the original


is

equation only on the assumption that the equation of condition

satisfied

and a

0.

Again suppose

ax

+ b -x +

=
we have

Expanding

with reference to
(a

x,

and
(b

collecting coefficients,

c)

x +

c)

-x

and the equation


(a

of condition is

+ c)(b +

c)

a b + a
b

+b

= ab + =

The

solution

is

c x c -a -c
c
0,

But, by 5-72, the equation of condition gives

and hence -c =

1.

Hence the complete solution may be written


b

c x c -a
is

But here

again, the solution b

ex c-a

equivalent to the original equation

only on the assumption, contained in the equation of condition, that c = 0. This example may also serve to illustrate the fact that in any equation
is 0, any terms which do not involve x or -x may be dropped without affecting the solution for x. If a x + b-x + c = 0, then by 5-72, a x + b -x = 0, and any addition to the solution by retaining c will

one member of which

be indeterminate.

All

terms which involve neither the unknown nor


constituent"

its

negative belong to the "symmetrical

of the equation

to be

explained shortly.

Poretsky
is

Law
-x

of

Forms

gives immediately a determination of x which

equivalent to the given equation,

whether that equation involve x or not.

7 23

x+

B
=

=
x +

is

equivalent to x

= -A

x +

-x.

[7-15]

-x

is

equivalent to

(A x +

-x) -x + (-A x +

-B

-r)

r- B

-x +

-Ax

The Classic, or Boole-Schroder, Algebra of Logic

149
of Jevons. 10

This form of solution

is

also the one given

by the method
is

Although

it

is

mathematically objectionable that the expression which


-.r,

gives the value of x should involve x and


logically simple

this

in reality a useful

and

form
is

of the solution.

It follows

from 7-2 and 7-23 that

= -A

x +

B -x

equivalent to

B
is

cx c

-.1.

Many

writers on the subject have preferred the form of solution in

which the value of the unknown

given in terms of the coefficients and an

undetermined (arbitrary) parameter.


form.

This

is

the most

"mathematical"

7-24
+

If

AE =
is

0,

as

the equation

x+

B -x =
= B

requires,

then

Ax
is

-x

=
-x

satisfied

by x

= B -u + -A
is

u, or x

+ u -A, where u

arbitrary.

x +

B
.

And this = there

solution
is

some value

complete because, for any x such that of u such that x = B -u + -A u = B

+u

-A

(a)

By
if

6-4,

if

-u + -A
?/,

u,

then -x

= -B -u + A

u.

Hence

x +

B -x
if

= B -u + -A then = A (B -u + -B u) + B
= B -u

(-B -u +

A u) = AB

-?/

AB

= A

Hence
of u,

x+

A B = and B -x = 0.

.r

-A

u,

then whatever the value

Suppose x known and such that A x + Then if x = B -u + -A u, we have, by 7-1,


(b)

-x

0.

(B -u +

-A

if)

-x + (-B -u +

n) x

= (A
The condition that

x +

-A

-x) u + (B -x +

-B
is,

x)

-u =

this equation hold for

some value
x + -.1

of

by 7 221,

(A x + -A -x) (B -x + -B x)
This condition
is

= A -B

-x

satisfied

if

A
(c)

(B + -B) x + (A + -A)

A x + B -x = 0, for then B -.r = AB + A -B x + -A B -x =


-x

and by 5-72,
If If

A -B x + -A B
0,

0.

AB =

then

A B = 0, then A B u Hence B -u + -A u = B -u + -A (B + -B) u + A B u = B -u + (A + -A) B u + -.1 -B u = B (-u + u)


But
10

B -u + -A = 0.

= B

u-A,

for:

-A -B u = B + -A -B

u.

[5-85]

B + -A -Bu =

B+n -A.

See above, p. 77.

150

Survey of Symbolic Logic

Only the simpler form

of this solution, x

= B+

u -A,

will

be used hereafter.

The above

solution can also be verified

by substituting the value given

for x in the original equation.

We

then have

A
And
if

(B -u + -A
0,

?/)

(-B -u +
is

Au) = A B-u + ABu = A B


the same as

AB =

the solution

verified for every value of u.

That the

solution, x

= B-u + -Au = B + u-A, means


if

B ex c-A,
parameter,
If If

will
u, is

be clear

we

reflect that the significance of the arbitrary

to determine the limits of the expression.

u =

B -u + -A u = B + u -A = B. = 1, B-u + -Au = -A and B + u-A = B + -A. But when


0,

AB
Hence x =
stated

0,

B + -A = -A B + -A =

-A.
fact,

B -u + -A u = B + u-A
of condition of the

by

B ex
1,

simply expresses the c-A, that the limits of x are B and -A.

otherwise

The equation

and the solution

for equations of the

form

C x + D -x =
7 25

and

form

x+

B -x = C x + D -x,
1 is

follow readily

from the above.

The equation

of condition that

C x + D -x =
C.

C+D =

1,

and the

solution of

Cx

D -x
6-3,

1 is

-D c x c

(a)

By
if

Cx + D-xcC + D.
any value
1.
==

Hence

there be

of x for

which C x +

D -x =
0,

1,

then

necessarily
If

C+

D =

Cx + D-x (6) -DcxcC.


7-26
x
If

1,

then [6-4]

-Cx + -D-x =
1

and

[7-2]

C+

D =

1,

then the equation


is

Cx + D-x =
equivalent to

is

satisfied

by

= -D

uC, where u
Since [6-4]

arbitrary.
1
is

Cx + D-x =
is

-Cx + -D-x =

0,

and C +

D =

equivalent to

-C -D =

0,

the theorem follows

from 7-24.
7-27
If

x+

B -x = C x
B -D

-x, the equation of condition

is

(A -C +

-A C)(B -D

-BD) =0
-C, or

and the

solution

is

-B

cx c

A C + -A
+
+

= B -D +-B D By
7 13,

+ u (A

C + -A

-C), where

is

arbitrary.

x +

-x

= Cx

-x

is

equivalent to

(A-C + -AC)x

(B-D

-Bd-x =

0.

The Classic, or Boole-Schroder, Algebra of Logic

151

Hence, by 7-221, the equation of condition And by 7-2 and 7-24, the solution is

is

as given above.

B -D + -B D c x c - (A -C + -A C) x = B -D +-B D + u--(A -C + -A C), where u And [6-4] -(A -C + -AC) = A C + -A -C.


,

or
is

arbitrary.

The

subject of simultaneous equations

is

very simple, although the


is

clearest notation

we have been

able to devise
in

somewhat cumbersome.

7-3

The condition that n equations

one unknown,

x+

-x

0,
is

A*x +

-x

0,

x+

-x
h

0,

may
)

be regarded as simultaneous,

the condition that

(A
fi,

Bk =
is

And

the solution which they give, on that condition,

^B
k

cxc
is

H-A
k

or x

^B
k

u>

II -^* where u
k

arbitrary.

By

6-42

and

5-72,

x+

5
1

-x

0,

A*x +

-x

0,

x+

-x

0,

are together equivalent to

1 (A +

A n )x +
,

(B + B* +

B")

-x

or
k

z +
k

-*

=
is

By

7 23, the equation of condition here

A* x
A:

B =
k
A

But

ZM* x Z
A-

fc

=
1

(.I

+
. .

,I

+...+.4

= A B
1

+
l

B- +

+
.
.

A Bn + A B
1

+ A*
+
A"

A*B

+ A*

+ A* B- +

B B

+ +

A B + An B
+
2

Z(^*).
h,k

And by 7-2 and

7-24, the solution here

is

or

*=
A-

A-

And by
It

5-95,

-{^M
k

==

II ~A
k

k
.

may

be noted that from the solution in this equation,

?i

partial solu

152
tions of the form

A
B
h

Survey of Symbolic Logic

c.r

c-A>

can be derived, for

Wc^B*
k
2 Similarly, 2
"

and

H -A*
k

c-A>

partial solutions

can be derived by taking selections of

members

of

k ^Z H

an d II

-A k

Symmetrical and Unsymmetrical Constituents of Equations. Some of the most important properties of equations of the form A x + B -x = are

by dividing the equation into two constituents the most comprehensive constituent which is symmetrical with respect to .r and -x, and a completely unsymmetrical constituent. For brevity, these may
clear

made

be called simply the


constituent
is
".

"symmetrical constituent"

and the

unsymmetrical

In order to get the symmetrical constituent complete, it necessary to expand each term with reference to every element in the

function, coefficients included.

Thus

in .1

x+

-x

it is

necessary to

expand the

first

term with respect to B, and the second with respect to A.

A
By

(B + -B) x + (A + -A)
5 72, this
.1
is

-x

= A Bx

AB -x + A -B x + -A B -x
.1

equivalent to the two equations,


-.r)
is

(x

.1

B =

and

-B

x + -.4

-x
is

=
the

The

first of

these

the symmetrical constituent; the second

unsym

metrical constituent.

The symmetrical

constituent will always be the equa

tion of condition, while the unsymmetrical constituent will give the solution. But the form of the solution will most frequently be simplified con

by

sidering the symmetrical constituent also.

The unsymmetrical
is

constituent

will always be such that its equation of condition the equation of condition of

satisfied a priori.

Thus

A -B
is (.1

x +

-A B -x =

-B)(-A B) = 0, which is an identity. By this method of considering symmetrical and unsymmetrical con stituents, equations which are indeterminate reveal that fact by having no unsymmetrical constituent for the solution. Also, the method enables
any term
us to treat even complicated equations by inspection. Remembering that in which neither x nor -x appears belongs to the symmetrical

constituent, as does also the product of the coefficients of x

and

-.1-,

the

separation can be

made
(c

directly.

For example,

+ x) d + -c -d + (-a + -x) b

The Classic, or Boole-Schroder, Algebra of Logic


will

153

have as

its

equation of condition
c

d + -c -d + -a

and the solution

will

be
b

c .r c -d
symmetrical constituent
is

Also, as

we

shall see shortly, the

always the

complete resultant of the elimination of x. one

The method does not readily apply to equations which do not have member 0. But these can always be reduced to that form. Plow it

extends to equations in more than one

unknown

will

be clear from the

treatment of such equations.


Eliminations.

The problem

of elimination

is

the problem, what equa

tions not involving x or -x can be derived from a given equation, or equa


tions,

which do involve x and

-.r.

In most algebras, one term can, under

favorable circumstances, be eliminated from two equations, two terms from three, n terms from n 1 equations. But in this algebra any number

terms (and their negatives) can be eliminated from a single equation; and the terms to be eliminated may be chosen at will. The principles whereby such eliminations are performed have already been provided in
of

theorems concerning the equation of condition.


7-4

A B =
By

contains

all

the equations not involving

.r

or

-.r

which can

be derived from

x+

-x

0.

7-24, the complete solution of

x+

B -.r =

is

= B

-?/

-A u
we have

Substituting this value of x in the equation,


.1

(B -u +

-A

11)

B
is

(-B -u +

u)

= A B -u

AB

= A B =
.r.

Hence
It is at

AB =

the complete resultant of the elimination of

once clear that the resultant of the elimination of x coincides

with the equation of condition for solution and with the symmetrical con
stituent of the equation.

7-41

If
,

n elements,
,T 3 ,
.
. .

x\,
0,

.r 2 ,

.r 3

xn

be eliminated from any equation,


is

F(XI, # 2

xn )

the complete resultant


in

the equation to

of the

product of the coefficients


(a)

F(x it x 2

.r 3

x n ).
is

By
If

6-1 and 7-4, the theorem

true for the elimination of

one element,
(b)

= 0. .r, from any equation, /(.r) the theorem hold for the elimination

of k elements, x\,

X2

154 xk

A
. . .

Survey of Symbolic Logic

from any equation,

$(.TI,

x Zt

...
.TI,

xk )
.T L
>,

=
.
.

0,
.

then
,

it will

hold

for the elimination of k

elements,

x k Xk+i, from any

equation, V(xi, x z

x ki xk +i)
...
,

0, for:
/(.ri,

By (Ml,

V(xi,

.1-2,

xk x k+ i) =

xz

x k )-x k+i xz
,
.

+f

(xi,

x k )--Xk+i.

And

the coefficients in

will

be the coefficients in / and /

By

7-4, the complete resultant of eliminating x k +i


/(.Ti, .To,
is
.
.

from
i

X k ) Xk+i + /
. .
.

(Xi,

XZ

Xk)~X k+ =
a *)

f(xi,

.T 2 ,

x k ) x/ Oi,
.

.r 2 ,

=
=
is

And by
to
<(.TI,

6-43,
2*2.,

/(.TI,
a;*;)

.T 2 ,

x k ) X/

(.TI, or 2 ,

a**)

equivalent
if

...

0,
2,

where $
3,

is

a function such that

the

coefficients

in/ be PI,

P P

etc.,

and the corresponding

coefficients
<

in/

will then the corresponding coefficients in be PiQi, P 2 2, PsQa, etc. Hence if the theorem hold for the elimina x k ) = 0, this tion of k elements, .TI, x Z) ... x k from $(.TI, x z

be Qi,

Qz Q
,

s,

etc.,

elimination will give


(P.Qi)(ft<2.)(PiC.).
.

(f iftftis

.ddQ..

0,

where PiP^Pzor in /

-QiQzQsi.

the product of the coefficients in


SF.
.TI,

3>,

and /
if

e.,

the product of the coefficients in

Hence
.
.

the theorem hold for the elimination of k elements,


<(.TI,

xz

x k from
,

xz

xk )
,

0, it will

hold for the elimination of


^(.TI,

elements,
i

Xi,

xz

...

x k x k+ i, from

xz

x k x k+ i)
,

0,

provided x k+ be the first eliminated. But since the order of terms in a function

is

immaterial, and for

any order of elements in the argument of a function, there is a normal form of the function, x k+ in the above may be any of the k + 1 elements in \F, and the order of elimination is immaterial.
i

(c)

Since (a) and

(b)

hold, therefore the

theorem holds

for the of

elimination of any

number

of elements from the equation to

any function

of these elements.
it is

By this theorem,

possible to eliminate simultaneously

elements from any equation, by the following procedure:


equation to the form in which one
(2) Develop the other

any number of (1) Reduce the

form;

member is 0, unless it member of the equation


(3)

already have that


as a normal-form

function of the elements to be eliminated;

Equate to

the product of

the coefficients in this function.


resultant.

This

will

be the complete elimination

The Classic, or Boole-Schroder, Algebra of Logic


Occasionally
it is

155

convenient to have the elimination resultant in the


1,

form of an equation with one member gives the resultant have that form.
7-42

especially

if

the equation which

The complete

resultant of eliminating
,

n elements,

x\

x2

...

xn

from any equation, F(XI, x 2 of the coefficients in F(XI, x 2


Let AI,

...
...
etc.,

xn) x n ).

==

1,

is

the equation to

of the

sum
x n ).

2,

A
=

3,

be the coefficients in F(XI, x 2


Xz, ...

F(XI, xz, ... x n )

1 is
,

equivalent to -[F(xi,

x n )]
,

0.

And
$

by 6-41, -[F(xi, x 2
that
will
if

...

x n )]

is

a function, $(xi, x 2

...

x n ), such

any
-A/c.

coefficient in

be AU, the corresponding coefficient in

be

Hence, by 7-41, the complete resultant of eliminating x\ x 2 from F(XI, Xz, ... x n ) = 1 is
9

...

xn

U-A
But
[5-95] -{

=0,
=--

or

-{ 11-^1

-=
1

11-4}

Z A.

Hence Q.E.D.

For purposes
tion
is

of application of the algebra to ordinary reasoning, elimina

a process more important than solution, since most processes of


"middle"

reasoning take place through the elimination of

terms.

For

example

If all b is x,

c x,

-x
ax

and no a
then

is x,

= =

0,

ax

+ b -x

0.

Whence, by elimination, a

0,

or no a

is b.

The complete solu unknown may be accomplished by tion of any equation in more than one eliminating all the unknowns except one and solving for that one, repeating
Solution of Equations in

more than one Unknown.

the process for each of the unknowns.

Such solution

will

be complete

because the elimination, in each case, will give the complete resultant which is independent of the unknowns eliminated, and each solution will be a

unknown, and complete, by previous theorems. How ever, general formulae of the solution of any equation in n unknowns, for each of the unknowns, can be proved.
solution for one

any equation in n unknowns is identical with the resultant of the elimination of all the unknowns; and this resultant is the condition of the solution with respect to each of the unknowns sepa
7 5

The equation

of condition of

rately.
(a)

If the

equation in n unknowns be of the form


z,
.

xn )

0:

156

Survey of Symbolic Logic


...

Let the coefficients in F(XI, x 2

x n ) be AI,

2,

3,

etc.

Then,

by 6 -32,

II A c F(xi, x 2
and
[5-65]

...

xn)

JJ

A =
=
0*2,

is

a condition of the possibility of

F(x lt

0-2,

z) Xi,

And
.

[7-41]
,

UA
.

is
.

the resultant of the elimination of


.

x2

x n from F(x\,
(6)

xn)

0.

If
.

the equation in n
.

unknowns have some other form than


it

F(XI, x 2)
of that

Xn)

0,

then by 7-1,

has an equivalent which


its

is

form, and

its

equation of condition and

elimination

resultant are the equivalents of the

equation

of

condition and

elimination resultant of

its

equivalent which has the form


,

F(x lf X 2
(c)

...

X n)

=0
the

The

result of the elimination of all

unknowns

is

the
,

equation of condition
because
(1)
:

with respect to any one of them, say x k

ing

all

(2)

The equation to be solved for Xk will be the result of eliminat the unknowns but Xk from the original equation; and The condition that this equation, in which Xk is the only
is,

unknown, have a solution for Xk result of eliminating Xk from it.

by

(a)

and

(6),

the same as the

Hence the equation


other

of condition with respect to Xk is the


first

same as
all

the result of eliminating, from the original equation,

the

unknowns and then x^


(6),

And by 7-41 and


independent

the result of eliminating the unknowns

is

of the order in
it

which they are eliminated.


be unnecessary to investigate separately

Since this theorem holds,

will

the equation of condition for the various forms of equations;

they are

already given in the

theorems concerning elimination.


,
. . .

7-51

Any
its

equation in n unknowns, of the form F(XI, x 2


as follows: Let x k be

x n)

0,

provided
of the

equation of condition be satisfied, gives a solution for

each

unknowns
and

any one

of the

unknowns;
,
. . .

let P],

2,

Pa, etc., be the coefficients of those terms in F(XI, x 2


is

x n ) in which Xk

positive,

Qi,

Q2 Q
,

etc.,

the coefficients of those terms in which Xk


is

is

negative.

The

solution then

II Q ex* c

2 -P,

or

Xk

= II Q

+ u-

2 -P,

where u

is

arbitrary.

The Classic, or Boole-Schroder, Algebra of Logic

157

(a)

By (Ml,
to

for

some / and some /


.r 2
,
. . .

F(x\,
.
.

.TO,
.

x, n )

=
--=

is

equivalent

/(.ri,

.r n _i)-.r n

+/
3,

(.ri, .T 2 ,

.r n _i)--.r n

0.

Let the coefficients

in

/ be PI,

P P
2,

etc., in

be ^i,

z,

Qs, etc.

Then
in

PI,

P
Xk
is

2,

3,

etc., will

be the coefficients of those terms in

F
in

which
Xk

is

positive, Qi, Q*,

Q3

the coefficients of terms in

which

negative.
.T n _])-.r n

If /(.TI,
.TI, .TO,
.

.T 2 ,

be regarded as a function of the variables,


,

n _i, its
.r 2
,
. .

coefficients will be Pi.r n


.

2 .r n ,

3 .T n

etc.
.TI,
o* 2 ,

And
. .
.

r n -i) -x n be regarded as a function of / Oi, x n -i, its coefficients will be Q -x n Q 2 -x n (? 3 -x n etc.


if
1
,

Hence, by 6-42,
/(.?!,

X2

X n -l) X n + / (1,
^(.TI,
.TO,
. .
.

.T 2 ,

Xn-i) -X n

=
a function in

is

equivalent to

x n -i)

0,

where

is

which the
+

coefficients
etc.

are

(PiX n + Qi-x n ),
is

(P 2 x n +

Q2
.

-.r n ),

(P 3 ^n

3-^ n ),
^(iCi,

And

T2

...

.r n _i)

equivalent to F(XI,

.r 2 ,

xn)
,
. .

0.

.r n _i 7-41, the complete resultant of the elimination of 0*1, .r 2 = will be the equation to of the product x n -i) from ty(xi, x 2

By

of its coefficients,

But any expression


tion of x n
.

of the

form

P nr n + Q

-x n

is

a normal form func

Hence, by 6 43,

By

7-2 and 7-24, the solution of

H P^n+

Qr -^

is

And
for

[5-951]

-{II P]

----

T.-Pis

(b)

Since the order of terms in a function

immaterial, and

is a any order of the variables in the argument normal form of the function, x n in the above may be any one of the .r n _ ) and / (.n, .r 2 .T n ), and /(.TI, .i 2 variables in F(x^ .T 2

of a function there

.T n

_0 each some function

of the

Therefore, the theorem

holds for

1 variables. remaining n one of the variables, x k any


.

That a
is

single equation gives a solution for

any number

of

unknowns

another peculiarity of the algebra, due to the fact that from a single equation any number of unknowns may be eliminated.

158

Survey of Symbolic Logic

As an example of the last theorem, we give the solution of the exemplar equation in two unknowns, first directly from the theorem, then by elimina
tion

and solution

for

(1)

x y+

-x y +

each unknown separately. C x -y + D -x -y = has the equation of condition,

ABCD
Provided this be
satisfied,

the solutions for x and y are


or

BDcxc-A+-C,
CDcyc-A+-B,
(2)

x
y

= BD

+ u (-A + -C)

or

= CD
is

+u

(-A+-B)

x y+

-x y +
(a)

C x -y + D -x -y =

equivalent to
-x) -y
-.r

(Ax

B -x) y+(Cx + D

=
=

and

to (b)
(a),

(A y + C -y) x + (B y +

D -y)

Eliminating y from

we have
+

(Ax
The equation

B-x)(Cx

D-x) =

ACx + BD-x
is,

of condition

with respect to x

then,

(AC)(BD) =
And
the solution for x
C),
is

ABCD
C).

BDcxc-(A

or
(6),

BD + u--(A

And -(A

C)

= -A

-C

Eliminating x from

we have
y+
is,

(Ay+C-y)(By + D-y) = A B
The equation
of condition with respect to y
is

C D -y =

then,

BCD

0.

And

the solution for y

CDcyc-(AB),
Another method

or

CD + v-(AB).
-y
in

And -(A B) = -A+-B


two unknowns, x and y, coefficients, with x and u
this value

of solution for equations in

would be to solve

for y

and

for

terms of the

as undetermined parameters, then eliminate y


of
it

by substituting

in the original equation,

and solve

for x.

By

a similar substitution,

then be eliminated and the resulting equation solved for This y. method may inspire more confidence on the part of those unfamiliar with

may

this algebra, since it is a general algebraic

method, except that

in other

algebras

more than one equation


solution of

is

required.

The
y

x y+

-x y +

(C x +

D -x)

+ u--(A x +

C x -y + D -x -y = for y is B -x) = (C + u -A) x +(D + u -B) -x

The Classic, or Boole-Schroder, Algebra of Logic

159

The

solution for -y

is

-y

= (A

x +

B -x)

v-(C x

D -x) =

(A +

-C) x + (B +

-D) -x

Substituting these values for y and -y in the original equation,

(A x +

B
=

-x)[(C + u -A) x + (D + n -B) -x] + (Cx + D -x)[(A +

(C + u -A) x +
7?

=
Hence

AC x + D -x

B (D = 0.

+ u -B) -x +

C (A +

-C) x + (B + v -D) -x\ -C) x + 1) (B + v -D) -x


v

B D ex c-A
it is

+ -C.
to equations in

Theoretically, this

method can be extended

any number

of

unknowns:
7 52

practically,

too cumbersome and tedious to be used at

all.

Any

equation in n unknowns, of the form


F(XI, x 2
,

...

xn)

f(xi,

x2

...

xn)

gives a solution for each of the


of the

unknowns
Qi, Qi,
3,

as follows:
Q.a,

Let x k be any one

unknowns;
If 2
,

let PI,
.
. .

P P
2,

3,

...

be the coefficients in F,

and

J/i,

3/3,

AI,

A2 A
,

... the coefficients of the

terms

in /, so that

and

1/V are coefficients of

corresponding terms in which x k is positive,


is

and Q r and

are coefficients of terms in which Xk


is

negative.

The

solu

tion for Xk then

II
or
Xk

(Qr

N
(Qr

+ -Q,

r)

Xk

C
+ U
r

(P r

M
^

+ ~P, - J/ r )

= II
r

-N
i,

+ -Qr
...

r)

(Pr

By
i,

7-13,
-I

7 X.i
n)

;i 2 ,

.r n )

=
is

#2,

0,

where $

/Oi, 2 ... .r n ) is equivalent to a function such that if A r and l? r


,

be coefficients of any two corresponding terms in coefficient of the corresponding term in $ will be

F and /, then the A -B + -A B,.


r r
r

Hence, by 7-51, the solution

will

be

II
Or

(Qr

"N r

+ -Qr

r)

C Xk C

-(P r

-M

-P

3/ r )

= II
[0-4]

(Qr

-N

-Q r
r)

r)

+ H

-(P r - J/ + -P,
r

J/r)

And
7-53

-(P r -J/ r + -P J/
r

= (P

J/ r +

-P -J/
r

r ).

equations in n unknowns, each of the form be regarded as simultaneous, is as follows: xn) 0, may F(XI, x, x n ) = 0, Let the coefficients of the terms in F 1 in the equation F (x lf .r 2 1 1 let the coefficients of the correbe Pi 1 , Pa 1 P3 1 QJ, Qs Qi

The condition that


. . .

160

A
in

Survey of Symbolic Logic

spending terms

F
Qs

2
,

in the
.

equation

(.ri, ,r 2 ,

.r n )

0,

be Pi 2

2 2
,

Pa

2
,

2
,
.

Q\~, Qz~,

Fm
z

in the
,

equation

F m (x

lt

the coefficients of the corresponding terms in m xz x n ) = 0, be PI W P 2 m P 3 W Qi Qzm


,
. .

....

The condition then


;

is

Or

if

*
r

be any coefficient, whether

or Q, in

P\

the condition

is

And

the solution which n such equations give, on this condition, for any
Xk, is

one of the unknowns,

as follows:

Let PI^, P/,

P
h

7i

...
0, in

be the coef

ficients of those terms, in

any one
,

of the equations

F =
is

which

x-k

is

positive,

and

let

Qi

h
,

Qi

h
3
,

...

be the coefficients of those terms, in


solution then

F =
h

0, in

which Xk

is

negative.

The

nizvic^czn-w]
r

ft

or

**

= ![ tZ
r

Qr

w-Z
r

III
h

-Pr"]

By
.
.

6-42,

equations in n unknowns, each of the form F(x\, x z


&(x-i,

x n)
xn)

0, 0,

are together equivalent to the single equation

xz

where each of the


1
,

coefficients in
2

<

is

the

sum

of the
l

F That is, if corresponding coefficients in F F F P r m be the coefficients of corresponding terms in P1 F 2


3
, ,
. .

P
.

P, Fm

2
,

then the coefficient of the corresponding term in

<

will
h

be

P, +
1

P
Q
r

2 r

+ P, .-,

or

2 Pr
h

and
1
,

if
2
,

Qr
.

Qr

2
,
.
.

m be the coefficients of corresponding terms in


<

F F
will

Fm

then the coefficient of the corresponding term in

be
h

The equation

of condition for
0,

<>

0,

and hence the condition that


is

F =
1

0,

F =
2

...

Fm =

may

be regarded as simultaneous,
that
is,

the equation to

of the product of the coefficients in $;

or

nizjvixiiizcr i-o h
r

And by

7-51, the solution of $(#1, T 2,


-

#n)

for Xk

is

II

S Qr

h
]

c TA c

The Classic, or Boole-Schroder, Algebra of Logic


or

161

%
5-95, -[
h

And by
7 54

P\ =
h

-P/.

The condition that

m
,

equations in n unknowns, each of the form


...

F(XI, x 2

x n)

f(xi,

x2

...

xn)
l
,

may

be regarded as simultaneous, is as follows: Let the coefficients in F l 1 in the equation F 1 =f be Px 1 P, Pj 1 and let the QJ, Q Q. in the equation F = f l be coefficients of the corresponding terms in /
1 1
,
, ,
,
. .
.

Mi MJ, Ms
1
,

1
,
.
. .

Ni As AY,
,

let
,

the coefficients of the corresponding

terms in

2
,

in the

equation

F =f
2

be P^,

2 3
,
.
.

and
.
.

let
2
,

the coefficients of the corresponding terms in /

QS, be

M M
2
i
, ,
.

Q 22 Q 32
,

the coefficients of the corresponding terms 2 , 3 Ni m = m P m P m m be m m W in the equation F Pi 2 3 Qi Qz $3 f m let the coefficients of the corresponding terms in / m be Mi m z T m m AT m .... The condition then is A
.

A A
2

let

Ms 2 in F m
,

.,
,

and

Ms m

A
h

II
r

H (Pr

-MS + ~Prh

MS)] X JJ
r

E
h

(Qr*

-N, +

-(? r

A/)]

=
represent
is

or

if

h
r

represent any coefficient in


h
,

Fh
+

whether

P
=

or Q,

and

h
r

the corresponding coefficient in /

whether

M or A, the condition
B
h
r

II

IZ
m

(Ar

-BS

-A

h
r

)}

And

the solution which

one of the unknowns,

.TA-,

such equations give, on this condition, for any h h be the coefficients is as follows: Let P r and r

of those terms, in any one of the equations

Fh = f h

in

which a^

is

positive,

and
is

let

h
r

and

h
r

be the coefficients of the terms, in


is

F = fh
h

in

which x k

negative.

The

solution then

or

a*

- II
By

Z
.

(Qr

-A7/ +

-9,"

AV)] +
.

E [II
a? 2 ,

7-13, ^(.TI,
. .

#2,
0,

ar n )

= / A (-TI
is

...

n)

is

equivalent to
if
h Q and
r

i, .r 2 ,

xn)

where

a function such that

h
r

be coefficients of corresponding terms in of the corresponding term in * will be

Ph
r

and Jf r A be
h

coefficients of

and / h the coefficient * h Q r -Nr + -Q r AV, and if h h the corresponding terms in F and /
,

Fh

coefficient of the corresponding

term
h
r

in
h
r

^ will
+

be

P h -M h + -P h
r r
r r

h
r
.

And -(P r -MS


12

-P

h
r

h
r )

= P

-P/

-M \

Hence the theorem

follows from 7 53.

162 x n)

A
F(XI,
.1*2,
. .

Survey of Symbolic Logic

f(xi,

xz

a- n )

is

a perfectly general equation, since

F
of

and / may be any expressions any number


and
this

in the algebra,

developed as functions of

the variables in question.

7-54 gives, then, the condition and the solution

of simultaneous equations, in

any number
is

of

unknowns,

for

each of the unknowns.


tion,
is its

This algebra particularly lends


It

itself to

generaliza

most general theorem.

the most general theorem

concerning solutions in the whole of mathematics.

Boole

General Problem.

Boole proposed the following as the general


11

problem of the algebra of logic. Given any equation connecting the symbols x, y, ... w, z, .... Re quired to determine the logical expression of any class expressed in any

way by
.
. .

the symbols x, express this:


t

y,

... in
t

terms of the remaining symbols u\

z,

....
z,

We may
)
;

Given

f(x, y,
z,

.)

and
is

$>(.r,

y,

.)

ty(w,

to determine
. .
.

in
.

terms of w,
.

....

This

perfectly general, since

if
is,

x, y,

and w, z, 7-1 and 5-72, a by

for solution
=

V(w,

2,

by any number of equations, there The rule single equation equivalent to them all. = f(x, y, .) and .) may be stated: Reduce both t y, to the form of equations with one member 0, combine them .)
.

are connected

3>(.r,

by addition into a single equation, eliminate x, 7-1, the form of equation with one member
form.

y,
is

and solve

for

t.

By

equivalent to the other

And by

5-72, the

sum

of

equivalent to the equations added.

two equations with one member is Hence the single equation resulting
will contain all the data.
is

from the process prescribed by our rule


result of eliminating will be the

The

complete resultant which

independent

of these,
of
t

and the solution


z,
. . .

for

will

thus be the most complete determination

in

terms of w,

afforded by the data.

Consequences

of

Equations in General.

word

of caution with refer

ence to the manipulation of equations in this algebra may not be out of As compared with other algebras, the algebra of logic gives more place.

room

for choice in this matter.

Further, in the most useful applications

of the algebra, there are frequently

problems of procedure which are not

resolved simply by eliminating this and solving for that.

The

choice of

method must,

then, be determined with reference to the end in view.

But

the following general rules are of service:


(1)

Get the completest possible expression


==

0,

or the least inclusive

possible expression

1.

a+b+

c+

gives

0, 6

0, c

0,

a+ 6

0,

a+

0,

of Thought, p. 140.

The Classic, or Boole-Schroder, Algebra of Logic


etc.

163

But a

will
. .

a+ b

1,

a+

==

not generally give a + b = 0, etc. Also, a == == 1 will not 1, but a + b generally give a == 1.
of equations, with

gives

(2)

Reduce any number


is

which

it is

necessary to deal,

by first reducing each to the form in which and then adding. The various constituent equations can always be recovered if that be desirable, and the single equation gives
to a single equivalent equation,

one

member

other derivatives also, besides being easier to manipulate.

Do

not forget

that

it is

possible so to
If

than the data.


or a + b

we have a =

combine equations that the result is less general and 6 = 0, we have also a = b, or a b = 0,

0,

according to the

mode

of combination.

But a +

is

equivalent to the data, while the other

two are

less

comprehensive.

A
fact,

general

desired

method by which consequences of a given equation, in any 12 and is, in terms, may be derived, was formulated by Poretsky,

a corollary of his

Law

of

Forms, given above.

We

have seen that

this

law

may

a-b + -ab =

be formulated as the principle that if a = b, and therefore and a b + -a -b = 1 then any t is such that a -b + -a b c t
,

and tcab + -a-b, or any


member" of this set.
"zero

the

which contains the

"zero

member"
"

of the set of equations equivalent to a

b,
t,

and

is

contained in the

whole

Now
Law

if

x c

t,

ux c

for

any u whatever, and thus the


arbitrarily
if

member" of

the

of

Forms may be multiplied by any


Similarly,

chosen u which we choose to introduce.

tcy, then

tcy + v,
by the

and the

"whole

member" in

the

Law
v.

of

Forms may be

increased

addition of any arbitrarily chosen

This gives the


t

Law

of Consequences.

7-6

If

b,

then

= (ab + -a

-b + u)

v (a

-b + -a

b) -/,

where u and

are arbitrary.
[7
1

12]

If a

b,

then a -b + -ab
+
v (a

= =

and ab + -a-b =
(1

1.
--

Hence

(a b +

-a -b +
all

it) t

-b + -a

b) -t

?/) t

+ v-0--t

t.

This law includes

First, let us see that it is

the possible consequences of the given equation. more general than the previous formulae of elimina

tion
for
t.

and

solution.

Given the equation

.r

B -x =

0,

and choosing

.1

we should

get the elimination resultant.


0,

ItAx + B -x =

then

AB

(-A x + -B -x + u)

AB

Since u and
fore
12

are both arbitrary

and

may assume

the value

0,

there

AB =

0.

Sept

lois, etc.,

Chap. xn.

164

A
this
is

Survey of Symbolic Logic

But

only one of the unlimited expressions for

AB

which the law

gives.

Letting

u =

0,

and

1,

we have
x+

A B = A -B
Letting u

-A B -x

= A and

B,

we have

AB
And
so on.

AB + -AB-x
AB

be found that every one of the equivalents of which the law gives will be null.
it will

But

Choosing x for our


If

t,

we should

get the solution.

x+

-x

0,

then x

= =

(-A x + -B -x + u) x + (-A + u) x + v B -x.


0,

(A x +

B -x)

-x

Since u and v

may

both assume the value


x

= -A

x,

or

x c

-A

(1)

And

since

u and

may both assume the value 1, x = x + B -x, or B -x ex

But

if

B -x c x,
(1)

then
(2),
1,

Hence,

and
v

B -x = (B B cxc-A.
Law
x
of

-.T)

0,

or

cx

(2)

When u =
of

and

the

Consequences becomes simply the

Law

Forms.

For these values

in the above,

= -A

x+

B -x
in the

which

is

the form which Poretsky gives the solution for x.


of the arbitraries,

The introduction
all

u and
of

v,

Law

of

Consequences
it

extends the principle stated by the

Law

Forms

so that
all

covers not

only

equivalents of the given equation but also

the non-equivalent

As the explanation which precedes the proof suggests, this is accomplished by allowing the limits of the function equated to t to be expressed in all possible ways. If a = b, and therefore, by the Law of
inferences.

Forms,
t

(a b

+ -a -b)

(a

-6 + -a
6,

b) -t

the lower limit of


1,

t,

0, is

expressed as a -b + -a
-b.

and the upper

limit of

/,

is

expressed as

ab + -a
its

In the

Law

of Consequences, the lower


is,

limit, 0, is expressed as v

(a-b + -ab), that


in

can be derived from

expression as
is,

expressed as a b + -a -6 + u, that

ways which a-b + -ab; and the upper limit, 1, is all possible ways which can be derived
in all possible

from

its

expression as a b + -a -b.
t

Since an expression of the form


t

(a b

+ -a -b)

(a

-b + -a

b) -t

The Classic, or Boole-Schroder, Algebra of Logic


or of the form
t

165

(a b

+ -a -b + u)

v (a

-b + -a

b) -t

determines

only in the sense of thus expressing


all

its limits,

and the Law


it

of

Consequences covers
all

possible

ways

of expressing these limits,

covers

possible inferences
is,

from the given equation.

The number

of such

inferences

of course, unlimited.

The number

expressible in

terms of n

elements will be the number of derivatives from an equation with one

member and the other member expanded with reference to n elements. The number of constituent terms of this expanded member will be 2 n and the number of combinations formed from them will be 2 2 Therefore,
,
".

since pi + pz + p$ +

=0

number
of

of consequences of

p z = 0, p s = 0, etc., this is the a given equation which are expressible in terms


gives pi

0,

n elements.

Poretsky gives the sixteen determinations of a in terms of the three elements, a, b, and c, which can be derived from

As one

illustration of this law,

the premises of the syllogism in Barbara:


If all

13

is b,

a-b =
b -c

0, 0,

and

all 6 is c,

then a-b +
a

b -c

0,

and hence,

=
=

a
5 b

= a (-b + c) = a + 6 -c = a b = a (b c + -b -c) _ c + a (b c + _5 _ c = a c = b -c + a c = a (-b + c) + -a b -c = a b c -c + ab c = a (b c + -b -c) + -a b -c = a c + -a b -c = a 6 c + -a b -c


(b

+ -c)

(b

c)
)

The Inverse Problem


that

of

Consequences.

Just as the

Law

of

Conse

quences expresses any if a-b + -ab =

inference from a
0,
1
;

then (a-b + -ab)v

by taking advantage = 0, and if ab + -a-b =


any equation which

of the fact
1,

then a b + -a -b + u
the inference
fl( a 5

so the formula for

will give

+ _a a -b + - a

_fr)

= =

b
i}

can be expressed by taking advantage of the fact that if then ab + -a-b = 1, and if a-b + -ab + u = 0, then
thus get Poretsky
s

l>

o.

We

Law
of v

of Causes, or as
14

it

would

be better translated, the

Law

of Sufficient Conditions.

7-7

If for

some value
t

of

u and some value


t

(ab + -a -b)

+ (a -b + -a b + u)

-t,

then a

b.

[v

= v(ab + -a-b)t+(a-b + -ab + u)-t, = (a b + -a -b) t + (a -b + -a b + u) -t] -t


If
t (

then

[7-1, 5-72]

a -^ + - a

1)

+ u ) -t

= (a-b + -a

b) -t

+ u -i

13
14

Ibid., pp. Ibid.,

98 /. Chap. xxm.

166

A
Hence
(a

Survey of Symbolic Logic

-b + -a

b) -t
t

=
+ -a -b)
t
t,

(1)

Hence

also [5-7]

v (a b

and

[4-9]
b)

t--[v (a b

+ -a -b)]
(a

=
b)
t

(-v

+ a -b + -a

t-v + (a -b + -ab)

Hence

[5

72]

-b + -a
(

=
(t

(2)

By

(1)

and

(2),

a -b + -a
b.

b)

-t)

a -6 + -a

6.

Hence

[7-1] a

Consequences and the Law of Sufficient Conditions are more general than the Law of Forms, which may be derived from either.
of

Both the Law

Important as are these contributions of Poretsky, the student must not


be misled into supposing that by their use any desired consequence or

equation can be found automatically. The only sense in which these laws give results automatically is the sense in

sufficient condition of a given

which they make it possible to exhaust the list of consequences or conditions And since this process is expressible in terms of a given set of elements.
ordinarily too lengthy for practical purposes, these laws are of assistance
principally for testing results suggested
"intuition
".

One has

to discover for himself

by some subsidiary method or by what values of the arbitrages

u and

v will

give the desired result.

V.

FUNDAMENTAL LAWS OF THE THEORY OF INEQUATIONS


.

The

In this algebra, the assertory or copulative relations are = and c denial of a = b may conveniently be symbolized in the customary way:

8-01

4= b is

equivalent to

"a

"a

b is false

".

Def.

We
or
"a

lent to a b

might use a symbol also for = a and to a -b = 0,


4= 0.
",

b is

false.".

But

since a

b is

its

negative

may
b"

be represented by a b

equiva 4= a

by a -b
c

It is less

b is false

since

"a

is

necessary to have a separate symbolism for not contained in is seldom met with in logic
in
true,
is

except where a and b are mutually exclusive,

which case a
then Q
is

=
",

0.

For every proposition of the form

"If

P
it.

is

true

there

is

Q is false, then P is false ad absurdum,oT the simplest form


"

another,

If

".

This

the principle of the rednctio

of

In terms of the relations

and

=|=,

the more important forms of this principle are:


"If

(1)

(2)

"If

= = =

b,

then then
c

6,

= gives = d&ndh =
d",
".

also,
k",

"If

c 4= d,

then a
=(=
<Z,

4= b

".

gives also,

"Ifc

then a

4=

b",

and
then

"If

4= k,

then a
b

4= b

(3)

"If

and
"If

c c

c 4=

d",

and

d,

then h

=
k",

gives also,
4= b
".

"If

and h

4= k,

d and h

4= k,

then a

The Classic, or Boole-Schroder, Algebra of Logic


(4)
"a

167

= =

b is equivalent to c

d"

gives also,

"a

4=

b is equivalent

toe

4=

d".

(5)
"a

"a

b is equivalent to the set, c

=
<7,

k,
".

..,"

gives also,

4= 6 is

equivalent to

Either

c 4=

</

or h

4= A

or ...

16

The
But the
from

general forms of these principles are themselves theorems of the


of propositions"

"calculus

the application of this algebra to propositions.

calculus of propositions, as an applied logic, cannot be derived

this algebra

without a

circle in

the proof, for the reasoning in

demon

stration of the theorems presupposes the logical laws of propositions at

every step.

We

must, then, regard these laws of the reductio ad absurdum,

like the principles of proof previously used, as given us

by ordinary logic, which mathematics generally presupposes. In later chapters, 16 we shall discuss another mode of developing mathematical logic the logistic

method

which avoids the paradox of assuming the principles of logic in order to prove them. For the present, our procedure may be viewed simply
as an application of the reductio ad

absurdum

in

ways

in

which any mathe

matician

feels free to

make

use of that principle.

Since the propositions concerning inequations follow immediately, for the most part, from those concerning equations, proof will ordinarily be

unnecessary.

Elementary Theorems.
sitions are as follows:

The more important

of the elementary propo

8-1

If

ac

4=

be, then a

4= b.

[2-1]

8-12

If

a+c

=t=

c,

then a

4= b.

[3-37]

8-13

4= b is

equivalent to -a

4"

-b.

[3-2]

8-14

a + b

4= b,

ab

4= a,

-a + b

4= 1,

and a -b

4=

are

all

equivalent.

[4-9]

8-15

If

a+b
[5-7]

x and b

4= x,

then a

4=

8-151

If

=
=

and

b 4= x,

then a + b

4= x.

[5-7]

8-16

If

ab

x and b

4=

.r,

then a

4=

1.

[5-71]
15

"Either

...

or

..." is

here to be interpreted as not excluding the possibility that


v.

both should be true.


1(5

Chap,

iv, Sect, vi,

and Chap.

168

A
If

Survey of Symbolic Logic

8-161

and

b 4= x,

then a b

4= x.

[5-71]

8-17

If a

+ b

4=

and a =
and a

0,

then b

4= 0.

[5-72]

8-18

If

a b

4= 1

1,

then

b 4= 1.

[5-73]

8-17 allows us to drop null terms from any sum


a rule by which an equation and an inequation for example, a + b 4= and x = 0.

4= 0-

In

this, it gives

may

be combined.

Suppose,

a+ b

=
if

(a + b)(x

Hence ax + bx
But
x
[8

+ -x) =ax+bx+a-x+b -x. + a-x + b-x 4= 0.

0,

then a x

and

0.

Hence
8-2
If

17]

a-x +

-x

4= 0.

4= 0,

then a + b

4= 0.

[5-72]

8-21

If

4= 1,

then

064=!.
4=

[5-73]

8-22

If

a 6

4= 0,

then a

and

b 4= 0.

[1-5]

8-23

If

+ 6
[4-5]

4=

1,

then a

4= 1

and

b 4= 1.

8-24

If a b 4-

x and a

x,

then

b 4= x.

[1-2]

8-25

If a 4=

and a c6, then 64-0.

[1-9] If

ac6, then a
4=

a.
4= 0.

Hence Hence
8-26
a+ 6
4-

if

and a c

6,

then a 6

[8-22] 6
is

0.
"Either

equivalent to

4=

or 6

4=

".

[5-72]

8-261
as
4=

ai

+ a 2 + a3 +
".

4=

is

equivalent to

"Either

ax

4=

or a 2

4=

or

0, or ...

8-27

a 6

4= 1 is

equivalent to

"Either

4=

or 6

4= 1

".

[5-73]

8-271
3

aia 2
1

+1
=

is

equivalent to

"Either

ai 4= 1

or

a2

4= 1

or

or ...

".

The
tions

difference
b

between 8-26 and 8-27 and their analogues for equa


is

5-72 a +

equivalent to the pair, a

and

0,

and

The
5-73 a
b

Classic, or Boole-Schroder, Algebra of Logic

169

1 is

equivalent to the pair, a

==

and

points to a neces

sary difference
inequations.

between the treatment of equations and the treatment of Two or more equations may always be combined into an

equivalent equation;

two

or

more inequations cannot be combined

into

an equivalent inequation. and b =|= 0. pair, a 4=

But, by 8-2,

a+

b 4=

is

a consequence of the

Equivalent Inequations of Different Forms.

The laws

of the equiva

lence of inequations follow immediately from their analogues for equations.

8-3

4= b is

equivalent to a -b + -a b 4

0.

[7-1]

8-31

4=

1 is

equivalent to -a

=h 0-

[7-12]

8-32

If

&(xi,

2,

...

x n ) and V(xi,

.r 2 ,

x n ) be any two functions of

the same variables, then


,

X2
.

Xn)

4=

is

equivalent to F(XI,

.T 2 ,

xn)
Ai,

4= 0?

where
3,

is

a function of these same

variables
f

and such
,

that,

if

z,

etc.,

be the coefficients in

$ and

Bi 2 3 etc., be the coefficients of the corresponding terms in ^, then the coefficients of the corresponding terms in F will be Ai-Bi + -AiBi,
,

A, -B 2 + -A z

B A -B
2,
3

+ -At B,, etc.

[7-13]

Poretsky
8-33
a
4=

Law

of

Forms

for inequations will be


t

is

equivalent to

3=

a -t+ -at.

[7-15]

Or

in

more general form


a
4= b is

8-34

equivalent to

=h

(ab + -a -b)

(a

-b + -a

b) -t.

[7-16]

Elimination.

The laws governing the

elimination of elements from an

inequation are not related to the corresponding laws governing equations

by the

reductio ad absurdum.

But these laws

follow from the

same theorems

concerning the limits of functions.

8-4

It

x +

B -x

4= 0,

then

[6-3\Ax + B -x
8-41
If

A + B + 0. cA+B. Hence

[8

25]

Q.E.D.
,
.
. .

the coefficients in anv function of n variables, F(XI, x z

x n ),

170
be Ci,

A
C
2,

Survey of Symbolic Logic

To, etc.,

and

if

F(x 1}

.r 2 ,

.r n )

=f=

0,

then

ZC=NO
[0-32]

F(x lf

.r,

a- n )

cC.

Hence

[8-25]

Q.E.D.

Thus, to eliminate any number of elements from an inequation with one member 0, reduce the other member to the form of a normal function
of the elements to be eliminated.

The

elimination

is

then secured by

putting

=f=

the

sum

of the coefficients.

The form

of elimination resultants
It is

for inequations of other types follows

immediately from the above.

obvious that they will be analogous to the elimination resultants of equa tions as follows: To get the elimination resultant of any inequation, take
the elimination resultant of the corresponding equation and replace

= by

=f=

and x by +

A
a
is
is

universal proposition in logic


"

is

represented by an equation:

"All

by a -b

"

0,

No

"

is b

by

a b

0.

Since a particular proposition

always the contradictory of some universal, any particular proposition may be represented by an inequation: "Some a is by a b =f= 0, "Some
b"

by a -b =j= 0. The elimination of the "middle" term from an which represents the combination of two universal premises equation But elimina gives the equation which represents the universal conclusion.
is

not

b"

tion of terms from inequations does not represent


process.

an analogous

logical

Two

particulars

give

no conclusion:

a particular conclusion
is

requires one universal premise.

The drawing

of a particular conclusion

represented by a process which combines an equation with an inequation,

by 8-17, and then


All

simplifies the result,

by 8-22.
. .

For example,
c

aisb,
a
is c,

a-b =
a
c
=J=

0. 0.

a -b
a b
c

0.

Some

+ a -b
4= 0.

c =h 0.

abc

[8-17] [8-22]

Some
"

fr

is c.

6c4=0.

Solution

"

of

an Inequation.
of

An

inequation

may

be said to have a

solution in the sense that for

inequation one

member
4=
is

any inequation involving x an equivalent which is x can always be found.


4=

8-5

x +

-x

equivalent to x

-A

x +

-x.

[7-23]

8-51
i.

A
to

x+

-x

4=

is

equivalent to
or

"Either

-x

or

",

e.,

"Either

B ex

is false

xc-A

is false

".

[7-21

The Classic, or Boole-Schroder, Algebra of Logic


Neither of these
"

171
"

"solutions"
"

determines x even within limits.


x";

B ex
is

is

false

does not
x".

mean
"

is

excluded from
is

it

means only
is
false"

"

not

wholly within

Either

Bex

false

or

xc-A

does not

determine either an upper or a lower limit of x; and limits x only by ex cluding B + u-A from the range of its possible values. Thus "solutions"
of inequations are of small significance.

Consequences and

Sufficient Conditions of

an Inequation.

By

Poret-

sky s method, the formula for any consequence of a given inequation follows from the Law of Sufficient Conditions for equations. 17 If for some value
of

u and some value


t

of

v,

x (a b

+ -a -6)

+ (a -b + -a b + u) -t

then a 8-52

=
If

b.

Consequently, we have by the reductio ad absurdum:


4= b,

then

^= v (a b

+ -a -b)

+ (a -b + -a

+ u)

-t,

where

and

v are arbitrary.

[7-7]

The formula
lows from the

for the sufficient conditions of

an inequation similarly
If

fol

Law
t

of

Consequences for equations.


(a b

b,

then

+ -a -b + u)

v (a

-b + -a

b) -t

where u and
8-53
If for

are arbitrary.
of

Consequently, by the reductio ad absurdum:


of
v,

some value
t

u and some value


t

3=

(a b

+ -a -b + u)

v (a

-b + -a

b) -t

then a

=|=

b.

[7-6]

an Equation and an Inequation. If we have an equation in one unknown, x, and an inequation which involves x, these may be combined in either of two ways: (1) each may be reduced to the form in which one

System

of

member
either.

and expanded with reference to all the elements involved in Then all the terms which are common to the two may, by 8 17,
is

be dropped from the inequation; (2) the equation and this value substituted for x in the inequation.

may
C

be solved for

x,

8-6

If

B -x = and Cx + D -x * 0, then [5-8] If Cx + D-x


x+
=(=

0,

then -.1

x +

-BD -.r

4= 0.

ACx + -ACx + BD-x + -B D


17

-.r

See Poretsky, Theorie des non-egaliUs logiques, Chaps. 71, 76.

172

A
[5-72] If

Survey of Symbolic Logic

Ax + B-x =
and

0,

then
0.

and

B -x =

0,

and hence

A Cx =
Hence
[8

BD

-x
+

=
-B

17] -.4

Cx

-x

4= 0.

The

result here is not equivalent to the data, since

for

one reason
x +

the equation

A Cx+BD
mode

-x

is

not equivalent to
is

B -x =

0.

Nevertheless this

of combination

the one most frequently useful.

8-61

The condition that the equation


4=

x+

B -x =
is,

C x + D -x + -B D 4=

may
x
4=

be regarded as simultaneous
of x

AB =
is

and the inequation and -A C

0,

and the determination


(-A -C +

which they give


-x

A-D)x + (BC + -B D)
is

[7-23]

A x + B -x =

equivalent to x

= -A

x+

-x.

Substi

tuting this value of x in the inequation,

C (-A x + B
or (-A
[8-4]

-x) +

(A x +
4= 0.
is

-B

-x)

4=

C + A D) x+(B C + -BD) -x
condition of this inequation

4=0,
or

(-A+B) C+(A

But the equation


that

+ -B) D 4= 0. x + B -x = +A =

requires that

^45 =

0,

and hence

-A + B = -A and -B
if

-B.

Hence

the equation be possible and

AB =

0,

the condition of the

inequation reduces to

-A C + -B D

4= 0.

[8-4] If the original inequation

this condition is

be possible, then already present in -A C + -B D 4=


if

C+

4= 0.

But
and

0, since

-A C c C

and hence
hence
to
"

[8-25]

if

-B

4= 0,

-.4(74=0, then C 4= 0, and then D 4= 0, while [8-26] C + D 4=

-BDcD
is

equivalent

Either

"Either

C 4= -A C 4=

or

4=

",

and -A C + -B
".

4=

is

equivalent to

or

-B D

4=

Hence the

entire condition of the

system

is

expressed by

AB =
And
[8
5]

and

-A C + -B

4=

the solution of the inequation,


4= 0,
is

(-AC + AD)x + (BC + -BD)-x


x
4=

(-A-C + A-D)x

(BC + -BD)-x
x, in

This method gives the most complete determination of

the form of

an inequation, afforded by the data.

The Classic, or Boole-Schroder, Algebra of Logic


VI.

173

NOTE ON THE INVERSE OPERATIONS,


"DIVISION"

"SUBTRACTION"

AND

It is possible to define

"subtraction"

and
a.

"division"

in the

algebra.

Let a
a.

b be

x such that

+x

And

let

be y such

that

by =

However, these inverse operations are

more trouble than

they are worth, and should not be admitted to the system. In the first place, it is not possible to give these relations a general meaning. We cannot have in the algebra: (1) If a and 6 are elements
in
A",

then a
b is

b is

an element
in

in

K; nor
If

(2) If

a and 6 are elements in K,


y,

then a
it

an element
b
if

K.
a.

must be true that


5-2, a cb.

by

Thus

y a and

But
b

then for some y a, then, by 2-2, a cb y and, y be so chosen that a cb is false, then a b
a
:

b is b

an element,

if

cannot be any element in K. To give a b a general meaning, be required that every element be contained in every element
:

it

would
is,

that

that
x, in

all

elements in
for

be identical.

Similarly,
6

if

b
a.

be an element,

A, then

some

x, it

must be true that

+x

=
if

But

if

+x

a,

then,

by

2-2, b

xca

and, by 5-21,
b
:

be a.

Thus
It

a and b be so chosen
in
A".

that b c a

is false,

then a

cannot be any element

Again, a since a + -a
1

and a

are ambiguous.
1

might be expected that,

=
=

1,

the value of

is

satisfied
1,

by any x such that -a c


is

a would be unambiguously -a. But x. a = x is equiva For 1

lent to x + a

which

equivalent to

-(x + a)

-1

=
x.

-a -x
it

And -a -x =
that, since a -a
:

is

equivalent to -a c
:

Similarly,

might be expected

y,

or a

= 0, the value of a would be unambiguously -a. = y = 0, is satisfied by any y such that y c -a. a y

But
and

y c -a are equivalent.
Finally, these relations can always be otherwise expressed.
of a
:

The value

b is

the value of y in the equation, b y

a.

is

equivalent to

-a b y + a -b + a -y

=
0.

The equation
condition,
is

of condition here is a -b

And

the solution, on this

+u

(a

+ -b)

= ab + u-a-b, where u

is

undetermined.
b

The value
is

of a

b is the value of

x in the equation,

+x

a.

+x

equivalent to

-a

+ -a x + a -b -x

174

Survey of Symbolic Logic

The equation
condition,
is

of condition here

is,

-a

0.

And

the solution, on this

a -b + v a

a -b + v a

6,

where

v is

undetermined.

of the expression,

In each case, the equation of condition gives the limitation of the meaning and the solution expresses the range of its possible values.

CHAPTER
There are four applications

III

APPLICATIONS OF THE BOOLE-SCHRODER ALGEBRA


of the classic algebra of logic

which are

commonly
relations.

considered:

(1) to spatial entities, (2) to the logical relations

of classes, (3) to the logical relations of propositions, (4) to the logic of

The

application to spatial entities


line,

may

be

made

to continuous

and

discontinuous segments of a

or to continuous

and discontinuous regions

in a plane, or to continuous

dimensions.

Segments

of a line

and discontinuous regions in space of any and regions in a plane have both been

used as diagrams for the relations of classes and of propositions, but the
application to regions in a plane gives the

more workable diagrams,

for

obvious reasons.

And

since

it

is

only for diagrammatic purposes that

the application of the algebra to spatial entities has any importance,


shall confine

we

our attention to regions in a plane.

I.

DIAGRAMS FOR THE LOGICAL RELATIONS OF CLASSES


a, b, c, etc.,

For diagrammatic purposes, the elements of the algebra,


will

denote continuous or discontinuous regions in a given plane, or in a 1 represents the plane (or circumscribed circumscribed portion of a plane.
portion)
itself.
is

the null-region which

every region.
a,
i.

e.,

not-a.
b.

For any given region, a, The "product", a xb or a


If
b,
b,

is supposed to be contained in -a denotes the plane exclusive of


b, is
",

that region which


is

is

com
0.

mon
The
over.

to a and
"sum",

a and 6 do not

"overlap

then a b

the null-region,

a+

determining a +

denotes the region which is either a or b (or both). In the common region, a b, is not, of course, counted twice
a + b

a -b + a b + -a

b.

This

is

a difference between + in the Boole-Schroder Algebra and the

= b, signifies that a and b denote the same equation, a ac6 signifies that a lies wholly within b, that a is included or region. contained in b. It should be noted that whenever a = b, a c b and b c a.
of arithmetic.

The

Also, a

c a holds always.
.

Thus the

relation

is

analogous not to

<

in

arithmetic but to

175

176

A
While the laws of

Survey of Symbolic Logic

this algebra hold for regions, thus denoted,

however

those regions

may

be distributed in the plane, not every supposition about


is

their distribution
classes.

equally convenient as a diagram for the relations of

All will be familiar

with Euler
"All
"

diagrams, invented a century


is

earlier

than Boole

algebra.
"

is

6"

represented by a circle a a and


b,

wholly within a
intersect;

circle b;

No

is b

by two
a
is

circles,

which nowhere

"Some

is

6"

and

"Some

not

6"

by

intersecting circles,

sometimes with an asterisk to indicate that division of the diagram which The defects of this style of diagram are obvious represents the proposition.
:

All a

is

No

is

cuto;
Some
a
is &

x-*>c^\

^-^^\
Some
a
is

not 6

FIG. 1

the representation goes beyond the relation of classes indicated by the propo
sition.

In the case of

"All

is

&",

the circle a
b is

falls
a",

within b in such wise

as to suggest that

we may

infer

"Some
"No

not
b"

but this inference

is

not valid.

The representation
b",

of

a
is

is

similarly suggests

"Some

things are neither a nor

which

also

unwarranted.

With these
is

dia

grams, there

is

no way

of indicating

whether a given region


is

null/

But
Yet

the general assumption that no region of the diagram


misinterpretations

null leads to the


similar.

mentioned,

and

to

others

which are

Euler
still

diagrams were in general use until the invention of Venn, and are
specifically to represent the relations

doing service in some quarters.

The Venn diagrams were invented


of these

of logical classes as treated in the Boole-Schroder Algebra. 1

The

principle

diagrams is that classes be represented by regions in such relation to one another that all the possible logical relations of these classes can be
indicated in the same diagram.
for

any

possible relation of the classes,

then be specified
null.

the diagram initially leaves room and the actual or given relation, can by indicating that some particular region is null or is notis,

That

Initially the

or

1.

diagram represents simply the "universe of discourse", For one element, a, 1 = a + -a. 2 For two elements, a and b,
1

(a

+ -a) (b + -b)

a b + a -b + -a b + -a -b

See Venn, Symbolic Logic, Chap. v. The first edition of this book appeared before Schroder s Algebra der Logik, but Venn adopts the most important alteration of Boole s the non-exclusive interpretation of a + b. original algebra 2 See above, Chap, n, propositions 4-8 and 5-92.

Applications of the Boole-Schroder Algebra

177

For three elements,


1

a, b,

and

c,

(a

+ -a) (b + -b) (c + -c)

a b c + a b -c + a -b

+ -a b
b -c

+ a -b -c
c

+ -a

+ -a -b

+ -a -6 -c

Thus the

"

universe of

discourse"
r(

for

any number

of elements, n,

must

correspond to a diagram of 2

expansion of

1.

If

divisions, each representing a term in the the area within the square in the diagram represent

-a
FIG. 2

-a-b

the universe, and the area within the circle represent the element a, then the remainder of the square will represent its negative, -a. If another element,
6, is

to be introduced into the

same universe, then

may

be repre

sented by another circle whose periphery cuts the first. The divisions, (1) into a and, -a, (2) into b and -b, will thus be cross-divisions in the uni
verse.
If

a and b be classes, this arrangement represents a


b,

all

the possible

subclasses in the universe;

those things which are both a and b;

a-b, those things which are a but not b; -ab, those things which are 6 but not a; -a -6, those things which are neither a nor b. The area which
represents the product, a
b,

will readily

be located.

We

have enclosed

by a broken line, in figure 2, the area which represents a + b. The negative of any entity is always the plane exclusive of that
For example, -(a b + -a -6), in the above, two divisions of the diagram, a -b + -a 6.
If it

entity.

will

be the

sum

of the other

be desired to introduce a third element,

c,

into the universe,


into

it is

necessary to cut each one of the previous subdivisions

two

one

part which shall be in c and one part which shall be outside c. This can be be accomplished by introducing a third circle, thus It is not really necessary to draw the square, 1, since the area given to the
figure, or the

whole page,
is

may

as well be taken to represent the universe.


it

But when the square


13

omitted,

must be remembered that the unenclosed

178
area outside

A
all

Survey of Symbolic Logic


is

the lines of the figure


-ft,

a subdivision of the universe

the entity -a, or -a


involved.

or -a -b -c, etc., according to the

number

of elements

-a-b-c

FIG. 3
If

a fourth element,

d,

be introduced,

it is

no longer possible to repre

sent each element

by a

circle, since

a fourth circle could not be introduced

each previous subdivision into two parts one part in d and one part outside d. But this can be done with ellipses. 3 Each
in figure 3 so as to cut

-a-b-c-d

FIG. 4
3

We have deformed the ellipses slightly

and have indicated the two points

of junction.

This helps somewhat in drawing the diagram, which is most easily done as follows: First, draw the upright ellipse, a. Mark a point at the base of it and one on the left. Next,

Applications of the Boole-Schroder Algebra

179

one of the subdivisions in figure 4 can be "named" by noting whether it is in or outside of each of the ellipses in turn. Thus the area indicated by
6
is

a b c -d, and the area indicated by 12


it

is

-a -b

c d.

With a diagram

of
c,

four elements,

requires care, at

first,

to specify such regions as a +

These can always be determined with certainty by developing each term of the expression with reference to the missing ele
a
c
d,

+b

+ -d.

ments. 4

Thus

ac +

= ac
a b
c

(b

+ -b) (d + -d) +
b c

b
c

d (a + -a) (c + -c)
d + a -b
c

d+ a

-d + a -b

-d + a

b -c

d + -a

b c

d + -a b -c d 4 by the

The terms
divisions

of this

sum,

in the order given, are represented in figure


6,

numbered

10,

9, 5,

14,

11,

15.

Hence ac + bd
one

is

the region

which combines these.

With a
it

little

practice,

may

identify such
b

regions without this tedious procedure.


easily identified
1, 4, 5, 8.

Such an area as
2, 3, 6, 7,

+ -d

is

more
and

by inspection:

comprises

10, 11, 14, 15,

Into this diagram for four elements,


e, if

it is

possible to introduce a fifth,


5.
it

we

let e

be the region between the broken lines in figure


diagram"

The

principle
all

of the

"square

(figure 6) is the

same

as

Venn

s:

represents

FIG. 5

draw the horizontal ellipse, d, from one of these points to the other, so that the line con necting the two points is common to a and d. Then, draw ellipse 6 from and returning to the base point, and ellipse c from and returning to the point on the left. If not done in
this

way, the

first

attempts are likely to give twelve or fourteen subdivisions instead of

the required sixteen. 4 See Chap, n, 5-91.

180

Survey of Symbolic Logic

the subclasses in a universe of the specified

number

of elements.

No

diagram

is

really convenient for

more than four elements, but such are


-a

-a

-a

a-b

-a -b

FIG. 6

frequently needed.
slightly the square

The most convenient


-a

are those

made by modifying
Figure 7

diagram of four terms, at the right

in figure 6. 5

b<

-b

h
-e

-e

-h
FIG. 7

gives,

by

this

method, the diagrams


for seven

for five

and

for six elements.


is

We

give

also the

diagram

(figure 8)

since this

frequently useful and

not easy to

make in any other way. The manner in which any function

in the algebra

may

be specified in a

diagram of the proper number of divisions, has already been explained. We must now consider how any asserted relation of elements any incluwe
See Lewis Carroll, Symbolic Logic, for the particular form of the square diagram which adopt. Mr. Dodgson is able, by this method, to give diagrams for as many as 10 terms, 1024 subdivisions (p. 176).
6

Applications of the Boole-Schroder Algebra


sion, a

181

b,

or

any equation, a =
relation, or

b,

or inequation, a

=)=

may

be repre

sented.

such relations, can be completely in these diagrams by taking advantage of the fact that they specified

Any such

any

set of

or to the can always be reduced either to the form of an expression = form of an expression 4= 0- Any inclusion, a c b, is equivalent to an equa
tion,

a -6

one of the

And every equation of the form a = b form a -b + -a b = O. 7 Thus any inclusion or


O.
6

is

equivalent to

equation can be

represented by some expression form a --r b is equivalent to one


asserted relation

any inequation of the Thus any of the form a-b + -ab =H O. 8 whatever can be specified by indicating that some region
Similarly,
is null,

= 0.

(continuous or discontinuous) either

=
{

Oj, or

is

not-null,

0}.
in

We

can illustrate

this,

and at the same time indicate the manner

which such diagrams are


Given: All a
.

useful,

by applying the method to a few syllogisms.


a -b
b -c

is 6,
is c,

a cb,
b cc,

and

All b

= =

0.

0.

6 7 8

See Chap, n, 4-9. See Chap, n, 6-4. See Chap, n, 7-1.

182

A
have here indicated
striking
it

Survey of Symbolic Logic


the a which
T

We
by
that

(figure 9) that a -b
lines).

is

not

is

null

out (with horizontal

Similarly,

w e have

indicated

by striking out b -c (with vertical lines). Together, the two have eliminated the whole of a -c, thus indicating that a -c = 0, operations
all b is c

or

"All

is c

".

FIG. 9

For purposes

of comparison,
9

we may

derive this

same conclusion by

algebraic processes.

Since a -b

and and

b -c

= =

= =

therefore, a -b c

+ -c) = a -b c + a -b -c, b -c (a + -a) = a b -c + -a b -c, + a -b -c + a b -c + -a b -c = 0,


a -b
(c

[5-72] a b -c + a -b -c

a -c

(b

+ -b)

a -c.

The equation in the third line, which combines the two premises, states The last exactly the same facts which are represented in the diagram.
equation gives the conclusion, which results from eliminating the middle term, b. Since a diagram will not perform an elimination, we must there
"look for"

the conclusion.
illustration of this kind
is b,
:

One more

Given: All a

a -b
b c

and No

b is c,
is

= =

0.

0.

The

first

premise

indicated (figure 10) by striking out the area a -b (with

horizontal lines), the second

by

striking out

be (with

vertical lines).
c,

To

gether, these operations have struck out the whole of a

giving the con

clusion
9

ac =

0,

or

"No

is

c".

orem

in

Throughout this chapter, references in square brackets give the number Chap, u by which any unobvious step in proof is taken.

of the the

Applications of the Boole-Schroder Algebra


In a given diagram where
all

183

the possible classes or regions in the uni

verse are initially represented, as they are by this

method
null or

of

diagramming,

we

cannot presume that a given subdivision

is

is

not-null.

The

actual state of affairs

may

require that
null

some regions be
not.

null, or

that some

be not-null, or that

some be

and others

Consequently, even

when we have struck out the


the premises.
All

regions which are null,

we cannot presume
null-regions,

that all the regions not struck out are no/-null.

This would be going beyond

we can

say,

when we have struck out the


then,

is that, so far as the premises represented are concerned,

any region not

struck out

may
is

be not-null.

If,

we wish

to represent the fact that a

given region
there
is

definitely not-null

that a given class has members, that


indicate this
it is

mark
a

in
b

some expression 4= the diagram. For


=}=

we must
this purpose,

by some
b.

distinctive

convenient to use asterisks.

That a

0,

may

be indicated by an asterisk in the region a


If

But here

further difficulty arises.

say, a,

a
it

b -c.

the diagram involve more than two elements, a b c and b, and c, the region a b will be divided into two parts, a b 4= 0, does not tell us that a b c =}= 0, and Now the inequation,
tell

does not
then,

us that a 6 -c

4= 04=
it

It tells us

only that a b

+a

b -c

=|=

0.

If,

we

wish to indicate a b

by an asterisk
that
is all

in the region a b,

we
c.

shall not be

warranted

in putting

either inside the circle c or outside

It belongs in one or the other or both

we know.

Hence

it is

convenient to indicate a b
of a b

4=

by placing an
by a broken

asterisk in each of the divisions

and connecting them


is

line,

to signify that at least one of

these regions

not-null (figure 11).

ab =0

FIG. 11

We

shall

show

later that a particular proposition

is

best interpreted

by
0.

an inequation;

"Some

is

6",

the class a b has

members, by ab

184

A
we have:
is b, is c,

Survey of Symbolic Logic

Suppose, then,

Given:

All a

a-b =
a c
b is
c",

0.

and Some a

H= 0.
is

The

conclusion,

"Some

indicated (figure 12) by the fact that

one of the two connected asterisks must remain the whole region a b c + a -be cannot be null. But one of them, in a -b c, is struck out in indi
cating the other premise, a -6

0.

Thus a

b c

=}=

0,

and hence a

c 4= 0.

FIG. 12

The
by

entire state of affairs in a universe of discourse

may

be represented

by and remembering that any region which is neither struck out nor occupied by an asterisk is in doubt. Also, the separate subdivisions of a region occupied connected asterisks are in doubt unless all but one by
are not-null,
of these connected asterisks

striking out certain regions, indicating

asterisks that certain regions

any regions which are


course, be

left in

made

specifically

occupy regions which are struck out. And doubt by a given set of premises might, of null or not-null by an additional premise.

In complicated problems, the use of the diagram is often simpler and more illuminating than the use of transformations, eliminations, and solu
tions in the algebra.
tions, the

All the information to be derived


(for

diagram gives

one who can

"see"

it)

at a glance.

from such opera Further


diagrams in con

illustrations will be unnecessary here, since

we

shall give

nection with the problems of the next section.


II.

THE APPLICATION TO CLASSES

interpretation of the algebra for logical classes has already been 10 explained. a, b, c, etc., are to denote classes taken in extension; that is 10 Chap, n, pp. 121-22.

The

Applications of the Boole-Schroder Algebra


to say, c signifies, not a class-concept, but the aggregate of
all

185
the objects

denoted by some class-concept. Thus if a b, the concept of the class a not be a synonym for the concept of the class b, but the classes a and b may

must
nifies

consist of the

same members, have the same extension,


of the class a
is

acb

sig
b.

that every
product",

member
b,

also a

member

of the class

The

"

bers of a

and members

denotes the class of those things which are both mem of b. The "sum a + 6, denotes the class of those
",

things which are either

members

of a or

members

of b (or

members

of both).

denotes the null-class, or class without members.

Various concepts

may

an empty class "immortal men", "feathered invertebrates", "Julius Caesar s twin," etc. but all such terms have the same extension; they denote nothing existent. Thus, since classes are taken in extension,
denote
there
is

but one

null-class, 0.

Since

it

is

a law of the algebra that, for

every

x,

the null-class
bers of

ex, we must accept, in this connection, the convention that is contained in every class. All the immortal men are mem
class, since there are
"universe

any
the

no such.
",

represents the class


"universe
".

"every

thing
is

",

of discourse

or simply the

This term
if

pretty well understood.


of the class a,
"universe
X"

But
A"

it

may

be defined as follows:

a n be any

member
then the
"a

and

represent the class-concept of the class x,

of

discourse" is

the class of
If

all

the classes,

x,

such that
blind"

is

an

is

either true or false.


"fixed

"The

fixed stars are


class

is

neither true nor false, then

stars"

and the

"blind"

do not

belong to the

same universe
of a, -a,
is

of discourse.

The negative
of discourse:

a class such that a and -a have no

members
+ -a
==

in

common, and a and -a between them comprise everything


a -a

in the universe

0,

"Nothing is
-a".

both a and

not-a",

and

1,

"Everything is

either a or
6,

Since inclusions, a c

equations, a

b,

and inequations, a

4= b,

repre

sent relations which are asserted to hold between classes, they are capable
of being interpreted as logical propositions.

And
would

the operations of the


are

algebra

transformations,

eliminations,

and
It

solutions

capable

of

interpretation as processes of reasoning.

hardly be correct to

say that the operations of the algebra represent the processes of reasoning

from given premises to conclusions: they do indeed represent processes of reasoning, but they seldom attain the result by just those operations which are supposed to characterize the customary processes of thinking.
In fact,
it is

the greater generality of the symbolic operations which


*

makes

their application to reasoning valuable.

186

A
The representation

Survey of Symbolic Logic

of propositions

by

inclusions, equations,

and

in

equations, and the interpretation of inclusions, equations, and inequations


in the algebra as propositions, offers certain difficulties,

due to the fact

that the algebra represents relations of extension only, while ordinary logical

In discussing the representation of the four typical propositions, we shall be obliged to


propositions quite frequently concern relations of intension.

consider some of these problems of interpretation.

The
(1) (2)

"

universal affirmative,
a

All a

is

6",

has been variously represented as,

= =

b,

fl-c6,

(3)

v b,

where

v is

undetermined,

(4) a -b

0.

All of these are equivalent. 11

The only
its

possible

doubt concerns

(3)

v b,

where

v is

undetermined.

But

equivalence to the others

may

be

demon

strated as follows:
[7- 1]

v b is

equivalent to a--(v

b)

+ -a
v b

v b

0.
v b.

But a
Hence

-(v b) + -a v b
[5-72]
if

a (-v + -b) + -a

a -v + a -b + -a

vb, then a -b

0. v
(i.

And

if

b}

then for some value of


of
"All

e., v

a),

v b.

These equivalents
(1)

is

6"

would most naturally be read:


s

The a
is

are identical with those things which are a


b:

and

6 s both.
b.

(2) a
(3)

contained in
class a
is

every

member

of a

is

also a

member

of

The The

identical with

some (undetermined) portion


of a but not

of the

class

6.

(4)

class of those things

which are members

members

of 6

is null.

If

we examine any one

of these symbolic expressions of

"All

is

6",

we

not only may it hold when a 0, but it always = 0-6, c 6, and 0--6 = 0, will be true for every when a = 0. = b for some value of element b. And is always true for v = 0. = means that a has no members, it is thus clear that the algebra Since a
shall discover that

holds

"0

v"

requires that

"All

is

6"

be true whenever no members of a

exist.

The

actual use of language

that

"All

ambiguous on this point. We should hardly say sea serpents have red wings, because there aren t any sea ser
is

pents";
11

yet

we understand the hero


%

of the novel

who

asserts "Whoever

See Chap. n. 4-9.

Applications of the Boole-Schroder Algebra


enters here
assert that
is

187

must pass over


any one

my

dead

body".

This hero does not mean to

will enter the

defended portal over his body: his desire

that the class of those


is

who

enter shall be null.


serpent"

The

difference of the

two cases
concept
portal"

this: the concept

"sea

does not necessarily involve the


"those

"having

red

wings",

while the concept of

who

enter the

as conceived

by the hero

does involve the concept of passing


of

over his body.

We

readily accept and understand the inclusion


of the

an

empty

class in

some other when the concept

of the other

when

the relation

is

one of intension.

one involves the concept But in this sense, an


but in some only.
.r,
ex",

empty

class is

not contained in any and every

class,

In

order to understand this law of the algebra,

"For

every

we must

bear in mind two things:


only,

(1)

that the algebra treats of relations in extension

that ordinary language frequently concerns relations of intension, and is usually confined to relations of intension where a null The law does not accord with the ordinary use of language. class is involved.

and

(2)

This

is,

however, no observation upon

its

truth, for
is

it is

a necessary law

of the relation of classes in extension.

It

an immediate consequence of
"All

the principle,
are also

"For

every

y,

cl",

that

is,

members

of

any

class, y,

members

of the class of all

things".

One cannot accept

this last

without accepting, by implication, the principle that,


class is contained in

in extension, the null-

every

class.
is

The

interpretation of propositions in which no null-class


difficulty,

involved

is

not subject to any corresponding

both because

in

such cases the

relations predicated are frequently thought of in extension and because the relation of classes in extension is entirely analogous to their relation in

intension except where the class

or the class
all cases,

is

involved.

But the

interpretation of the algebra must, in

be confined to extension.
in the algebra as stating

In brief:

"All

is

6"

must always be interpreted


class,
b",

a relation of classes in extension, not of class-concepts, and this requires


that,

whenever a

is

an empty
"No

"All

is

6"

should be true.

The
do not

proposition,
6",

is

is

represented by

ab =
of a

"Nothing is

both a and

or

"Those
"No

things which are

members

and

of b both,

exist".

Since

is

6"

is

equivalent to

"All

is not-6", it

may
v is

also be represented

by

a -b

-b, a

c-6, 6c-a,

or a

=
is

v -b,

where

undetermined.

In the case of this proposition, there


the ordinary use of language.

no discrepancy

between the algebra and

has been a problem to representation of particular propositions not clearly conceived the symbolic logicians, partly because they have

The

188

Survey of Symbolic Logic


tried to stretch the algebra to cover traditional
If
"Some

relations of classes relations

and have

which hold

in -intension only.

is

6"

that

it is false

when the
"All

class a
6",

has no members, then


"All
"

"Some

be so interpreted a is will
6"

not follow from

a
"

is

for
is

is

6"

is

true whenever a
0,

0.

But
diffi
6"

on the other hand,


culties:
(1) this

if

Some a

be true

when a =

we have two

does not accord with ordinary usage, and (2) "Some a is will not, in that case, contradict "No a is For whenever there are no
6".

members
"Some

of a
is
6"

(when a = 0), "No a can be true when a =

is

6"

(a b

0) will be true.

Hence
a
is

if

0,

then

"Some

is

b"

and

"No

6"

can both be true at once.


that
"Some

is

b"

solution of the difficulty lies in observing as a relation of extension requires that there be some a
of the class a exist.
"Some

The

that at least one

member

are interpreted in extension,


is
6",

is

6"

does not follow from

Hence, when propositions "All a


is
6"

precisely because

whenever a =
"All

0,

"All

will be true.

But

"Some

is

b"

does follow from

is 6,

and members
a

of a

exist".

To
is

interpret properly

"Some

a
6".

is

6",

we need only remember that


"No

it

the contradictory of

"No

is

Since
0,

is

6"

is

interpreted

by

a b
are

0, "Some

is

6"

will

be a

=(=

that

is,

"The

class of things

which

members

of a

and

of b

both

is not-null".

It is surprising

what blunders have been committed


"Some

of particular propositions.

is

y"

in the representation has been symbolized by x y = v,

where

v is

undetermined, and by u x

v y,

where u and

v are

undetermined.

Both

of these are incorrect,

and

for the

same reason: An
1

"undetermined"

element

may have

the value

or the value

or

any other

value.
all.

Conse
are

quently, both these equations assert precisely nothing at both of them true a priori, true of every x and y and in

They
But

all cases.

For

them

to be significant,

u and

must not admit the value

0.

in that

case they are equivalent to x y =J= 0, which is much simpler defined laws which are consonant with its meaning.

and obeys

well-

Since

we

are to symbolize
"Some

"All

is

b"

by a -b =
=|=

0, it is clear

that

its

contradictory,

is

not

6",

will

be a -b

0.

To sum
follows:

up, then:

the four typical propositions will be symbolized as

A. All a

is b,

a -b

E.
I.

No

is b, is b, is

= 0. = 0. ab
a b
b,
=|=

O.

Some a Some a

0.
=}=

not

a -b

0.
12

Each
12

of these four has various equivalents: See Chap. 11, 4 9 and 8 14.

Applications of the Boole-Schroder Algebra


A. a -b
all

189

0,

b,

-a +

= =
1

1,

-a + -b

-a, a

b,

and -b c-a are

equivalent.

E. a b
all

=
4
1

0,

=
4=

a -b, -a + -b

1,

-a +

-a,

ac-6, and be -a
s

are

equivalent.
I.

a b

0, 0,

a -&, -a + -6 4

1, 1,

and -a +

4 -a are
4=

all all

equivalent. equivalent.

O. o -6

=}=

4=

6,

-a +

and -a + -b

-a are

The reader will easily translate these equivalent forms for himself. With these symbolic representations of A, E, I and O, let us investigate
the relation of propositions traditionally referred to under the topics,
"The

Square of

Opposition",

and

"Immediate Inference".

That the
is

traditional relation of the


If

two

pairs of contradictories holds,


=f=

at once obvious.

a -b
is

=
is

is

true, then a -6

is

false;

if

a -b

is false,

then a -b

=)=

true.

Similarly for the pair, a b


defined:

and a

b 4= 0.

The

relation of contraries

Two

propositions such that both

may

be false but both cannot be true are

"contraries".

This relation

is

traditionally asserted to hold between

A
a

and E.

It does not hold in

ex

tension:

it fails

to hold in the algebra precisely


is

the two propositions

a null-class.

If

0,

whenever the subject of and a b = O. 13 then a -b =

That

is

to say,
"All

if

no members of a
6"

extension,

is

and

"No

is

6"

exist, then from the point of view of But if it be assumed are both true.

or stated that the class a has

members

(a

=|=

0),

then the relation holds.

(b
if

+ -b) a

a b + a -b.

Hence

4= 0,

then a

+ a -b

4= 0.

[8-17] If a b + a -b 4=

and a -b =

0,

then a

b 4= 0.
4= 0.

(1) (2)

And

if

a b + a -b

=h

and a

b
:

0,

then a -b

We may
(1)
is false.

read the last two lines


there are

If

members

of the class a

and

all

is b,

then

"No

is

6"

(2)
is false.

If

there are

members

of the class a

and no a

is b,

then

"All

is

6"

By

tradition, the particular affirmative should follow

from the universal

affirmative, the particular negative

from the universal negative.


0.

As has

been pointed out, this relation fails to hold when a = T ever a 4= 0. e can read a b 4= 0, in (1) above, as

But
a

it

holds
6"

when

"Some

is

instead

"

of of
"

No
13

is
is

b b

is
is

false",

and

a -b 4= 0, in (2), as

"Some

is

not

6"

instead

All a

false

".

We

then have

See Chap, n, 1-5.

190
If If

A
(1) (2)
"

Survey of Symbolic Logic


of a, of a,

there are there are

members members
"

and

all

is b,
is b,

then some a

is b. is

and no a

then some a

not

b.

Subcontraries

are propositions such that both cannot be false but

both

may

be true.

Traditionally

"Some

is

6"

and

"Some
=j=

is

not

6"

are subcontraries.
false,
"

But whenever a =
and a -b

0,

a b
a
"

4=

and a -6
0, it

are both

and the
is

relation fails to hold.


b
is false
",

is

Some a
(1)

When = is

holds.
is

Since a b
1

=
we

Some
a

not b

is

false",

can read
(1)
is

and

(2)

above:
of a,

If there are
b.

members

and

"Some

is

6"

is false,

then some a

not
(2)

If there are

members

of a

and

"Some

is

not

6"

is false,

then some

is 6.

To sum
class

up, then: the traditional relations of the

"square

of

opposition"

hold in the algebra whenever the subject of the four propositions denotes a

which has members.

When

the subject denotes a null-class, only

the relation of the contradictories holds.


are, in that case,

The two

universal propositions

both true, and the two particular propositions both false. The subject of immediate inference is not so well crystallized by tradi

tion,

and

class

for the good reason that it runs against this very difficulty of the without members. For instance, the following principles would all
:

be accepted by some logicians


"No

is

6"

gives gives

"Xo

b is b

a".

"No

b is
b is

a"

"All

is not-a".

"All

not-a"
not-a"

gives gives
gives
(6)

"Some
"Some

b is

not-a".

"Some

b is
"Xo

not-a

is

6".

Hence

is

6"

"Some
"

not-a

is

6".

Xo cows
implies

(a) are inflexed

gasteropods

implies
(a)

"Some

non-cows are
"

inflexed gasteropods":
"Some

"Xo

mathematician

has squared the circle


circle".

(b)

non-mathematicians have squared the

These

infer

ences are invalid precisely because the class b


cessful circle-squarers
is

inflexed gasteropods, suc


it was presumed Those who consider the

an empty
"Some

class;
b is

and because

that

"All

b is

not-a"

gives

not-a".

algebraic treatment of null-classes to be arbitrary will do well to consider

the logical situation just outlined with some care.


particular

The

inference of any

proposition

from

the

assumption that either the

class

corresponding universal requires the denoted by the subject of the particular

proposition or the class denoted by its predicate ("not-6" regarded as the is a class which has members, predicate of "Some a is not
6")

Applications of the Boole-Schroder Algebra

191

The
tive
is

"

conversion"

of the universal negative

and
is

of the particular affirma


",

validated by the law


a".

ab =
a
is

"

b a.
6",

Xo
4= 0,

a b

0,

gives b a

0,

"No

b is

And
a
is

"Some

a 6

gives b a
4= 0,
is

4= 0, "Some b is
"Some

a".

Also,

"Some

not
of

6",

a -6 4

0,

gives -6 a

not-6

is

a".

The

"converse"

the universal affirmative

simply the

"converse"

of the corresponding particular, the inference of

which from the universal

has already been discussed.

What

are called

"obverses"

i.

e.,

two equivalent propositions with


is

the same subject and such that the predicate of one


predicate of the other
or

the negative of the

are merely alternative readings of the

same equation,
is

a depend upon the law, -(-a) a is is also "Xo a which is


"All
6",

14
.

Since x y

=
And

"

is

Xo x
a

",

a -b
is

0,

is not-6".

since a b

equiva

lent to a -(-b)

0,

"Xo

is

6"

is

equivalent to

"All

is not-6".

A
S

convenient diagram for immediate inferences can be

made by

putting

(subject)

and

(predicate) in the center of the circles assigned to them,


divisions of
-*S,

-S between the two


divisions.

and -P between

its

two constituent
which the dia-

The

eight arrows indicate the various

ways

in

(Mverl Prop,

Converse

-S-P

FIG. 13

and thus suggest all the immediate inferences which are valid. For example, the arrow marked "converse" indicates the two terms which will appear in the converse of the given proposition and the

gram may be

read,

order in which they occur.


14

In

this diagram,

we must

specify the null

and

See Chap, n, 2-8.

192

Survey of Symbolic Logic


if

not-null regions indicated

wish
If

add the
"No

by the given proposition. And we may qualification that the classes, S and P, have members. is and S and P have members:
P",

we

SP

=
5.

4=

-4.

-S-P

7.

6.

FIG. 14

Reading the diagram of figure 14 in the various possible ways, we have: 1. No S is P, and 1. Some S is not P. (According as we read what
is

indicated by the fact that

is null,

or

what

is

indicated by the fact

that S
2.
3.

-P
All

is

not-null.)
is

not-P,

and

2.

Some S

is

not-P.

All

4. 5.
6. 7.

P is not-S, and 3. Some P is not-S. No P is S, and 4. Some P is not-S.


P.
S.

Wanting.

Some not-S is Some not-P is


Wanting.

8.

Applications of the Boole-Schroder Algebra


if

193

Similarly,

"All

is

P",

and S and
0,
f.

P
0,

have members:

S -P =

P
5.

4=

FIG. 15

Reading from the diagram


1.

(figure 15),

we have:

All

is

P,

and

1.

Some S

is

P.

2.

No S

is

not-P.

3.

Wanting.

4.
5. 6. 7.

Some P is S. Some not-$ is


Wanting. No not-P
All
is is

not-P.

S.

8.

not-P

not-S,

and

8.

Some not-P

is

not-S.

The whole
trivial.

so simple as to be almost subject of immediate inference is Yet in the clearing of certain difficulties concerning null-classes

the algebra has done a real service here.


results of syllogistic reasoning algebraic processes which give the in those examples we carried out the have already been illustrated. But

The
14

194

Survey of Symbolic Logic

operations at unnecessary lengths in order to illustrate their connection

with the diagrams.

The premises

of

concerns, altogether, three classes.

any syllogism give information which The object is to draw a conclusion

which gives as much of


the
"middle"

this information as
is

term.

This

can be stated independently of the kind of result which elimination exactly


is

gives in the algebra.

And
it

elimination
is

very simple.
O.
15

eliminating x from
of a syllogism
is

A x + B -x
may
of

AB =
is 0,

The result of Whenever the conclusion

universal,

be obtained by combining the premises

in a single equation

one

member

which

and eliminating the

"middle"

term.

For example:

No

is y,

x y
z

0. 0.

All z

is x,

-x

Combining

these, x y + z -x
x,
is
"No

=
=
or

0.
0.

Eliminating

y y
is

Hence the

valid conclusion

z",

"No

z is

y".

Any

syllogism with a universal conclusion


"If

may
a
b,

also be symbolized so
b cc,

that the conclusion follows from the law,

c6 and

then

ace".

By

this

method, the laws, -(-a)

a and

"If

a c

then -b

c-a",

are

some

times required also. 16

For example:

No

is y, is x,

x c -y.
z

All z

ex.
"No

Hence
There
is

c-y, or

"No

z is

y",

and y c-z, or

is

z".

no need to treat further examples of syllogisms with universal conclusions they are all alike, as far as the algebra is concerned. Of course,
:

there are other


clusion,

and of getting the con but the above are the simplest. When a syllogism has a particular premise, and therefore a particular conclusion, the process is somewhat different. Here we have given one

ways

of representing the premises

equation
(1)

1=0) and

one inequation

{+0j.

We

proceed as follows:

expand the inequation by introducing the third element; (2) multiply the equation by the element not appearing in it; (3) make use of the prin If a + b =f= and a = 0, then b 4= ciple, to obtain an inequation with
"

0",

only one term in the


the
"middle
15 16

literal

member;

(4) eliminate the element representing

term"

from

this inequation.

Take, for example,

1 1 in

See Chap, n, 7-4. See Chap, n, 2-8 and 3-1.

Applications of the Boole-Schroder Algebra


the third figure:
All x
is z,

195

x -z x y
-z.
0.

=
=}=

0.

Some x
xy

is y,

0.

[1-5]
[8

+ -z) = x y z + x y Since x -z = 0. x y -z =
x y
(z

Hence, x y

+ x y -z

=f=

0.

17] Since

xy z + xy -z
yz
4= 0,

4=

and
y
is

-.r

0,

therefore x y z

=f=

0.

Hence

[8-22]

or

"Some

a".

An

exactly similar process gives the conclusion for every syllogism with a

particular premise.

have omitted, so far, any consideration of syllogisms with both premises universal and a particular conclusion those with "weakened" in the third and fourth figures. These conclusions, and A A I and E A
are
all

We

invalid as general forms of reasoning.


is

They
If

involve the difficulty

which

now

familiar:

a universal does not give a particular without an

added assumption that some class has members.


of such syllogisms the

we add

to the premises

assumption that the

class

denoted by the middle


Take, for

term

is

a class with members, this makes the conclusion valid.

example,

AA

I in the third figure:


is y,

All x

x-y =
x -z

0, 0.

and x has members,

=|=

0.

All x

is z,

Since x

=k Q,

xy + x

-y

4= 0,

and

since x -y

0,

x y 4

0.

Hence x y
Since x -z

+ x y -z 4= 0, x y -z
(2),

0.

(1)
0.

(2)
z
=f=

By

(1)

and

xy

0,

and hence y

0,

or

"Some

is

z".

Syllogisms of this form are generally considered valid because of a tacit assumption that we are dealing with things which exist. In symbolic
reasoning, or

any other which

is

rigorous,

any such assumption must be


is

made
If

explicit.

An alternative treatment
we take

of the syllogism

due to Mrs. Ladd-Franklin. 17

the two premises of any syllogism and the contradictory of its three conclusion, we have what may be called an "inconsistent triad" that if any two of them be true, the third must be false. propositions such For if the two premises be true, the conclusion must be true and its con17

See

"On

the Algebra of

Logic",

in Studies in Logic by

University, ed.

by

Peirce;

also articles listed in Bibl.


in a modified form,

We

members of Johns Hopkins do not follow Mrs. Franklin s

symbolism but give her theory

due to Josiah Royce.

196

A
And
false.
if

Survey of Symbolic Logic


the contradictory of the conclusion be true,
i.

tradictory false.
if

e.,

the conclusion be

false,

and

either of the premises true, then the other

premise must be

sponds to three valid syllogisms.

As a consequence, every inconsistent triad corre Any two members of the triad give the
For example:

contradictory of the third as a conclusion.

Inconsistent Triad
1.

All

a:

is is

y
z
is

2. 3.

All y

Some x

not

z.

Valid Syllogisms
1.

Allan s?/
All

1.

M\

is

y
is is

2.

All y

is 2 is is

2.
.
.

is

3.
. .

All x

is 2.

Some x Some y

not

z
z.

3.
.

not

Some x Some x

not 2 not
y.

Omitting the cases

in

which two universal premises are supposed to

give a particular conclusion, since these really have three premises and

are not syllogisms, the inconsistent triad formed from any valid syllogism
will consist of

two universals and one

particular.

For two universals

will

give a universal conclusion, whose contradictory will be a particular; while

one premise be particular, the conclusion will be particular, and its contradictory will be the second universal. Representing universals and
if

particulars as
sistent triad,

we have done, this means that if we symbolize any incon we shall have two equations = Oj and one inequation 0}.
{

={=

And

the two universals

as a conclusion.

must give the contradictory of the This means that the contradictory of the
{

Oj

particular particular

must be

expressible as the elimination resultant of

a x + b -x

0,

because we have found

all

an equation of the form conclusions from two universals

to be thus obtainable.
will

Hence the two universals

be of the form a x

=
is

and
a b

-x
0,

any inconsistent triad The elimination respectively.


of
will

resultant of a x + b -x

= =

whose contradictory
:

be a

0.

Hence every inconsistent


ax

triad will
0,

have the form

-x

0,

a b

=f

where a and

b are

any terms whatever

positive or negative,

and x

is

any

positive term.

The
from

validity of

any syllogism m-ay be tested by expressing

its

proposi
it

tions in the form suggested, contradicting its conclusion

by changing

{=0}

to

{=t=.0} or the reverse,

and comparing the resulting triad

Applications of the Boole-Schroder Algebra

197

with the above form.


considering

And

the conclusion of any syllogism

may

be got by

how

the triad must be completed to have the required form.

Thus,

if

the two premises are

No
and

is y,
is y,

x y
-z -y

= = =

All not-z

the conclusion must be universal.

The

particular required to complete


is

the triad

is

x -z
it

4= 0.

Hence the conclusion

x -z

0,

or
is

"All

is

z".

(Incidentally

may

be remarked that this valid syllogism


Again,
if

in

no one of

the Aristotelian moods.)

the premises should be x y

and

yz =

0,

no conclusion

is

to the possible, because these two cannot belong

same inconsistent

triad.

We can, then, frame a single canon for all strictly valid syllogistic reason
ing:

The premises and the contradictory of the conclusion, expressed = 0} or 4= 0}, must form a triad such that symbolic form,
{ {

in

(1)

There are two universals

(2)

and one particular The two universals have a term in common, which
{

0}

{4= 0).
is

once positive

and once negative.


(3)

The

particular puts

4=

the product of the coefficients of the

com

mon term

in the

two

universals.

few experiments with traditional syllogisms will make this matter clear The validity of this canon depends solely upon the nature to the reader. three terms, three propositions and upon the law of

of the syllogism

elimination resultants,

"If

.r

+ b -x

0,

then a

0".

Reasoning which involves conditional propositions hypothetical argu same process, if we first ments, dilemmas, etc. may be treated by the A reduce them to syllogistic form. For example, we may translate cases in which where x is the class of is J5, then C is by "All x is All cases in which .1 in which C is Di. e., A is B, and y the class of cases
"If

D"

y",

"

is

B
"

are cases in which


All 2
is

is

D".

And we may

translate

"But

is

B"

by

",

where

z is

the hypothetical argument: is represented by the syllogism: C is


"If Z)",

the case or class of cases under discussion. Thus But ,1 is B. Therefore, .1 is B, C is D.

"All
"

cases in which
all

is

are cases in which

is

D.
is

But

the cases in question are cases in which

B.
is
Z)."

"Hence all

the cases in question are cases in which

And

in some other arguments of this type are reducible to syllogisms to Thus the symbolic treatment of the syllogism extends similar fashion.
all

198

A
also.

Survey of Symbolic Logic


is

them

But conditional reasoning


of the algebra

more

easily

and simply treated


lies in

by another interpretation

the interpretation for propositions.

The
its

chief value of the algebra, as

an instrument of reasoning,

liberating us

from the limitation to syllogisms, hypothetical arguments,


logic.

dilemmas, and the other modes of traditional


the narrowness of formal logic
still

Many who

object to

do not

realize

logical point of view) its limitations are.


etc.,

how arbitrary (from the The reasons for the syllogism,


be worth while to exemplify

are not logical but psychological.

It

may

this fact.

We

shall offer

two

illustrations designed to show,


logical possibilities

each in a

different
logic.

way, a wide range of


first

undreamt

of in formal

The

of these turns

upon the properties of a triadic relation


to think in terms of dyadic relations:

whose

significance

was

first

18 pointed out by Mr. A. B. Kempe.

It is characteristically

human

we

habitually break up a triadic relation into a pair of dyads.


is

In fact, so

ingrained

this disposition that


It

some

will

be sure to object that a triadic

relation is a pair of dyads.


all

dyadic relations are triads with a null

would be exactly as logical to maintain that member. Either statement is


simply one of point of view
psychological
logical sense

correct

enough
If

the difference

is

preference.

there should be inhabitants of

Mars whose

coincided with our own, so that any conclusion which seemed valid to us would seem valid to them, and vice versa, but whose psychology otherwise
differed

prejudice in favor of triadic relations.

from ours, these Martians might have an equally fundamental We can point out one such which
the

they might regard as


or inclusion.

elementary relation of logic


all

as

we regard

equality

In terms of this triadic relation,

their reasoning

might

be carried out with complete success. Let us symbolize by (ac/b), a-bc +

diagrammed
+ -ab -c
is

as in figure

16,

since

-ab-c = 0. This relation may be a-bc + -ab-c = is equivalent to


(ca/b) are equivalent, since
c.)

accbc(a + c).
This relation

(Note that

(ac/b)

and

a -be

symmetrical with respect to a and


(ac/b)

precisely the information which we discard in drawing a syllogistic conclusion from two universal habitually If all a is b and all b is c, we have premises.

represents

a -b

=
+ -c) +

and
(a

6 -c b -c

=
0,

Hence a -b
18

(c

+ -a)

the Relation of the Logical Theory of Classes and the Geometrical Theory of Points/ , Proc. London Math. Soc., xxi, 147-82. But the use we here make of this relation is due to Josiah Royce. For a further discussion of Kempe s triadic relation,
"On

See his paper

see below, Chap, vi, Sect. iv.

Applications of the Boole-Schroder Algebra


Or, a -6 c + a -b -c + a b -c + -a b -c
[5-72] This equation
(1)
is

199

0.

equivalent to the pair,

a -b -c + a b -c

and
(1) is

(2)

a -b

+ -a

b -c
"All

= =
a

a -c
0.
is

(b

+ -b)

a -c

0,

the syllogistic conclusion,


feel

c";

(2) is (ac/b).
all

Perhaps most
the information
that precisely

of

us would

that a syllogistic conclusion states

given by the premises: the Martians might equally well

feel

FIG. 16

what we overlook
curious
"illogical"

is

the only thing worth mentioning.

And

yet with this

be capable of understanding and of getting for themselves any conclusion which a syllogism or a hypo
prejudice, they
still

would

thetical

argument can

stateable in terms of

and many others which are only very awkwardly a is our formal logic. Our relation, a c6, or
give,
"All
6",

would

be, in their terms, (Ob/ a).


1

(Qb/a)

is

equivalent to

a -6 +

-a 6

=
"

a -b

hence (Oc/a) This would, in fact, be only a special case of a more general principle which is one of those we may suppose the Martians would ordinarily rely upon

Hence the syllogism

in

Barbara would be

(06/a)

and

(Oc/6),

".

for inference:
principle holds,

"If

(xb/a)

and

(xc/b),

then

(arc/a)".

That

this

general

is

proved as follows:
(xb/a)
(xc/b)
is

-x a -6 + x -a -x
b -c

b
c

is

+ x -b

= =
=

These two together give


-x a-b - x a _5
(e

or?

+ -c) + x -a b (c + -c) + -x b -c (a + -a) + x -b c (a + -a) + - x a -b -c + x -a b c + x -a b -c + -x a b -c + -x -a b -c + x a -b


c

0,

+ x -a -6 c

0.

200
[5-72] This equation
(1)
-.r

A
is

Survey of Symbolic Logic


equivalent to the pair, -ab c + x -a -b

a b -c + -x a -b -c + x

c c (b

= =
(2)

-x a-c

(b

+ -b) + x -a

+ -b)

-x a -c +
c

a*

-a

0.

x -a

b -c

+ -x -a b -c + x a -b

+ -x a -b c
b -c (x

= =
(1) is (xc/a), of

-a

+ -x) + a -b
c

c (x

+ -x)

-a

b -c

+ a -b

0.

which our

syllogistic conclusion is a special case;

(2) is a

similar valid conclusion,

though one which we never draw and have no

language to express.

Thus these Martians could deal with and understand our formal
by treating our dyads
as triads with one

logic

member

null.

In somewhat

similar fashion, hypothetical propositions, the relation of equality, syllo

gisms with a particular premise, dilemmas, etc., are all capable of state ment in terms of the relation (ac/b). As a fact, this relation is much more
powerful than any dyad for purposes of reasoning. Anyone who will trouble to study its properties will be convinced that the only sound reason for not using it, instead of our dyads, is the psychological difficulty of

keeping in mind at once two triads with two members in

common

but

Our differently placed, and a third member which is different in the two. But the operations of the algebra are inde attention-span is too small.
pendent of such purely psychological limitations that is to say, a process too complicated for us in any other form becomes sufficiently simple to be
clear in the algebra.
logic

The algebra has a

generality and scope which

"

formal

"

cannot attain.

traditional modes of reasoning.

This illustration has indicated the possibility of entirely valid nonWe shall now exemplify the fact that by
are not so remote from familiar processes of reasoning,

modes which

any

number of non-traditional conclusions can be drawn. we make use of Poretsky s Law of Forms: 19
x
This law
is

For

this purpose,

is

equivalent to

-x +

-t

-tx
"All

=
a

-t-0
is
6"

evident enough: if x = 0, then for any t, t-x = t-1 = t, and = t. Let us now take the syllogistic premises, + 0, while

and

"All

6 is

c",

and

see

what

sort of results can be derived

from them by

this law.

All a
All b
19

is b,
is c,

a-b =
b-c =

0. 0.

See Chap, n, 7-15 and 7-16.

Applications of the Boole-Schroder Algebra

201

Combining

these,

a-b +

b -c

And

[3-4-41]

-(a-b + b-c)

= 0. = -(a

-b) --(b -c)

(-a + b)(-b +

c)

=
Let us make substitutions, in terms of
a+ b
a, 6,

-a -b + -a
t

+b

c.

and

c,

for the

of this formula.

= (a + b) (-a -b + -a c + b c) = abc+-abc + bc = be
is

+ -a -b (a-b + b -c)

What

is

either a or 6

identical with that


"All

which
6

is

both

and
c".

c.

This

is

non-syllogistic conclusion from

is

and

all

b is

Other such

conclusions
a

may be got by similar substitutions in the formula. + c = (a + c) (-a -b + -a c + b c) + -a -c (a -b + b -c) = a b c + -a -b c + -a c + b c + -a b -c = a b c + -a (b + c)


is

What
is

is

either a or c

identical with that

which

is a, b,

and

c, all

three, or

not a and either


-5
c

b or c.

= -b c (-a -b + -a c + b c) + (b + -c) (a = -a -b c + b -c + a -b -c = -a-b c +


but not
c is identical
c.

-b + b -c)
(a

b)

-c

That which
or
is

is b

with what

is c

but neither a nor b

either a or b but not


"All

The number
"All

of such conclusions to be got


is

from
of

the premises,

is

6"

and

b is

c",

limited only
c,

by the
this

number

functions which can be formed with


stitutions in terms of these
is,

a, b,

and

and the limitation to sub

of course, arbitrary.

By

method, the
is

number

of conclusions

which can be drawn from given premises

entirely

unlimited.

In concluding this discussion of the application of the algebra to the more involved logic of classes, we may give a few examples in which problems

than those usually dealt with by formal logic are solved. The examples chosen are mostly taken from other sources, and some of them, like the
first,

are fairly historic.

Example

I.

20

certain club has the following rules:

(a)

The

financial

committee
shall

shall be

chosen from

among
(c)

the general committee;

(b)

No

one

be a

member both
on the

of the general

and

library committees unless he be also


of the library

on
be

the financial committee;

Xo member

committee

shall

financial committee.

Simplify the rules.


10

See Venn, Symbolic Logic, ed.

2, p.

331.

202 Let /
g
/

A
= member = =
" "

Survey of Symbolic Logic

of financial committee.
"

general
"

library

The premises then become:


(a)
(6)
(c)

fcg,
(gl)cf,

or or

/ -g

0.
0.

-fgl =

// =

0.

We

can discover by diagramming whether there

is

redundancy here.
(c)

In

figure 17, (a) is indicated

by

vertical lines, (6)

by

horizontal,

by

oblique.

(a)

and

(c)

both predicate the non-existence of / -g

I.

To

simplify the

rules, unite (a), (6),

and

(c)

in a single equation:

-g)
[5-91]

= f -g +

-f g

+ fg

+ f- g

=f-g+(-f+f)gl=f-g
is

+ gl

0.

And
Thus the
(a
}

[5-72] this

equivalent to the pair, f -g


:

and g

0.

simplified rules will be


financial

The

committee

shall be

chosen from

among

the general

committee.
(6
)

No member

of the general

committee

shall be

on the library com

mittee.

Applications of the Boole-Schroder Algebra

203

Example
a
is

2.

21

The members
s

of a certain collection are classified

in three

ways

as

or not, as 6

or not,

and as

c s

or not.

It is
c s

then found that the class 6


c s

made up
is

precisely of the a s

which are not

and the

which are not a

s.

How

the class c constituted?


b b

Given:

a -c + -a
b (c

c.

To
6 c

solve for
b -c.
c.

c.

22

+ -c) = Hence, b c + b -c

a -c + -a
c

Hence [7-27] a -6 + -a b c Or [2-2] c = a-b + -ab.

c a -b + -a

b.

The

c s

comprise the a

which are not

6 s

and the

6 s

which are not a


b

s.

Another solution of
+ -ac to the form
[7
1]
{

this

problem would be given by reducing

a -c

0}

and using the diagram.


c is

a -c + -a

equivalent to

6 --(a -c +

-a

c)

+ -b
c

(a -c

+ -a

c)

And

[6 4]

-(a -c + -a

c)

+ -a

-c.
c

Hence, a

b c

+ -a

6 -c

+ a -6 -c + -a -6
c

0.

We

observe here (figure 18) not only that

a -b + -a

b,

but that the

FIG. 18

relation of a,

6,

and

c,

stated
c.

by the premise

is

totally symmetrical, so that

we have
21

also a

b -c + -6
of

Adapted from one


i

Venn

s, first

printed in an article on

"Boole

System

of

Logic",

Mind,
22

(1876), p. 487.
intelligible
if

This proof will be

the reader understands the solution formula referred

to.

204

A
3.
23

Survey of Symbolic Logic

Example
If

x that
is

is

not a

is

the same as

6,

and a that

is

not x

is

the same as

c,

what

x in terms of
b

a, 6, c

and

c?
-.r.

Given:

= -ax and

To

solve for

.r.

[7-1] 6

=
6

-a

.i

is

equivalent to

-(-a

.T)

+ -a -b x

(a

-.r) 6

+ -a -6

a:

=
And
c

a b + b -x + -a -6

.r

(1)

=
c

a -x

is

equivalent to

-(a -x)

+ a -c -x

(-a +

.T)

+ a-c -x

= -a c
Combining
(1)

x + a -c -x

(2)

and

(2),

a b + -a c + (-a -b + c) x + (b + a -c) -x

(3)

Hence

[5

72]

(-a -b + c) x + (b + a -c) -#

(4)

[7-221] This gives the equation of condition,

(-a -b + c)(6 + a -c)


[7-2]

6 c

(5)

The

solution of (4)
(b

is

+ a -c) c # c -(-a -6 +

c)

And by

(5),
c)

-(-a -6 +

=
x

-(-a -6 +

c)

+ 6

(a

b)

-c + b c

=
Hence
[2-2]

a -c + b

(c

+ -c)

+ a -c

b + a -c.

FIG. 19
2S

See Lambert, Logische Abhandlungen,

i,

14.

Applications of the Boole-Schroder Algebra

205

This solution

is

verified

combines

all

the data.

by the diagram (figure 19) of equation Lambert gives the solution as


x
(a

(3),

which

b) -c

This also

is

verified

by the diagram.

Example

4.

24

What

is

the precise point at issue between two disputants, one of

whom,

A, asserts that space should be defined as three-way spread having points


as elements, while the other, E, insists that space should be defined as

three-way spread, and admits that space has points as elements.

Let

asserts: s
s

=
p

p.

= space, t = three-way spread, p = having points as elements. B states: s = and s cp.


s
t

is

equivalent to

s--(tp)+-stp =
s

s -t

-p + -s

(1)
(2)

cp

is

equivalent to

-p =
t

And
(2)

is

equivalent to s -t + -s

=
=

(3)

and

(3)

together are equivalent to


s -t

-p + -s

(4)

(1)

represents

.4 s

assertion,

and

(4)

represents

s.

The

difference between

FIG. 20

the two

is

that between -s tp

=
t

and -*
-s
t

=
t

0.

(See figure 20.)

-s
24

p + -s

-p

Quoted from Jevons by Mrs. Ladd-Franklin,

loc, tit., p. 52.

206

A
difference
is,

Survey of Symbolic Logic

The

then, that

would be easy to misinterpret

way

spread not having points nificantly assert this, for he has denied the existence of any space not having
".

-st-p = 0, while A does not. It is t-pcs, "Threethis issue, -st-p = But B cannot sig as elements, is space
asserts
s

points as elements.

Both assert

p.

The

real difference is this:

definitely asserts that all

space, while

has

left

three-way spread has points as elements and is open the possibility that there should be three-way

spread not having points as elements which should not be space.

Example

5.

Amongst the
enough to sink a
unless
it is

objects in a small
useful.

boy

pocket are some bits of metal

which he regards as
either

But

all

the bits of metal which are not heavy

fishline are bent.

And

he considers no bent object useful


is

heavy enough

to sink a fishline or
fishline,

not metal.

And

the

only objects heavy enough to sink a


are bits of metal that are bent.

which he regards as

useful,

Specifically

what has he

in his pocket

which

he regards as useful?

Let x

= y = z = w =

bits of metal,

objects he regards as useful,

things heavy enough to sink a fishline,

bent objects.

Symbolizing the propositions in the order stated, we have

xy *
x -z c w,
or

x -z -w
or

y w c (z + -.i ), or zy ex ID,

x y -z w = -x y z + y z -w =
z

Expanding the inequation with reference to


x y
z

and w,

w+x

-w + x y

-z

w+

x y -z

-w

=H

Combining the equations,


x -z -w (y + -y) + x y -z w + -x y
or
z

(w + -w) + y
z

-w

(x

+ -x)
z

=
-w =

x y -z

-w + x -y

-z

-w + x y

-z

w + -x y

w + -x y z-w + x y

terms of the inequation appear also in this equation, with the exception of x y z w. Hence, by 8- 17, x y z w 4= 0. The small boy has
All the

Applications of the Boole-Schroder Algebra

207

some bent
useful.

bits of

metal heavy enough to sink a

fishline,

which he considers

This appears in the diagram (figure 21) by the fact that while

FIG. 21

some subdivision
It

of x y

must be

not-null, all of these but x y z

is null.

appears also that anything else he may have which he considers useful may or may not be bent but is not metal.

Example

6.

25

The annelida
and
either

consist of all invertebrate animals having red blood in a

double system of circulating vessels.

And

all

annelida are soft-bodied,

naked or enclosed

in a tube.

Suppose we wish to obtain the


in tubes are placed (by virtue

relation in

which soft-bodied animals enclosed

with respect to the possession of red blood, of an external covering, and of a vertebral column.
of the premises)

Let a

= = s n = = i = r =
/

annelida,

soft-bodied animals,

naked,
enclosed in a tube,
invertebrate,

having red blood,

etc.
t

and a cs (n + t), with the implied condition, n eliminate a and find an expression for s t.
Given: a
i

0.

To

25

See Boole, Laws of Thought, pp. 144-46.

208
a

A
=
i r is

Survey of Symbolic Logic

equivalent to
-(i r) a + -a
i r

a -i + a -r + -a

i r

(1)

cs (n+t)
n+
s
t)

is

-0

= 0. equivalent to a --(s n + s t) = (-s + -n) (-s + -0 -(s ii) -(s t)


-

-s + -n

-t.

Hence,

a-s + a-n-t =
(1)

(2)

Combining

and

(2)

and n

0,

a -i + a -r + -a

i r

+ a -s + a -n -t + n

(3)

Eliminating
(-i

a,

by 7

4,

+ -r + -s +

-7i -2

+ n

t) (i

+n

t)

= n
26

i r

-s +

i r

-n

-t

The

solution of this equation for s


its

is

i r

cs.

And

solution for
[5-3]

tisir-nctc-n.
s
t

Hence

ir-nc.stc.-n, or

= ir-n+u--n, where u

is

un

determined.

The

soft-bodied animals enclosed in a tube consist of the invertebrates

FIG. 22
26

See Chap, n, Sect,

iv, "Symmetrical

and Unsymmetrical Constituents

of

an Equa

tion".

Applications of the Boole-Schroder Algebra

209

which have red blood

in a double system of circulating vessels and a body with an undetermined additional class (which may be covering, together This solution may be null) of other animals which have a body covering.

by the diagram of equation (3) the square formed by the two crossed
verified

(figure 22).

In this diagram,

is

rectangles.

The lower

half of this

inner square exhibits the solution.


i r

Note that the


is

qualification, -n, in

-n c s

t,

is

necessary.
i

In the top row

a single undeleted area repre


s
t.

senting a portion of

r (n)

which

is

not contained in

Example

7.

27

"All

Demonstrate that from the premises a is either b or c is no conclusion can be drawn which involves only two
"All
a",

c",

and

of the

classes, a, 6,

and
(b

c.

Given: a c

c)

and

a.

To prove
is

that the elimination of any one element gives a result which

either indeterminate or contained in one or other of the premises.

a c

(b c

c) is
is

equivalent to a -6 -c

0.

And

= 0. Combining these, a -b -c + -a c = 0. = 0. Eliminating a [7-4], (-b -c) c = 0, which is the identity, = 0. Eliminating c, (a -b) -a = 0, or c + a -c) -a c = -a c = 0, which is the second Eliminating b, (-a
ca
equivalent to -a c
premise.

Example

8.

A
green,

set of balls are all of

them spotted with one

or

more

of the colors, red,

and

blue,

and are numbered.


All the

And

all

the balls spotted with red are


balls,

also spotted with blue.

odd-numbered blue

and

all

the even

numbered balls which are not both red and green, are on the table.
scribe the balls not

De

on the table.

Let

e r

g
t

Given:
27

(1)

-r -b -g

= = = = = =

even-numbered, -e
spotted with red,

= odd-numbered,

spotted with blue, spotted with green,


balls
0.
Logic, p. 123.

on the

table.

See

De Morgan, Formal

15

210
r

A
(2)
(3)

Survey of Symbolic Logic

-6

0.

[-eb + e-(rg)]ct,
x,

or

(-e b + e -r + e -g) -t

0.

To

find

an expression,

such that -t c x, or -t x

-t.

Such an expression
-t

should be as brief as possible.


respect to
e, r,

Consequently we must develop


all

with

b,

and

g,

and eliminate
-t,

null terms.

(An alternative
is

method would be
-t

to solve for
(r

but the procedure suggested

briefer.)

-t(e + -e)

+ -r) (b + -6) (g + -g)

= -t(erbg+erb-g+er-bg + e-rbg + -erbg + er-b-g


+ +
e -r b
e

-g + -e

r b
r

-g +

-r -b g + -e -r b g + -e

-b g
(4)

-r -b -g + -e

-b -g + -e -r b -g + -e -r -6 g + -e -r -b -g)

From

(1), (2),

and

(3),

-t (-e b

-r +

-g +

-b + -r -b -g)
(5),

(5)

Eliminating from
-t

(4)

terms involved in
g),

-t (e r b g

+ -e -r -b

or

-t

(e r b

g + -e -r -b g)
all

All the balls not

on the table are even-numbered and spotted with colors or odd-numbered and spotted with green only.

three

g...

Applications of the Boole-Schroder Algebra

211

In the diagram (figure 23), equation (1)


(2)

is

indicated by vertical lines,

by oblique,
9. 28

(3)

by

horizontal.

Example

Suppose that an analysis of the properties of a particular class of sub stances has led to the following general conclusions:
1st.

That wherever the properties a and


c,

b are

combined, either the

property
2d.

or the property d,

is

That wherever the properties

present also; but they are not jointly present. b and c are combined, the properties

a and d are either both present with them, or both absent.


3d.
ties c
c

and d are both absent

That wherever the properties a and b are both absent, the proper also; and vice versa, where the properties

and d are both absent, a and b are both absent also. Let it then be required from the above to determine what may be con cluded in any particular instance from the presence of the property a with
respect to the presence or absence of the properties b and regard to the property d.
c,

paying no

Given:

(1) (2) (3)

a b c

(c

-d + -c

d).

bcc(ad + -a-d). -a -b = -c -d.


a.

To

eliminate d and solve for


(1) is (2) is

equivalent to a b--(c -d + -c d) equivalent to b c--(a d + -a -d)

= =

0.
0.

But [6-4] -(c-d + -cd) =cd + -c-d, and -(a d + -a -d) = -a d + a -d. Hence we have, a b (c d + -c -d) =abcd+ab-c-d = Q and b c (-a d+ a -d) =-abcd+abc-d = ()
(3) is

(4)
(5)

equivalent to
(c

-a -b

+ d) +

(a

b)

-c -d

=
Combining
function of
(4), d, (5),

-a -b
(6),

+ -a -b d + a -c -d + b -c -d
result the

(0)

and

and giving the

form of a

(-a -b c + -a -6 + a b c + -a b

c)

d
c

+ (-a -b
28

+ a -c +

b -c

+ a b -c + a b

c)

-d

See Boole, Laws of Thought, pp. 118-20.

Franklin, loc. cit., pp. 51-61, Venn, Symbolic Logic, Chap, Logik: Vol. I, Dreizehnte Vorlesung.

For furfher problems, see Mrs. Larldxm, and Schroder, Algebra dcr

212

A
Or, simplifying,
(_ a
-fc

Survey of Symbolic Logic

by 5-4 and 5-91,


+ b
c)

d + (-a -b
d,

+ a -c +

b -c

+a

b c)

-d

Hence

[7

4]

eliminating
I)

(_ a

-I)

c ) (-a

-b

+ a-c +

b -c

ab
c

c)

= -a-b

ab

Solving this equation for a [7-2], -b

cac

(-b + -c).

always present when c is present and b absent, and when ever a is present, either b is absent or c is absent. The diagram (figure 24) combines equations (4), (5), and (6).

The property a

is

FIG. 24

is

As Boole correctly claimed, the most powerful application of this algebra to problems of probability. But for this, additional laws which do not

belong to the system are, of course, required.

Hence we omit

it.

Some

thing of what the algebra will do toward the solution of such problems will be evident if the reader imagine our Example 8 as giving numerically the

proportion of balls spotted with red, with blue, and with green, and the

quaesitum to be

"If

a ball not on the table be chosen at random,

what
r

is

the probability that

it will be spotted with all three colors? that it w ill be with green?" The algebra alone, without any additional laws, spotted answers the last question. As the reader will observe from the solution,

all

the balls not on the

ta*ble

are spotted with green.

Applications of the Boole-Schroder Algebra


III.
If,
uct",

213

THE APPLICATION TO PROPOSITIONS


a, b, c, etc.,

in

our postulates,

represent propositions, and the

"prod

b,

represent the proposition which asserts a and b both, then

we

have another interpretation of the algebra. Since a+b is the negative of is false that a and b are both false", or -a -6, a + b will represent
"It

"At

least
"

one of the two, a and


6",

6, is

true".

It

has been customary to read


is

a+

b,

Either a or

or

"

Either a
"Either

is

true or b
or

true

".

But

this

is

some
in

what misleading,
algebra
is

since

...

..."

frequently

denotes,

ordinary use, a relation which

is

to be understood in intension, while this

incapable of representing relations of intension.


affirm
"Either

For instance,
"

we should hardly

parallels

meet at

finite intervals or all


.

men

are

mortal".

We

might well say that the


"

"Either

or

relation here predicated fails to hold because the


irrelevant.

But

at least one of the two,


mortal",

Parallels

two propositions are meet at finite intervals"

and

"All

men

are

is

a true proposition.

by +
and

in the algebra holds

between them.
in

The relation denoted Hence, if we render a + b by


necessary connection of a

"Either

a or

6",

we must bear

mind that no

6,

no relation of

"relevance"

or

is "logical import",

intended.
"a

The negative
false".

of a, -a, will be its contradictory, or the proposition

is

It

might be thought that -a should symbolize the


that
if

"contrary"

of a as well,

a be

"All

men

are mortal

",

then

"

Xo men

are mortal

"

But if the contrary as well as the contradictory be denoted then -a will be an ambiguous function of a, whereas the algebra by -a, 29 requires that -a be unique.
should be -a.

The
and
its

interpretation of

and

1 is

most

easily

made

clear

by considering
a,

the connection between the interpretation of the algebra for propositions


interpretation for classes.

The

prepositional sign,

may

equally

well be taken to represent the class of cases in which the proposition a is a b will then represent the class of cases in which a and b are both true,
true; -a, the class of cases in

which a

is false,

and

so on.

The

"

universe", 1,

will be the class of all cases, or all

"actual"

cases, or the universe of facts.


is

Thus a =
"a

represents

"The

cases in which a
"a

true are

all

cases",

or

is

true in point of
cases,

fact",

or simply
"a

is

true".

Similarly
or
"a

is is

the class

of

no
It

and a

will

mean

is

true in no

case",

false".

might well be asked:


die

May

not

a, b, c, etc.,

represent statements which


"Today

are sometimes true or


"The

and sometimes
ace"?

false,

such as

is

Monday"
is

shows an

May

not a symbolize the cases in which a

29

See Chap, n, 3-3.

214

A
and these be not
1

Survey of Symbolic Logic

true,

all

but only some of the cases?


true",

And

should not

be read

"a

is

always
"a

as distinguished from the less

com

prehensive statement, thus suggested can be

is true"?

The answer
But symbolic

is

that the interpretation

made and

that Boole actually

made

it

in his chapters

on

30

"Secondary

Propositions".

logicians

have come to

distinguish between assertions which are sometimes true and sometimes In the sense in which "Today is Monday" is false and propositions.

sometimes true and sometimes

false, it is called

a propositional function

and not a proposition.


first place,

There are two principal objections to interpreting


In the

the Boole-Schroder Algebra as a logic of propositional functions. the logic of propositional functions
is

this algebra,

and

in the

second place,

it is

algebra to propositions by the additional


if

much more complex than much more useful to restrict the a H= 0, then a = 1, and law
"If

=^ 1,

then a

0",

and avoid any confusion

of propositions with asser

tions

which are sometimes true and sometimes

false.

In the next chapter,

we

shall investigate the

tions but not for classes or for propositional functions.

consequences of this law, which holds for proposi We need not pre
:

sume

this

law at present

the Boole-Schroder Algebra, exactly as presented

in the last chapter, is applicable

remember that a proposition


tion
is

is

throughout to propositions. But we shall either always true or never true if a proposi
:

true at

all,

it is

always
==

true.

Hence
"a

in the interpretation of the


true"

algebra for propositions, a


indifferently
false"

means

is

or

"a

is

always
either

true"

the two are synonymous.


is

And

means

"a

is

or

"a

always
b,

false".

The
is false
true".

relation a c

since

it

is

equivalent to a -b
",

=
"If

0,

may
is

be read

"

It
is

that

is

true and b
like

is false is

or loosely,

true, then b

But acb,
and
==

a+

b,

here a relation which does not signify


"logical

"relevance"
=

or a connection of
"Christmas
is

import".

Suppose a
that

"2

+2
"If

4"

b =

holiday".

We
Yet

should hardly say


"2

4,

then Christmas
is

is

holiday".

it is false

+2

and Christmas hence a c b will


it is

not a

holiday":

in this

example a -b
called

is

true,

and
";

hold.

This relation, a c

b, is

"material

implication

a relation of extension, whereas

we most frequently
"a
6"

interpret

"implies"

as a relation of intension.

But

acb

common
true,
30

case in which a

with our usual meaning of is true but b is false does not occur.
b will

has one most important property in when a c b is true, the implies


If

a c b holds, and a

is

then
Laws

not be

false,

though

it

may

be irrelevant.

Thus

"material

of Thought, Chaps, xi-xiv.

Applications of the Boole-Schroder Algebra


implication" is

215
of ordinary

a relation which covers

more than the


"a

"implies"

logic: a

c6

holds whenever the usual


in

implies

6"

holds;

it

also holds in

some cases

which

"a

implies

6"

does not hold. 31


is

The

application of the algebra to propositions


its

so simple,

and so

resembles

application to classes, that a comparatively few illustrations

will suffice.
sitions,

We

give

some from the elementary

logic of conditional

propo

and conclude with one taken from Boole.

Example
If

1.

is

B,
is

C
B.

is

D.

(1) (2)

And A
Let x

= A

is B-,

= C

is

D.
are
:

The two premises then


(1)
(2)

xcy,
x

or [4-9] -x + y
or -x

==

1.

1,

=
1
is

0.

[5-7] Since

-x + y
"

=
C

and -x =

0,

1.

1 is

the conclusion

".

Example
(1) (2)

2.

If

is

B,
is

is

D.

But C

not D.

Let

= A
(1)

is

B- y

= C

is

D.

.re?/, or

-x + y

=
1

1.

(2)

0.

[5-7] Since

-x + y

=
"A

and y =

0,

-r
or

1.

-x

1 is

the conclusion

is 1? is

false",

"A

is

not

Example
(1)
If

3.

(3)

A isB,C isD; But either A is B


x

and
or

(2)

if

is

F,

is

H.

is Z).

Let

w = A isB;
(1)

(2) (3)
31

= ClsD; y = E = 10. icc.x, or [4-9] w x = # ycz, or 2/3 w + y - 1.


discussed

isF;

= G

is //.

"Material implication" is

more at length

in

Chap,

iv, Sect,

i,

and Chap.

v, Sect. v.

216
Since w + y

A
=
1,

Survey of Symbolic Logic


icx +

= w and y z = Hence [4 5] -w x + -w x + y z + -y z = 1 + -w Hence # (ir + -w) + z (y + -y) = x + z = 1.


and w x

?y,

yz =

1.

x + -y

1.

#+3 =

is

the conclusion
if

Either

is

Z)

or

6?

is

//".

This dilemma

may = (1) w-x

be diagrammed
0,

we put our equations


0,

in

the equivalent forms

(2)

y-z =

(3)

-10

-y

0.

In figure 25,

-z

is

struck

FIG. 25

out with horizontal

lines,

y-z with
is

vertical,
is

-H>

-# with oblique.

That

everything which remains

either x or z

evident.

Example
(1)
(2)

4.
.1 is 7?

Either Either
Either

or

is

not
F.

7).

is 7)
is Z?

(3)

or E or E

is is

not
2

/?.

Let x

= A
(1)

is /*;

= C

is 7);

= E
0.

is

^.

(2) (3)

* + -2/= 1. y + s = 1, or -y-z + -3 = 1, or -xz


(1),
a;

= =

0.

By

+ -#
(2),

(3

Hence by

+ -3) = x + x + -yz = 1
z

And by

(3),

-x -y

0.

+ -y -3 = 1. = x + -y z (x + -x) = x + x -y Hence x + x -y z = x = 1.
-?/ z

+ -x -y

z.

give the categorical conclusion "A is 5", indi cating the fact that the traditional modes of conditional syllogism are by

Thus these three premises

no means exhaustive.

Applications of the Boole- Schroder Algebra

217

Example

5.

32

Assume the premises:


a necessary being, either the property of gravitation necessarily present, or it is necessarily absent.
1.

If

matter

is

is

2.

If

gravitation

is

necessarily absent,

and the world


exist.

is

not subject to

any presiding
3.

intelligence,

motion does not

If gravitation is necessarily present, If


If

vacuum

is

necessary.

4.
5.

vacuum
is

matter

not a necessary being. a necessary being, the world is not subject to a presiding
is is

necessary, matter

intelligence.

Let x

= Matter is a necessary being. = Gravitation is necessarily present. y z = The world is not subject to a presiding w = Motion exists. = Gravitation is necessarily absent.
t

intelligence.

A vacuum
are
f),
:

is

necessary.

The premises then


1

(1)
(2)

xc(y +
t

or

x-y-t =
0.

0.

(3)

(4)
(5)

z c -w, or tzw = y cv, or y -v = 0. v c -x, or v x = 0. x c z, or x -z = 0.

And

since gravitation cannot be both present


(6)

and absent,

0.
:

Combining these equations

x-y
From
between

-t

+ tz

w+
it

ij

-v + v x + x -z + y
first,

=
any

(7)

these premises, let

be required,
being",

to discover

collection

x, "Matter is

a necessary
it

and

y, "Gravitation is necessarily

present".

For

this purpose,

is

sufficient to discover

whether any one


since these are

of the four, x y

0,

x -y

0,

-x y

0,

or

-.T

-y

0,

the relations which state any implication which holds between x, or -.r, and ?/, or -y. This can always be done by collecting the coefficients of

y,

x -y, -x

y,

and -x
(7),

such as equation
32

comprehensive expression of the data, and finding which of them, if any, reduce to 1. But
-y, in the

See Boole, Laws of Thought, Chap. xiv.


s

The premises assumed

are supposed to be

borrowed from Clarke

metaphysics.

218

Survey of Symbolic Logic


is

sometimes, as in the present case, this lengthy procedure

not necessary,

because the inspection of the equation representing the data readily reveals

such a relation.

From
Hence
If
if

(7), [5-72]

vx + -vy =

0.
(v

[1-5] v

x y + -v x y

+ -v) x y

x y
is

0,

or x c -y, y c -x.

matter

is

a necessary being, then gravitation


is

not necessarily present;

gravitation necessarily present, matter is not a necessary being. Next, let any connection between x and w be required. Here no such
is

relation

easily to be discovered

by

inspection.

Remembering that

if

0,

then a b

and a -b
(-y -t +
t

=
;

From

(7),

+ y -v +
z +

+ -z + y

t)

wx
-w x
(8)

y -v + y t) w -x + (-y -t + y -v + v + -z + y + (y-v + yt) -iv -x = +


(t

t)

Here the

coefficient of

wx
y+

reduces to
v,

1,

for [5-85],
t

-i)

and
-y
-t

+ -z

+ -z

and hence the

coefficient
)

is

But

[5-90]

(-y -t + y +
0, or

+ y + t + v + -z + y t. + -z + y t = l+v + -z + yt

1.

Hence w x =

c -x, x c -iv.

7-

-IV

-y-

--{

FIG. 26

Applications of the Boole-Schroder Algebra

219

None
being;

of the other coefficients in (8) reduces to

1.

Hence the conclusion


is

which connects x and


if

is:

"If

motion

exists,

matter

not a necessary

matter

is

a necessary being, motion does not

exist".

Further conclusions, relating other terms, might be derived from the same premises. All such conclusions are readily discoverable in the dia

gram

of equation (7).

In fact, the diagram

is

more convenient

for such

problems than the transformation of equations in the algebra. Another method for discovering the implications involved in given data is to state the data entirely in terms of the relation c and, remembering
,

that
-b

"If

ac6 and bcc,


to seek directly

then

ace",

as well as

"acb

is

equivalent to

c-a",

sitions which are in question.

any connection thus revealed between the propo Although by this method it is possible to
exists, the

overlook a connection which

danger

is

relatively small.

IV.

THE APPLICATION TO RELATIONS


is

The

application of the algebra to relations


is

relatively unimportant,

because the logic of relations

immensely more complex than the Boole-

Schroder Algebra, and requires more extensive treatment in order to be of We shall, consequently, confine our discussion simply to the service.
explanation of this interpretation of the algebra.

A relation,
etc.,

taken in extension,

is

the class of

all

couples, triads, or tetrads,

which
of"

have the property of being so related.


is

That

is,

the relation
is

"father

the class of

all

those couples,

(x\y)>

such that x

father

of y: the dyadic relation

the relation

R to

which have the


(x; y)

R is (.r; R y. The extension of a relation is the class of things y, We must distinguish between the class of couples relation.
the class of all couples
y) such that x has

and the

class of couples (y- x), since not all relations are


differs

symmetrical

and

.r

R y commonly

from y

R x.

Since the properties of relations,

so far as the laws of this algebra apply to them, are the

same whether they

are dyadic, triadic, or tetradic, etc.,


sufficient.

the discussion of dyadic relations will be

The
(x; y)

"

product

",

x S, or

S, will represent the class of all those couples


",

such that
all

xRy and
couples
;

the class of

R + S, will be x S y are both true. The "sum .r R y and y) such that at least one of the two,
of R, -R, will be the class of couples

x S

y, holds.

The negative
is false.

0;

y) for

which x

Ry
(t\

The
couples

null-relation, 0, will be the null-class of couples.

If

the class of

u) for

which

Ru

is

true,

is

a class with no members, and the

220

Survey of Symbolic Logic


v

class of couples (u; w) for

which

Sw

is

true

is

also a class with

then

R
t

and S have the same extension.

It is this

no members, extension which repre


t

sents.

Thus

R =

signifies that there are no two things,

and

u,

such

that

Ru

is

true

that nothing has the relation


1, is

R to

anything.

Similarly,

the universal-relation,
course).

the class of

all

couples (in the universe of dis

The
for

inclusion,

RcS,

represents the assertion that every couple (x; y)

which x R y is true is also such that x S y is true; or, to put it otherwise, that the class of couples (x; y) for which x R y is true is included in the
class of couples (u; v) for

which u S

v is true.

Perhaps the most satisfactory

reading of

RcS
$".

"

is

The presence

of the relation

R implies

the presence of

the relation
signifies

R =

S, being equivalent to the pair,

RcS

and S c R,

that

and S have the same extension

that the class of couples

(x; y) for

which x
v is true.

Ry

is

true

is

identically the class of couples (u; v) for

which u S
It is

obvious that

all

the postulates, and hence

all

the propositions, of

the Boole-Schroder Algebra hold for relations, so interpreted.


1-1 If

and S are
true,

relations (that

is, if

there

is

a class of couples (x; y)

such that x

Ry is

and a

class of couples (u; v)

such that u S

-K

is

true),

then

R xS is a relation (that is, z) w R z and w S z are both true). If R and S be such that there is no couple (w; z) for which w R z and w S z both hold, then R x $ is the null-relation,
there
is

a class of couples (w;

such that

i.

e.,

the null-class of couples.

1-2

The

class of couples (x; y) for

which x

Ry

and x

Ry

both hold

is

simply the class of couples for which x Ry holds. 1 3 The class of couples denoted by R x S is the same as that denoted

by

SxR
1-4

namely, the class of couples


m

(x

y)

such that

xRy
y,

and

xSy
all

are both true.

The

class of couples (x

y) for

which

xRy,

x S

and x T y

hold
i.

is

identically the same in whatever order the relations be combined

e.,

Rx(SxT) = (RxS) xT. 1-5 R xO = e., the product


i.

of the class of couples for

which x

Ry

holds and the null-class of couples


1

is

the null-class of couples.


class of couples

for

For every relation, R, there is a relation -R, the which x Ry is false, and -R is such that:
1-61
If

the relation
is

Rx-S
Sy
is

is

null (that

is,

if

there

is

no couple such
is,

that x

Ry

true and x

is false), is

then

R xS = R

(that

the class of

couples for which x

Ry

true
;

identically the class of couples for which

Ry

and x S y are both true)

and

Applications of the Boole-Schroder Algebra

221

1-62

If

couples for couples for

R xS - R and Rx-S = R, then R = e., if the class of which x R y and x S y are both true is identically the class of which x R y is true, and if also the class of couples for which
i.

x x

R y is true and x S y is false is identically the class of couples for which R y is true, then the class of couples for which x R y is true is null. 1-71= -0 e., the universal class of couples is the negative of the
i.

null-class of couples, within the universe of discourse of couples.

1-8

R+S =

-(-Rx-S)-i.

e.,

the class of couples


y, is

(x\y)

such that

at least one of the two, x


of couples for

Ry

and x S

true

is

the negative of the class

which x

Ry

and x S y are both


i.

false.

1-9
for for

RxS =
R Ry

which x

Ris equivalent to R c S y and x S y are both true is


is

e., if

the class of couples (x y)


;

identical with the class of couples

which x
if

true, then the presence of

implies the presence of

and
x

the presence of

R
is

implies the presence of S, then the class of couples


true
is
33

(x; y) for

which x

Ry

identical with the class of couples for

which

Ry
33

and x S y are both


For a further discussion

true.

of the logic of relations, see

Chap,

iv, Sect. v.

CHAPTER
We

IV

SYSTEMS BASED ON MATERIAL IMPLICATION


are concerned, in the present chapter, with the
"

calculus of propo
its

sitions"

or calculus of

"material

implication",

and with

extension to

pro positional functions.


procedure, and
it is

We

shall

discover here two distinct

modes
by

of

part of our purpose to set these

two methods

side

side.

procedure takes the Boole-Schroder Algebra as its foundation, the elements of this system as propositions, and adds to it a interprets
first

The

postulate which holds for propositions but not for logical classes.
result
is

The

what has been


then x

called the "Two-Valued

Algebra",

because the

additional postulate results in the law:

and

if

=t=

0,

1.

For any x, if x =|= 1, then x = 0, This Two- Valued Algebra is one form of the
where x n
is

calculus of propositions.

The extension
<px

propositions of the
{

form
,

of the Two-Valued Algebra to an individual member of a class

a* 3 etc., gives the calculus of prepositional functions. composed of x II and 2 functions have a special significance in this system, and the relation
,
a>,

of of

"formal implication", it,

Hx

(<px

c\j/x), is

particularly important.

In terms

the logical properties of relations

including the properties treated

in the last

chapter but going beyond them

can be established.

This

is

the type of procedure used by Peirce and Schroder.

The second method


is

that of Principia Mathematica

begins with the

calculus of propositions, or calculus of material implication, in a form

which

simpler and otherwise superior to the Two-Valued Algebra, then pro ceeds from this to the calculus of propositional functions and formal impli
cation,

and upon
of

this last bases not only the


classes".

treatment of relations but also

the

"calculus

It is especially

important for the comprehension of the whole subject

of symbolic logic that the agreement in results of these

and the difference of method, two procedures, should be understood. Too often they appear to

the student simply unrelated.


I.

THE TWO-VALUED ALGEBRA


.
. .

a, b, p, q, etc., represent propositions, and a x b or a b represent the joint assertion of a and b, then the assumptions of the

If

the elements

summary

See Schroder, Algebra der Logik: n, especially Fiinfzehnte Vorlesung. is contained in Schroder s Abriss (ed. tiller), Teil n.

An

excellent

222

Systems Based on Material Implication

223

and p = 1, is true". Since and 1 are unique, it follows that any two propositions, p and q, such that and q = 0, p = or such that p = 1 and q = 1, are also such that = q. p = q, in the p
sent
"p

Boole-Schroder Algebra will all be found to hold for propositions, as was 2 As was there made clear, p = explained in the last chapter. will repre
is
false",
"p

algebra, represents a relation of extension or

"truth

value",

not an equiva

lence of content or meaning.

-p symbolizes the contradictory or denial of p. The meaning of p + q is readily determined from

its definition,

P+
p+q
"At

-(-P -q)
q
is
false",

is

the denial of

"p

is

false

and
q,

or

it

is

the proposition

least

one of the two, p and


is

is

true",

p+q

may

be read loosely,

"Either

p
is

true or q

is

true".

The

possibility that both

p and

q should

be true

is

not excluded.
equivalent to p q

pc

= p and

to

p -q =

0.

;;

material implication.

We

shall consider its properties

c q is the relation of with care later in


q

the section.
"It

is

false

For the present, we may note simply that p c that p is true and q false". It may be read
(materially) implies
q".

means exactly
p
is

"If

true, q

is

true",

or

"p

With the

interpretations here given,

all

the postulates of the Boole-

Schroder Algebra are true for propositions. Hence all the theorems will also be true for propositions. But there is an additional law which holds
for propositions:

p =
"The

(P

=
p

i)
is

proposition, p,
this that

is

equivalent to

true

".

It follows

immediately

from

-P = (-P
"-p

=
".

1)

(P

0)

is

equivalent to

is

false

It also follows that

-p = -(p =

1),

and hence
-(P

1)
is

(P

0),

and

-(p

0)

=
is

(p

1)
is

=
"p

1 is

false

equivalent to p

=
==

0",

and

"

p =
is

false

equivalent

to

1".

Thus the
is

calculus of propositions

a two-valued algebra:

every proposition

either

or

1,

either true or false.

We

may, then,

proceed as follows:
2

All the propositions of the Boole-Schroder Algebra

However, many

and

solutions, are of little or

of the theorems, especially those concerning functions, eliminations, no importance in the calculus of propositions.

224

Survey of Symbolic Logic

which were given in Chapter II may be regarded as already established in the Two- Valued Algebra. We may, then, simply add another division of propositions the additional postulate of the Two-Valued Algebra and the
additional theorems which result

from

it.

Since the last division of the


shall

orems

in

Chapter
.

II

was numbered

8-,

we

number the theorems

of

this section 9

The
9-01

additional postulate

is:

For every proposition


for

p,

(p

1).
:

And
9-02

convenience we add the convention of notation

-(p

q) is

equivalent to p
of 9-01,

=|=

q>

As a consequence
-(p

we

shall

have such expressions as -(p

=
=|=

1)
1

and
and

=
0.

0).

9-02 enables us to use the more familiar notation, p

p*

It follows immediately from 9-01 that the Two-Valued Algebra cannot be viewed as a wholly abstract mathematical system. For whatever p and 1 may be, p = 1 is a proposition. Hence the postulate asserts that

any element,
interpretation

p, in the

system,

is

a proposition.

But even a

necessary

be abstracted from in one important sense no step in need be allowed to depend upon this interpretation. This is the proof

may

procedure we shall follow, though it is not the usual one. It will appear shortly that the validity of the interpretations can be demonstrated within
the system
itself.

In presenting the consequences of 9-01 and 9-02,

we

shall indicate

previous propositions by which any step in proof is taken, by giving the number of the proposition in square brackets. Theorems of Chapter II may, of course, be used exactly as if they were repeated in this chapter.

9-1

-p = (p =
[9-01]

0).

-p = (-p =

1).

And

[3-2]

-p =

1 is

equivalent to p

0.

9-12

-p = (p
[9-01]

+
=

1). 1).

p = (p =
(p

Hence

[3-2]

-p = -(p =

1)

(p

1).

9-13

(p*

1)

0).

9-14

(p

[9-M2] + 0) = (P =
[9-13, 3-2]

1).

9-13 and 9-14 together express the fact that the algebra

is

two-valued.

Every proposition

is

either true or false.

Systems Based on Material Implication

225
written the
.",

Up
logical
.
. .

to this point
relations
. .
.",

that
.

is,
,

"If

throughout Chapter II "Either then


.

we have
. .

.",

or

"Both

and

etc.,

not in the symbols of the system but just as they

in arithmetic or geometry or any other mathematical have had no right to do otherwise. That "... c ..." system. + is by inter and then is by interpretation or does not warrant us in identifying the "Either pretation

would be written

We

"If

".

."

",

.",

theorem

"If

a c

6,

then

-be

-a"

with
. . .

"

(a
,

c b) c
.

(-/;
.

c -a)

".

We

have

."in theorems with then had no more reason to identify c ..." than a geometrician would have to identify the period at "... end of a theorem with a geometrical point. The framework of logical the
"If

relations in terms of

which theorems are stated must be distinguished from

the content of the system, even

to interchange the joint But we can now prove with pcq, etc. We can then of p and q wath p xq, assertion p, demonstrate that if p and q are members of the class A then p c q is a And we can is equivalent to p cq. member of K, and that p, then
r

when that content that w e have a right


"If
<y",

is logic.

"If

</",

demonstrate that this

is

true not merely as a matter of interpretation but

by the necessary laws


"Both

of the

system
in
.
.

itself.

We
. .

can thus prove that writing


"Either
.

the logical relations involved

the theorems
,

or

.,"

...

and ..

.,"

"If

then

."in

terms of

+,

x,

c,

etc., is

a valid procedure.
in

which these things are proved are never needed here and their after, except in the sense of validating this interchange of symbols we need not give them any section number. Consequently interpretation.

The theorems

(1)

If

is

an element

in

A, p =
is

and

[9-01] If
If
is

is

an element in

= A, p = 1
/;

are elements in
is

an element

in
[9

A.
1]
/;

[1-0]

p is an element in A, -p an element in K.
q,

an element

in

A, and hence

(2)

The two, p and

are together equivalent to p x


[5-73]

q,

or p

q.

= (pq = 1). [9-01] pq = 1 and q = 1, and pair, p


(3)
If

pq =
is

is

equivalent
pair,

to

the

hence [9-01] to the

p and
in

q.

p and
[4-9]
if

q are

elements in A, then p c q

an element

A.

But
(4)

p cq is equivalent to p -q and q are elements in A, [1 6, p


"

0,
1

and hence
1]

[9-1] to -(p-q).

-(p -q)

is

an element

in

A.

-p
is

is

equivalent to

p
4=

is

false".

[9-12]
false",

-p =

(p

1),

and

[8-01]
"p

4=

is

equivalent to

=
"p

and hence

[9-01] to

is

false".

16

226

A
(5)

Survey of Symbolic Logic

pc

is

equivalent to

"If

p,

then
1,

g".

[5-64]
"If

p eg
g".

gives

"If

p =

then q

=
it

1",

and hence

[9-01]

p,

then
p,

And = g
[2-2]

"If

then

g"

gives

peg,

for [9-01]

gives

"If

p =
g

1,

then

1",

and

(a)

Suppose as a fact p

1.

Then, by hypothesis,

1,

and

peg.
Suppose that p
={=

(6)

1.

Then

[9-14]

p =

0,

and

[5-63]
in

peg.

(6)

If

p and
[7-1]

g are elements in

K, then

p =
+

is

an element
0,

K.
[9-1]

p =

is

equivalent to

p-g -pg =

and hence

to-(p-g + -pg).
(7)

Hence

[1-6, 1-1, 3-35]


is

Q.E.D.
g".

p =

g is equivalent to

"p

equivalent to
"peg

[2-2]

p =

g is equivalent to
"p

and

cyj".

By
if g,

(5)

above,
p".

eg and
this
is

cp"

is

equivalent to
"p

"If

p,

then

g,

and

then

And

equivalent to

is

equivalent to
in

g".

(8)

If

p and

g are elements in

K, then p
g).

4= g is

an element

K.

[9-02] (p

g)

=-(P =

Hence, by
(9)

(6)

above and 1-6, Q.E.D.


"p

=|=

is

equivalent to

is

not equivalent to

g".

By
(10) p + g
is

(4)

and

(2)

above, Q.E.D.
"At

equivalent to

least

one of the two, p and


equivalent to
is

g, is

true.

[1-8]

p+q =
and
(2)

-(-p-g).
above, -(-p-g)
is false)
g, is
".

By
(p

(4)
is

is

"It

is

false

that

false

and

And

this

equivalent to

"At

least

one

of the two,

p and

true".

In consideration of the above theorems, we can henceforth write


"

".

for

"If

then

...","...

=
. .

..."

for
.",

".

is

equivalent

to

...","...

..."

for

"Either,

or

etc.,

for

we have

and the proved that not only all expressions formed from elements in relations x and + are elements in K, but also that expressions which in
volve

c, and =, and

=j=

are elements in the system of the


of
"If
.

Two-Valued
. .

Algebra.
of
"Both

The equivalence
. .

then
.

..."

with

".

.",

no longer a matter = (p = 1). Also, we can of interpretation but a consequence of 9-01, p go back over the theorems of Chapter II and, considering them as propositions then of the Two-Valued Algebra, we can replace etc.,
.

and

..."

with

".

.",

etc., is

"If

.",

Systems Based on Material Implication

227

by the symbolic equivalents.


corresponding theorem which

is

Each theorem not wholly in symbols gives a wholly in symbols. But when we consider

the Boole-Schroder Algebra, without the additional postulate, 9-01, this procedure is not valid. It is valid only where 9 -01 is one of the postulates
i.

e.,

only in the system of the Two-Valued Algebra. Henceforth we shall write all our theorems with pcq for
q for
"If
.

"

If p,

then

7",

p =
use

"p

is

equivalent to

r/",

etc.

But
".

in the proofs
.

we

shall frequently

then

..."

instead of

.
",

etc.,

because the

symbolism sometimes renders the proof obscure and makes hard reading. (That this is the case is due to the fact that the Two-Valued Algebra does
not have what we shall hereafter explain as the true
"logistic"

form.)

9-15

+
(p4=

1.

=
.9-16
q)

0.

Hence [9-13]0
(-p

4= 1.

=
If

=
q

q)

(p

=
1,

-?).

(1)

And

if

p = = 1, p
If

and p

[3-2]

= = -p

then q

4= 1

and
q.

[9-13] q

0.

0.

Hence -p =
then [9-13] p

= 0, and [3-2] -p = =h 1, (2) p 4= q = 1 = -p. Hence if p q, then q 4= 0, and [9- 14] q If -p = q and q = 1, then -p = 1, and [3-2] p = 0. (3)
and p
=j=

1.

Hence

[9-15]
If

4= q.

= 0. (4) -p = q and q 4= 1, then -p 4= 1, and [9-13] -p Hence [3-2] p = 1, and p 4= qBy (1) and (2), if p 4= q, then -p = q. And by (3) and (4), = q are equivalent. Hence p 4= q and q-p = q, then p And [3 -2] (-p = q) = (p = -q).
=)=
->

if

This theorem illustrates the meaning of the relation, =, in the calculus or 7; = If p 4= (/. then either p = 1 and q = of material implication. = 0, and if p = 0, then -p = 1. Hence and q = 1. But if p = 1, then -p
the theorem.

Let p represent
home".

"Caesar died",
died"

and

q represent

"There
"There

is is

no place no place
is
{

like like

If

"Caesar

is

not equivalent to
die"

home",

then

"Caesar

did not
is

is

equivalent to
{

"There

no place

like

home".

The equivalence

one of truth values

0} or

1}

not of content or logical significance.

9-17

p =

(p

1)

(p

4= 0)

(-p

0)

(-p

4= 1).

[9-OM3-14-16]
9-18

-p

(p

0)

(p

4=

1)

(-P

1)

(-P

0).

[9-M3-14-16]

228
9-2

A
(p

Survey of Symbolic Logic

l)(p
[2-4]

0)

0. 0.

p-p =
is

And

[9-01]
false.

p =

(p

1);

[9-1]

-p = (p =

0).

No

proposition
(p

both true and

9-21

l)(p

0)

0.

9-22

(7;

= 0. And [2-4] -pp = l) + (p = 0) == 1.


[4-8]

[9-18J

-p =

(p

1);

[9-17]

p = (p

0).

p + -p =
is

1.

Hence [9-01-1] Q. E. D.

Every proposition
9-23

either true or false.


0)
== 1.

(p*

l)

(p*

[4-8, 9-01-1]

Theorems

of the

same

sort as the above, the proofs of

which are obvious,

are the following:

9-24

(p q)

9-25

= (p = l)(g = 1) = (p 4 0) (g = == - (p = l)(g 4= 0) = -(-p + -g) (p * 0)(g 1) = (-P + -g = 0) = [(p = 0) + (g = 0) = 0] = [(P * 1) + (g * 1) = 0], etc., etc. = (p + g = 1) = (p + g 4= 0) = (p = 1) + (g = 1) (p + g) = = -(-p -g) = [(p = 0)(g = 0) = 0] (P * 0) + (g 0) = [(P * l)(g * 1) 4= 1], etc., etc.
=
(p
1)

(p g

0)

=t=

0)

=(=

These theorems

illustrate the variety of

relation can be expressed in

ways in which the same logical the Two-Valued Algebra. This is one of the
redundancy
of forms.

defects of the

system

its

In this respect, the


neater calculus of

alternative method, to be discussed later, gives a


propositions.

much
.

We

turn

now

to the properties of the relation c

We

shall include here

some theorems which do not require the additional postulate, 9-01, for the sake of bringing together the propositions which illustrate the meaning of
"material implication".

9-3

(per/)

(-p +

q)

(p -q

0).
1).

[4-9] (peg) = (p-g = 0) == (-p+g = [9-01] (-p + g = 1) = (-p + g).


"p

materially implies
"It

g"

is
is

equivalent to
true

"Either

is

false or g is

true",

and to

is false

that p

and

false".
"

Since

pc

g has

been proved to be an element in the system,


g"

It is false

that p materially implies

may

be symbolized by -(peg).

Systems Based on Material Implication

229

9-31

-(p

cry)
[3 4]

(-p + q
q)

0)

=
-q.

(p-q).

-(-p +
</)

= p

And

[9

3]

-(p c

q)

-(-p +

q)

[9
"p

.02] -(-? +

(-/;+</

is

0).

does not materially imply


true",

q"

equivalent to

"It

is

false

that either p

is

false or q is

and to

"p

is

true and q

false".

9-32

(p

0)c(pc<7).

[5-03] Ocr/.
If

Hence Q.E.D.
ry,

is false,

the famous
proposition".

then for any proposition or notorious theorem:

p materially implies
false

q.

This

is

"A

proposition

implies any

9-33

(q=-

})c(pcq).

[5-01]

pel.

Hence Q.E.I).
"A

This

is

the companion theorem:

true proposition

is

implied by any

proposition".

9-34

-(pcq)c(p =
since
if

1).

The theorem
-(p
c<y),

follows from 9-32

then [9-32] p
q,

4= 0,

by the reductio ad absurdum, and [9-14] p = 1.


is

If

there

is

any proposition,
is

which p does not materially imply, then p

true.

This

simply the inverse of 9 32.


0).

A similar consequence of 9
=
0.

33

is:

9-35

-(pcq)c(q =
If

If

-(pcq), then [9-33] q =(= 1, and [9-13] p does not materially imply q, then q is false.
-(/;

9-36

q)

(7;

-q)

-(p c
cry),
0,

q)

c (-p c

q)
1

-(/;

q)

c (-p c

-q).

[9-34-35]
[3-2] If

If

-0;

then p

p =

1, -y;

and

if

= =

and

0.
=--

0,

then -q

1.

[9-32] If
[9-331 If
If

-p = -q =

0,
1,

then -p cq and then y;c-ry.


r/,

-;; c-ry.

p does not materially imply or denial, of and the negative


ry,

then
y>

materially implies the negative,


r/,

of

p implies

and the negative

of

p implies

the negative of
"The

ry.

If

"Today

is

Monday"

does not materially imply

made of green cheese", then "Today is Monday" implies "The is not made of green cheese", and "Today is not Monday" not Monday" implies "The moon is made of green cheese", and "Today is implies "The moon -is not made of green cheese". Some of the peculiar properties of material implication are due to the

moon moon

is

230

Survey of Symbolic Logic

fact that the relations of the algebra

were originally devised to represent


exhibits properties of material

the system

of logical classes.

But 9-36

implication which have no analogy amongst the relations of classes.


is

9-36

a consequence of the additional postulate,


"is

p =

(p

1).

For
b,

classes,

represents

contained

in":

but

if

is

not contained in

it

does not

follow that a
of
b.

is

contained in not-6

may

be partly in and partly outside

9-37 -(pcq) c(qcp).


[9

36] If

"(p

q),

then -p c
q, if

-q,

and hence

[3

1]

p.
q,

Of any two propositions, p and


materially implies p.

p does not materially imply

then q

9-38

(pq) c[(pcq)(qcp)].
[9-24]

pq =

(p

l)(q

==

1).

Hence

[9-33]

Q.E.D.

If

p and

q are both true, then each materially implies the other.

9-39

(-p-q)c[(pcq)(qcp)].
[9-24]

-p-q =

(-p

l)(-q

1)

(p

0)(g

0).

Hence
If

[9-32]

Q.E.D.

p and q are both false, then each materially implies the other. For any pair of propositions, p and q, there are four possibilities:
1)

2) 3)

4)

= p = p = p =
p

1,

q q
q q

0, 0,
1,

= = = =

1:

true, q true.
false, q false. false, q true.

0:
1
:

p
p

0:

true, q false.
1

Now

in

the algebra,

cO,

1,

and

1;

but

cO

is false.

Hence

in

the four cases, above, the material implications and equivalences are as
follows
:

1)

2) 3)

pc pc

q,
q,

qcp, p = qcp, p =

q.
q.

4)

pcq, -(qcp), p 4= ~(pcq), qcp, p ^

q.
q.

This summarizes theorems 9-31-9-39.


the content or meaning of p and
"implication"

These relations hold regardless

of

q.

Thus p cq and p =

q are not the

and

"equivalence"

of ordinary logic, because, strictly


"propositions"

speak
"truth

ing,

p and

q in the algebra are not

but simply the

values"

of the propositions represented.

In other words, material impli-

Systems Based on Material Implication

231

cation and material equivalence are relations of the extension of proposi whereas the "implication" and "equivalence" of ordinary logic are relations of intension or meaning. But, as has been mentioned, the material implication, pcq, has one most important property in common with
tions,

"q

can be inferred from


does not hold.

p"

in ordinary logic;

if

is

true and q false,

the relation of material equivalence, p connects a true proposition with a false one.

And

pcq
never

q,

These theorems should make as clear as it can be made the exact meaning and character of material implication. This is
important, since

many theorems whose

significance

would otherwise be very puzzling follow

from the unusual character of

this relation.

Two more propositions,


9-4
(p q

of

some importance, may be given:


[p

r)

(q

p c

r)

(q

r)]

[q

c (p c
q)
==

r)}.

[1-3]

pq =

qp.

Hence

[3-2]

-(p

-(q p), and

[-(/; q)

r]

= (p q c r), and [-(q p) + r] = And [3 41] [-(p q) + r = + -?) + r] = [-p + (-ry + + r] = Iqc(pcr)]. Similarly, [-(q p)
But
[9
3]

[-(p q) +

r]

(q
r)]

[(-/>

per). = (p c

(q

r)}

This theorem contains Peano


[(p
"If

Principle of Exportation,
[p

q)cr]c

c(qc r)]
r";

pq

implies

r,

then p implies that q implies

and

his Principle of

Importation,
[y;
"

c
r,

(q

r)]
if

[(p q)

r]

If

p implies that
[(pq)c
r]

q implies
[(p -r)

then

p and
c

q are

both true,

"

r is true.

9-5

-q]

[(q -r)

-y,].

[9 -3]

[(p q)

r]

= =

[-(p q)
[(-</

r]

=
-y;]

[(-p + -q) +

r]

[(-p +

r)

-ry]

[-Q; -r) + -q]

= =

[-(q -r) + -p].

[9-3]

[-(p- r ) +

-ry]

-[(p-r)c-g], and
[-(q-r)+-p]
l(q-r) c-p].

p and q together imply r, then if p is true but r is false, q must be false, and if q is true but r is false, p must be false. This is a principle first stated by Aristotle, but especially important in Mrs. Ladd-Franklin s theory of
If

the syllogism.

We

have now given a


to

sufficient

number

of

theorems to characterize the

Two-Valued Algebra

illustrate

the consequences of the additional

232

A
p =
(p

Survey of Symbolic Logic


of

p c q. Any further theorems of the system will be found to follow readily from the foregoing. A convention of notation which we shall make use of hereafter is the
postulate
1),

and the properties

following:

sign
;

=, unless enclosed
,

in parentheses, takes precedence

over

any other sign a sign c unless enclosed in parentheses, takes precedence and the sign + unless enclosed in parentheses, takes over any + or x
;

precedence over a relation x

This saves

many

parentheses and brackets.

II.

THE CALCULUS OF PROPOSITIONAL FUNCTIONS.


VARIABLE

FUNCTIONS OF ONE

The

calculus of prepositional functions

is

an extension of the Two-

Valued Algebra to propositions which involve the values of variables. Fol 3 lowing 3Ir. jlussell, we may distinguish propositions from prepositional
functions as follows:
or false;

proposition
is

is

any expression which

is

either true

an expression, containing one or more variables, which becomes a proposition when each of the variables is re
a propositional function

placed by some one of

its

values.
"

There

is

one meaning of
1,

Today
2,

is Monday"
.

ambiguously Jan.

or Jan.

or

etc.
is

for which today denotes For example, when we say


",

Today
Jan.
1 is

is

Monday

implies

Tomorrow

Tuesday

we mean that

if

is

Monday, then Jan. 2 is Tuesday; if Jan. 2 is Monday, then Jan. Tuesday; if July 4 is Monday, then July 5 is Tuesday, etc. Today
variables,

and tomorrow are here


etc.,

whose values are Jan.

1,

Jan.

2,

Jan.

3,

that

is,

all

the different actual days.

variable sense,

"Today is Monday" is

today is used in this sometimes true and sometimes false, today


,

When

or more accurately,

it is

true for some values of the variable

and

false for other values.

"Today is Monday" is

here a propositional function.


Monday" for

There
today
is

is

a quite different meaning of

"Today is

which
In

not a variable but denotes just one thing


if

Jan. 22, 1916.

this sense,

"Today

is

Monday" is
its

true

it

is

always true.
its
"

It is either

simply true or simply


change.

false:

meaning and
"

truth or falsity cannot

today Today is Monday is a proposition. When Today, meaning Jan. 16, 1916, is one value of the variable today this value is substituted for the variable, then the propositional function is
this

For

meaning

of

turned into a proposition.


3 See Principles of Mathematics, Chap, vii, and Principia Mathematica, i, p. 15. Mr. Russell carries out this distinction in ways which we do not follow. But so far as is here

in question, his

view

is

the one

we

adopt.

Principia Mathematica

is

cited hereafter as

Principia.

Systems Based on Material Implication

233

We may
functions, in

use

^r,

t(x, y),

f (*, y z )
t

e tc.,
x, or

to represent prepositional

y, or x, y, and z, etc. These propositional functions must be carefully distinguished from the functions discussed in Chapter II. We there used /, F, and the Greek

which the variable terms are

x and

capitals, $, *, etc., to indicate functions;


letters.

here

Also, for

any function
<px,

of one variable,
x.

we use only Greek small we here omit any parenthesis

around the variable

\l/y,

f(x), V(x, y), etc., Chapter II are confined to representing such expressions as can be formed from elements in the class A and the relations x and + If x and y in *(*, y) are logical classes, then *(.r, y) is some
.

in

logical class,

such as x + y or a x + b -y.

Or

if

x in

/(.r)
b.

is

a proposition,

then

/(.r)

is

some proposition such as


$(x
t

functions,
<px

<px,

y},

(x, y, z),

etc.,

propositional are subject to no such restriction.


it

a x or -.r +

The

becomes a proposition when a- is replaced by one of its values, but does not necessarily become any such x proposition as a x or -.r + b. x is a citizen of y, Monday, these are y is between x and z
propositional functions.

is

typical

They

are neither true nor false, but they

become

either true or false as soon as terms denoting individual things are sub stituted for the variables x, y, etc. All the functions in this chapter are

such propositional functions, or expressions derived from them.

A
is

fundamental conception
"range

of the theory of propositional functions


".

is

that of the

of significance

The range
class,

of significance of a function

determined by the extent of the


its

or classes, of terms which are


x, in
,

values of
<px,

variables.
,

All the

terms which can be substituted for


^.r.

and make sense


mortal

constitute the range of


is

If

<px

be

is

mortal

the range of this function

the aggregate of

all

the individual terms for


"range

which x
is

is

is

either true or false.

Thus the
"universe

of

significance"
is

to propositional functions

what the

of

discourse"

to class

terms.

Two

propositional functions,

<px

and
<px,

r/^,
is

may

be such that the

class of values of

x in

<px,

or the range of

identical with the class of

values of y in

\J/y,

or the range of $y.


.r

different ranges of significance,

is

Or the two functions may have a man and .r is a poet will have the
true will differ.
.r

same
for
is

range,

though the values of


is

.r

for

which they are


is

Any

.r

which x

man

is

either true or false,


.r

also such that


.r

is

a poet

either true or false.

But some

s
.r

for

which
is

is

a poet
.r

is

either true

or false are such that


.r

x precedes

nonsense,

is

a poet

and

precedes x

have different ranges. 4

It is

important to note that the

4 According to Mr. Russell s "theory of types" (see Principia, i, pp. 41-42), the one fundamental restriction of the range of a propositional function is the principle that nothing

234
range of
function.
is

A
<px

Survey of Symbolic Logic


x,

determined, not by

but by

<p.

<px

and vy are the same

we have a prepositional function of two variables, say x is a citizen of y we must make two substitutions in order to turn it into a proposition which is either true or false. And we conceive of two aggregates or classes
If
,

the class of values of the


variable, y.

first

variable, x,

and the

class of values of the

second

These two

classes
It

may,

for a given function, be identical, or


"John

they

may

be different.
Turkey" is

depends upon the function.


"Turkey is

Jones

is

citizen of
Jones" is

either true or false;


"3

a citizen of John
is

nonsense.
3".

"5

precedes

is either true or false, as But precedes The range of x and of y in $(x, y) depends upon
5"

also

^,

not

upon x and

y.
<px

xz
is

is by x\, x 2 convenient method of representing the values of x in This is not to presume that the number of such values of x in etc.

<px

finite, or even denumerable.

Any
If

sort of tag
all
,

these values as individual would serve


of
.TI,
,T 2 ,

which would distinguish the uses which we shall make


,

3,

etc., equally well.

<px-i,

<f>x*,

<p%3,

etc., will

x, x 2 x z be propositions; and
is

etc.,
<px

are individuals, 5 then


will
n is

be a proposition.
the values of x
,

<pxz

is

a proposition about a specified individual;

<px

a proposition about
if

a certain individual
in
\I/(x,

which
.r 3 ,

not specified. 6

Similarly,

y) be
,

Xi, x*,

etc.,
,

and the values


etc., are

of y be yi, y z ys, etc.,

then

$(%*>

2/s),

t(x z y n ), t(x m y n ),

propositions.
II

We

shall

now make

new

use of the operators

and

S, giving

them

a meaning similar to, but not identical with, the meaning which they had in Chapter II. To emphasize this difference in use, the operators are here
set in a different style of type.

We
x
.

shall let

2x

<px

represent

<pXi

<px

<f>Xz

to as

many
<pxi

terms as there are distinct values of x in


x
<px

<px.

And U x

<px

will represent

z
T

x
e

<px

to as

many

terms as there are distinct

values of x in

<px.

(W

have heretofore abbreviated

ax6

to a b or a -b.
is

But where prepositional functions are involved, the form

of expressions

that presupposes the function, or a function of the same range, can be a value of the func It seems to us that there are other restrictions, not derived from this, upon the tion. range of a function. But, fortunately, it is not necessary to decide this point here. 5 which is the only "Individuals" in the sense of being distinct values of x in
<px

conception of
6

"individual"

which we require.
<px

It

may be
is

urged that

n is

not a proposition but a prepositional function.

The

most difficult, and we cannot enter upon it. But this much may be said: Whenever, and in whatever sense, statements about an unspecified individual can be certain gentle If any object to this, we shall reply n is a proposition. asserted, man is confused". Peirce has discussed this question most acutely. (See above, pp.
question
"A
<px

93-94.)

Systems Based on Material Implication


likely to

235
chapter, always

be complex.

write

"products"

Consequently we with the sign x .)

shall,

in

this

might be an infinite set of values of x in does not affect the theoretical adequacy of our definitions. For nothing here depends upon the order of m ^x n p and it is only required that the values of x which are distinct should be identifiable or The ob "tagable".
<px <px

The

fact that there

<px

jection that the values of x might not be even denumerable

is

more

serious,

but the difficulty


Since
<pxi,
<px

may
z,

<f>x

be met by a device to be mentioned shortly. 3 etc., are propositions, #TI + 2 + 3 +


,
<px <px

is

proposition

the proposition,
"For

"Either

<pxi

or

<px

or #r 3 or ...
<px

etc.".

Thus

2x 2x
of
all

<px

represents
is

some value
Similarly,
3,

of x (at least one),

is

true".

And
"For

<px

a proposition.

^.TJ

<px

^.r 3

x ...

is

the joint assertion

<ATi

and

<px

and
<px

<px

etc.
",

Thus H t

<px

represents the proposition

values of x,
true",

is

true

We may
"

translate
<px

Z x ^.r loosely by
true".

"

<px

is

sometimes

and

lation fails of literal

by inasmuch accuracy
<px

nx

loosely

is

always

This trans
<px

as the variations of x in

may

not be confined to differences of time.

The conception
3,

of a prepositional function,

<px,

and

of the class of values

of the variable in this function, thus give us the


<px

(px n ,

2x

<px,

and

U x(px.

of proposition, Since the laws of the Two-Valued Algebra

new types

hold for propositions generally, all the theorems of that system will be true when propositions such as the above are substituted for ...
a, b,

p, q,

etc.

(We must,

of course,

remember that while

a, b,

... p, q, etc., in

the

Two-yalued Algebra represent propositions, x in yx, tion but a variable whose values are individual
to follow,

etc., is not

things.

a proposi In the theorems


in

we

shall

sometimes need a symbol for propositions

which no

variables are specified.

To avoid any

possible confusion,

we

shall represent

such propositions by a capital

letter, P.)

We

may, then, assume as already


cr,

proved any theorem which can be got by replacing any proposition of the Two-Valued Algebra, by

6,
<px

...
n,

p, q, etc., in
<px,

<px

3,

2x

or

Ti x

<px.

Additional theorems, which can be proved for propositions involving values of variables, will be given below. These are to be proved by reference to

Chapter II and in Section I of this chapter. As before, theorem by which any step in proof is taken will be given in square brackets. Since the previous theorems are numbered up to 9-, the additional theorems of this section will be numbered beginning with 10earlier theorems, in

the

number

of the

One
will

additional assumption, beyond those of the

Two- Valued

Algebra,

be needed.

The

propositions which have been proved

in sufficiently

236

Survey of Symbolic Logic

general form to be used where


are in question
all

sums and products

of

more than three terms

require for their demonstration the principle of


If,

mathe

matical induction.
of this section,

then, AVC wish to use those theorems

in the proofs

we

values of x in

<f>x,

by the difficulty that the number of may and hence the number of terms in 2 x #r and H x
are confronted
<px

not be

finite.

And any

use of mathematical induction, or of theorems


for proof, will

dependent upon that principle


nection.

then be invalid

in tttis

con

Short of abandoning the proposed procedure, two alternatives are open to us: we can assume that the number of values of any variable in a propositional function is always finite; or we can assume that any

law of the algebra which holds whatever finite number of elements be involved holds for any number of elements whatever. The first of these assumptions

would obviously be

false.

But the second

is

true,

and we

shall

make

it.

This also resolves our difficulty concerning the possibility that the number of values of x in might not be even denumerable, and hence
<px

that the notation


inadequate.
in

<pxi

<px

<px

and

<pxi

x
if

tpx 2

<px

might be
of values of
"2

We
(f>x

can make the convention that


<px,

the
<px

number
+
.
.

any function,
x
<px

be not
or

finite,

<fXi+

<px

or

<px,

and

<?Xi

Hx

<px,

shall

be so dealt with that any theorem to

be proved will be demonstrated to hold for any finite number of values and this being proved, our assumption allows us to extend the of x in
<px\

theorem to any case in which the values of the variable in the function are This principle will be satisfactorily covered by the infinite in number. x shall always + and convention that 3 x 2 x 3 + 2 +
. . . . . .

<pxi

<px

<px

<pxi

<px

<px

be supposed to have a finite but undetermined

number

of terms,

and any

theorem thus proved


distinct values of

shall

be presumed independent of the number of


operative, will be

7 any variable, x, which is involved. This postulate, and the convention which makes

it

supposed to extend also to functions of any number of variables, and to


sums, products, and negatives of functions. Xo further postulates are required, but the following definitions are

needed
10-01

2p.r

10-02 10-03
7

H(px
-<p.r

= 2x = Hx
=

=
<px

<pxi+

<px

+
x

tftf 3

+ ....
x
.

Def.

=
<px

<pxi

(f>x

<pxz

..

Def.

-{<f>x}.

Def.
is

This procedure, though not invalid,

far

from

ideal, as are

many

other details of

this general

criticisms together in the last section of this chapter. But it is a fact that in spite of the many defects of the method, the results which it gives are without exception valid.

method.

We

shall gather the

main

Systems Based on Material Implication


10-031
- xn 9

237

=
-{<px

n }.

Def.
Def.
Def.

10-04
10-05

-U x
-2 x

<px

<px

= -{n x = -{S,0*r}

<px}.

The
one

merely serve to abbreviate the notation. Elementary theorems concerning propositions which involve values of
variable are as follows:
2<px

last four

10-1

-n-^.r.
<f>xi+

[5-951]

<f>x+

<px

...

-{-^.n x-^.T 2 x-^i-3 x

J.

10-12

U<px

-2-<px.

[5-95]

^ix^r
"For

10-1 states that


denial of
"For

some values
<px

of x,
is

<px

is

true"

is

equivalent to the
"For

all
is

values of x,
true"

false".

10-12 states that

all

"For some values These two represent the extension of De Morgan s Theorem to propositions which involve values of variables. They might
<px

values of x,
<px

is

equivalent to the denial of

of x,

is

false".

be otherwise stated:
false that

"It

is

true that

all
is

"

is

is

equivalent to
"

"It

is

some x
is

is

not

";

and
is

"It

true that some x

is

is

equiva

lent to

"It

false that all


n.

not

10-2

Il(px

<px

[5-99]

<pxi

x x
x

<px-~

x x x

<**3

x x

c c c

<px

and and
10-21
<px

<pxi

(px%
<^r

<px

<px%

<px!

<px

^.r 3 ,

etc., etc.

c2<f>x.

[5-991]

^c^+^roH.^*...
<pXz

and

c c

(pXi
<pxi

(px^

<px$

+ +
. . .

and

<pXs

+ pxz +

<px

etc., etc.
it is

By
of

10-2,
or

if

<px

is

true for

all

values of
true of

.r,

then

true for

any given value


10-21, If
.r,
<px

.r,

"What is

true of

all is
.r,

any given
It

one".

By

is

true for one given value of


is

then

it is

true for some value of

or

"What

true of a certain one

is

true of
is

some".

implication stated by 10-21

reversible.
<px

might be thought that the But we do not have 2^.r c ^.r n


,

because

<px

example,

let

may =
<^.r

be

<p.r

2,

and 2

<px

would not hold generally.

For

"Today (x) is Monday".

Then

2^.r will

mean

"Some

day

is Monday",
"Today

but

<px

n will

mean

"Today

(Jan. 1)
"Some

is Monday",

or will

mean

(Feb. 23)

is Monday", etc.

day

is Monday"

does

238
not imply
"

.1

Survey of Symbolic Logic

Jan.

1 is Monday",

and does not imply


is

"Feb.

23

is

Monday
n

"-

does not imply that any one given day


certain value of
.r"

Monday.

x n in

<px

means

"

in a sense

which
<px

is

not simply equivalent to

"some

value of
respect.

.r".

Xo

translation of

will give its exact significance in this

10-22

II<>.r

cS^.r.

[5-1, 10-2-21]

Whatever
10-23

is

true of
is

all is

true of some.
"Whatever
3

H<px

equivalent to

value of x, in
(<pxi

<px,

xn x
.

may
.

be,

<px

".

Ii<px

=
<pxi

<px

x
<px

<px

x
<^r

=
. .

<px

<px

=1)
to

[9-01]
set

And

[5-971]

<pxi

x
1,

==1

is

equivalent

the

=
<pXi

1,

<pX

I,

pXz

....
<px

And

[9-01]
tt<px

<px

I is

equivalent to

Hence

is

equivalent to the set

<px

lf

<px

z,

<p.T

3,

....

This proposition is not tautological. It states the equivalence of the x tpxz x with the system of separate propositions 3 x product
<pxi
. .
<f>x

<pxi,

<f>x

z,

<f>x

s,

etc.

It is
tt<px

by virtue
as
"For

of the possibility of this proposition

that the translation of

all

values of x, ^x

is

true"

is

legitimate.

In this proof

we make

use of the principle,

p =

(p

1)

the only case in

which

it is

directly required in the calculus of propositional functions.

we can pass directly from any theorem of the TwoBy Valued Algebra to a corresponding theorem of the calculus of propositional functions. If we have, for example, pc.p+q, we have also "Whatever
virtue of 10 23

value of x, in
10-23,

<px,

xn

may

be,

<px

p.r n

P".

And hence we
importance of

have, by
this:
it

Ux

[<px

<px

+ P].

We

shall

later

see the

gives us,

for every

theorem concerning

"material

implication",

a cor

responding theorem concerning

"formal implication".

Next, we give various forms of the principle by which any proposition may be imported into, or exported out of, the scope of a II or S operator.
10-3

Z#c
2

<px

P = S^r + P). + P = +
(
<.ri

<px<2

<px

+...) +
+ P) +

P
...

P) +

(^

(^3 + P)+

[5-981]

10-31

P+Z^o- - ? X (P+
Similar proof.

p.r ).

10-3

may

be read;

"

Either for some

x,

<px

is

true, or

is

true

is

equiva-

Systems Based on Material Implication


lent to

239
10-31

For some
"Either

x, either
is

<?x

is

true or

is

true

".

And
is

may

be

read:

P
is

true or, for some x,


<px

<px

is

true

equivalent to

For

some
10-32

x, either

true or
(<^.r

is

true

".

U<f>x

P = nx II (pX + P =
+

+ P).
<pXz

(<pXiX

<pX

P
P) x... [5-941]

10-33

P+tt<px

= U X (P +

tpx).

Similar proof.
<(

Either

P is true
is

or, for
<px

every
true."

x,

<px

is

true"

is

equivalent to

"For

every

x,

either

true or

is

10-34

S X (^ + P) = S X (P
[4-3]
2<px

+*>*).

+P = P+

2<px.

Hence [10-3-31] Q.E.D.

10-35

ILOi

+ P)

= n x (P+

<px).

[10-32-33]

is

Exactly similar theorems hold where the relation of the two propositions x instead of + The proofs are so simple that only the first need be
.

given.

10-36

2<px

xP = S*(#E xP). 2 px x P = ^.TI +


(

<f>x

<px

P
xP) +
. . .

=
(<^.ri

xP) +

(^-2

xP) +

(cp.i-3

[5-94]

"

<px

is

true for some x, and


are both
true".

is

true",

is

equivalent to

"For

some

.r,

<px

and

10-361

10-37
10-371

- 2 X (P x x P). n^ xP = n Pxn^r = n x (Px^r).

xS^.i-

<r>.r).

j:

<px

10-38
10-381

S z (^ xP) = S,(P x

^.r).

Hx

(<px

xP) = n x (P

<^.r).

should perhaps expect that a proposition, P, might be imported into and exported out of the scope of an operator when the relation of P But here the matter is not to the other member of the expression is c
.

We

quite so simple.

240
10-4

A
Pc?<px

Survey of Symbolic Logic

= S,(Pc

<px).

[9-3]

PC

=
2(f>x

-P+2<px

= -P

[5-981]

= (P c ^rO = S X (P c
The
But
it is

+ (P c

^.r 2 )

+ (P c

^.r s )

[9

3]

<px)

relation

c, in the above,

is,

of course, a material implication.

tedious to read continually

"p

materially implies
9",

g".

We
g".

shall,

then, translate

pcq
"P

10-4 reads:
"For

implies simply by that for some implies


"p

or
<^.r

by
is

"If

p,

then
is

.r,

true"

equivalent to

some

.r,

implies that
:

is true".
^>.r

This seems clear and obvious,

but consider the next


10-41
Stf>.rcP

IlzO.rcP).
U-<px

[9-3]

S^ccP = -S#r + P =
2

P
. .

[10-12]

= (-^i x-^.r X-^.TS x .) + P = (-^d + P) x + P) x (-^3 + P) = (^.TiCP) X(^.T CP) X(^.T CP)... = U x (<pxcP)
(-<pxz

[5-941]

[9-3]

"

<px

is

true for
P".

some x

implies

P"

is

equivalent to

"For

every

x,

<p

implies

It is
If

the

first also:

<px

easy to see that the second of these two expressions gives is sometimes true, P must always implies P, then if
<px

be true.

It is
it

not so easy to see that


"If
<px

2<px

can put

thus:
P".

is

ever true,

P gives then P is
c

tt x (<px

c P).

But we
"

true"

must mean

<px

always implies
10-42

PcU<px

=
c

Tl x (Pc<px).

[9-3]

U<px

= -P + Il^.r = (-P +
[5-941]

==

-P +

(<pxi

<pxi)

x (-P + x (P c

<px

x ^r 2 x 3 x x (-P + #c 8 ) x z)
.

.)

...

- (P c = Ut (P c
"P

<pxj

<px

2)

x (P c ^c s ) ...

[9-3]

<px)

implies that
<px".

<px

is

true for every

x"

is

equivalent to

"For

every x

implies

10-43

n^TcP = 2 = -n^ + P [9-3] n^rcp


x (<pxcP).

S-^.T +
+<X3

P
. .

[io-i]
.

<X

+-

.1*

Systems Based on Material Implication

241

= (-^1

+ P) +

(-^
(<^

+ P) +

(-<pOr

+ P) +
+

[5-981]

C P) +

C P) +

(^3 C P)

[9-3]

"

<px

is

true for every x


implies
P".

implies that

is

true"

is

equivalent to

"For

some

x, (px

At

first

sight this

theorem seems to commit the


"fallacy

"fallacy

of

division"

going one way, and the

of composition"

going the other.

It suggests the ancient

and baldness.
out".

Suppose

<px

be

"If

x
is

is

example about the separate hairs a hair of Mr. Blank s, x has fallen

And
of

let

be

"Mr.

Blank

bald".

Then

H<px

cP

will represent
is
bald".

"If

all

Mr. Blank

hairs have fallen out, then


"There

Mr. Blank

And 2 x (#rcP)
that
Ti(px
if

will represent

is

some
is

hair of
bald".

Mr. Blank

such

this hair has fallen out,

Mr. Blank
(<px

In this example,
their equivalence
is

cP is

obviously true, but 2 x

cP)

is

dubious, and

seems likewise doubtful.

The explanation
".

of the equivalence
<px

this:
"

we
is

here deal with material implication, and false that n is true but P is false)
(<px
"

P means
means,

simply

It

U<px

cP

in this

example,
is

It is false
bald";

that

all

Mr. Blank

hairs have fallen out but


"There

Mr. Blank
s

not

and 2 I (#rcP) means


is

is

some one

of
is

Mr. Blank
is

hairs

such that

This hair has fallen out but

Mr. Blank

not bald

false".

No

necessary connection

predicated between the falling out of any single


is

hair

and baldness

material implication
last four

not that type of relation.

If

we compare the

theorems,

we observe

that an operator in

the consequent of an implication is not changed by being extended in scope to include the whole relation, but an operator in the antecedent is changed

from
to
-<p

to 2,
f

from 2 to

II.

This

is

due to the fact that

pcqis

equivalent

where the sign of the antecedent changes but the consequent remains the same; and to the law -n() = 2-(), -2() = n-().
+ q

The above
tional function

principles, connecting

any proposition, P, with a preposi

used in later proofs. In fact, all the proofs can be carried out simply by the various forms of this principle and theorems 10-1-10-23. Since P, in the above, may be any propo

and

its

operator, are

much

sition,

#r n 2#r, II#c, etc., can be substituted for P in these theorems. In order that + \l/x) and x^j-) are, of course, functions of x. must be significant and $x must be significant, + \I/x) be significant, ( par Such or $x" have meaning. further requisite that "Either and it is
,
(<px
(<px
<?x <px

considerations determine the range of significance of complex functions A value of x in such a function must be at + \I/x) and like x^.r).
(<px
(<px

17

242
once a value of x
(<px

A
in
<px

Survey of Symbolic Logic


of x in fa:

and a value

x n in

<px

and

in

fa n

in

+ fan), denotes identically the same individual.

10-5

<px

+ 2 fa

= Sx

^ + #r)
is

Since addition

associative

and commutative,

"

Either for some


<px

.r,

^,

or for

some

x,

fa"

is

equivalent to

"For

some

x,

either

or

fa".

If it

be supposed that the functions,


i.

<px

and fa, may have


is

different

ranges

e.,

that the use of the same letter for the variable

not indicative
(<p.r

of the range

then S^.r + 2 fa might have meaning when 2 x

+ fa) did

not.
will

But
in

in

not have meaning.

such a case the proposition which states their equivalence We shall make the convention that x n in n
<px

and x n

wherever

<px

n +fa n ) and n xfa n ), but fa n are identical, not only in and fa are connected, as in S^.r + S#r. Where there is no
(<px
(<f>x

such presumption, it is always possible to use different letters for the But even without this convention, the above variable, as S#e + 2^y.

theorem

will

always be true when

it is

significant

i.

e., it is

never

false

and a similar remark applies to the other theorems


10-51

of this section.

<px

Ufa = n z (
is

<px

x fa)

Since x

associative

and commutative, similar

proof.
it

We
"For "For

might expect 2<pxx2fa some .r, x is ugly, and for some

= 2 x (<pxxfa)
x, x

to hold, but

does not.
to,

is beautiful", is

not equivalent

some

x,

is

ugly and x

is beautiful".

Instead of an equivalence, we

have an implication:
10-52

Sx (
[5-21

<px

x fa)
n

( (p.i-1

x fa
,

(<px

xfa n ) c

<p.r

and

(<px

xfa n )cfa n
2,(<px

Hence Hence
Similarly,
x,

[5-31]
[5-34]
U<px

2 x (^.r xfa) cS^.r, and 2x xfa) c2<px xZfa


(<px

xfa) c^fa

Hfa = U x
is

(<f>x

+ fa)
",

fails
is
.T

to hold.

"Either

for every

is

ugly, or for every x, x


is

beautiful
".

either x
ful.

ugly or

.r

is

beautiful

Some

not equivalent to, "For every x, may be ugly and others beauti

But we have:

Systems Based on Material Implication


10-53
II

243

<px

Ufa ctt x
<px

(<px

+ #r)

[5-21]

c(<px

n +\I/x n ),
(<px

Hence
Hence

[5-3]
[5

H<px

cH x
+

and $x n c \f/x), and Il^.r


<px

cU x ((px +
lemma
in

if/x)

33] II

<px

n^x c n x (

+ $x)

In the proof of the last two theorems,


of writing
(<px

we

write a

for

<px

instead

it

for
<px

<pxi,

for

<pxz,

for

<px

s,

etc.

For example,

10-52 we write

x^.r n ) c

n,

instead of writing
(<pxi

<px3,

etc., etc.

The

proofs are

somewhat more obvious with

this explanation.

This method

of writing such

lemmas

will

be continued.
<px

With two

prepositional functions,

and

\f/x,

we can form two

impli

S x (#rc^) and IL^.r c i^.r). But 2 x (#rc#r) states is false or \j/x is true: is a value of x for which either only that there and this relation conveys so little information that it is hardly worth while
cation relations,
<px

to study

its

properties.
is

Ux
The

(<px

c^.c)

the relation of
is
<f>x

"jormal

implication"
is

"For

every

.r,

at least one of the two,


negative of
"It

false

and $x

true

is

a true statement"
(
<px

Ux

(<f>x

c^x)

is

S x (^.r
is

x-^.r), so that
.r

Ux
is

i//.r)

may
l

also

be read

is

false that there

any

such that
"At

<px

true and $x

false".

The material
false

implication,
is

peg,

states only
statement";

least
"It

one of the two,


false that

is

and
q

true

is

a true

or,
n

is

is

true
least

and

false".

The material
is
<p

implication,

<px

c\l/x n

states only
,

"At

one of the two,


or
"It

false of
<px

x n and

is

true of x n
false".

is

a true

statement";

is false

that

is

true and

^.r n is

But the formal impli


it

cation, n s (^.rc^.r), states that however x n be chosen,


is

is

false that
is

<px

true and $x n

is

false

in the

whole range of
false.

<px

and $x, there


II X

not a

case in which

<px

is

true and $x

To put

it

another way,

(^

^.r)

means

"Whatever

has the predicate ^ has also the predicate

^"

plies"

of "im This relation has more resemblance to the ordinary meaning it should than material implication has. But formal implication,

be remembered,

nx

(>.r

i//.r)

is

material implications; simply a class or aggregate of c ^.ri, ^.r 2 c ^.r, ^.r 3 c tx~, simply the joint assertion of ^,r,
is
is

etc.,
s

where each separate assertion


The whole question
"implies",

a material implication.

ing of

of material implication, formal implication, is discussed in Section v of Chap. v.

and the usual mean

244

A
The

Survey of Symbolic Logic

properties of formal implication are especially important, because


in the calculus of classes

upon this relation are based certain derivatives and in the calculus of relations.
10-6

n^arc^a-) = U x
[9

(-<px

+ fa)
<pX

= Hx

-(<px

x-fa).
n

3]

(pX n

Cfan = Q.E.D.
n

+ fan

-(

<pX

X -fa n )

Hence
10-61

[10-23]

tt x (<pxcfa)

c(<px

cfa n ).
materially implies fa n

[10-2]
If (px

formally implies fa, then


[II X (
<px

<px

10-611

c $x) x
10-61]

<px

n]

c fa^

[9-4,
If

<px

formally implies fa and $

is

true of

ar n ,

then ^

is

true of x n

This
is

is

for every x, a; one form of the syllogism in Barbara: for example, man implies x is a mortal and Socrates is a man, then Socrates
"If
,

is

mortal".

10 -62

nx nx

( ??ar

c #r) c S x (

<^x

c fa)

[10-22]

10 63

<px

c fa)

c(Il<pxc

Ufa).
then
<px

[10-61] If

U x (<pxcfa),

cfa n

Hence
10-631

[5-3]

Q.E.D.

[U x (<pxcfa)
[9-4,

xn^cn^o*.
<px

10-62]
is

If (px

always implies fa and

always true, then fa


.

is

always true.

10 -64

nx

<^.r

c fa)

c(2<px

c.

2 fa)

[10-61, 5-31]

10-641

[n x (

<f>x

cfa)*2
10-64]

<px]

c 2 fa.

[9-4,
If (px

always implies fa and


[n x (
<px

<px

is

sometimes true, then fa c

is

sometimes

true.

10 65

c fa) x
If

n x (^.T

fa;)]

nx

^.T

far).

[10-61]
iA.r n

n x (^.rc^.iO and U x (fa


.r,

far),

then

<px

cfa n and

far n .

Hence

[5-1]

whatever value of

xn

may

be,

<^.r

far B .

Hence

[10-23]

n x (^arcfar)

Systems Based on Material Implication

245
It is

This theorem states that formal implication


(

is

a transitive relation.
<px

= x is a another form of the syllogism in Barbara. For example let 10-65 will then read: Greek \f/x = x is a man and x = x is a mortal
,

l "If

for every x,

is

a Greek implies x
,

is

man
#

and

for every x,

x x

is
is

man

implies
".

.r

is

a mortal

then for every

x,

is

a Greek implies

mortal

10-65
10-651

may

also be given the form:

Ux

(<px

c #c) c [n x (#r c
10-65]

x)

Ux nx

(<px

fr)].

[9-4,

10 652

U r (ifrx
[9-4,

fa;)

c [U x (

<px

c $x) c

<px

far)].

10-65]

10 66

nx

#r c

[3-1]

= n x (-#c c - #x) (^ci/O - Htfc -?.!)


i/a*)
.

Hence

[2-2, 5-3]

Q.E.D.

Any

further theorems concerning formal implication can be derived

from the foregoing.


"Formal equivalence" is

reciprocal formal implication, just as material

equivalence

is

reciprocal material implication.

The

properties of formal

equivalence follow immediately from those of formal implication.

10-67

Ux
to

=
(<px

jx)

[tt x (vx

c $x) xll x (tx c

<px)].

Whatever value of x, x n may the pair, n c \//x n and \f/x n c


<px

be, [2-2]
n
.

<px

is

equivalent

<px

Hence
10-68
[U x

[10-23]

Q.E.D.

=
(<px

jx)

xn x (^.r =

f-r)]

cn

x (^.i

T-r).

<px

Whatever value of .r, .r n may = rr. Hence [10-23] Q.E.D.

be,

if

<px

= ^x and

i:

==

far,

then

10-681

Ux

=
(<px

^.r) c[II,(^.r

far)

[10-68, 9-4]

10-682

n z (^.r -

f.c)

c[H x (^r =

^.r)

[10-68, 9-4]

Formal equivalence, as indicated by the


relation.

last three

theorems,

is

a transitive

10-69

Ux

and
(<p

n x (^c -

[2-2,

10-61-62-63]

246
10-691

A
n(^c =
[3-2,
If

Survey of Symbolic Logic

$x)

= Ux

=
(-<px

-#r).

10-23]
to investigate the propositions
(<px

we wish

which can be formed from


sig

functions of the type of


nificance of
<px

may

differ

xfy) and from that of


tf

(<px

+ ty), where the range of

\l/y,

we

find that these will involve

2 Jl^r; ty), H S x (^r; ^), etc. And these are special + \l/y) are special cases of a function of two variables. x\f/y) and

two operators

(<px

(<px

cases of f (.T, y).


in general.

Hence we must

first

investigate functions of two variables

III.

PROPOSITIONAL FUNCTIONS OF

Two

OR

MORE VARIABLES
y), gives

prepositional function of
<p(x

two
y n ),

variables,

<p(x,

the derivative
y)-,

propositions

m y n ),
,

U*<P(X,

S*2

tf

p(a:, y),

2 y n x p(a:,

etc.

The

range of significance of
<p(x,

<p(x,

y) will

comprise

all

the pairs

(x,

y) such that

y)

is
.T 3 ,

either true or false.


etc.,
s

We
i/

here conceive of a class of individuals,


,

#1, xz,

and a

class of individuals, y lf y,, y s

etc.,

such that for any

one of the x

and any one

of the

s,

<p(x,

y)

is

either true or false.

As has already been pointed out, the function may be such that the class of values of x is the same as the class of values of y, or the values of x

may

be distinct from the values of


#",

y.

If,

for

example,

<p(x,

y)

be

"x

is

brother of

the class of

.r

for

which

<p(x,

y) is significant consists of

identically the
nificant.
all
9

same members as the

class of y s for

which
<p(x,

<p(x,

y) is sig

In such a case, the range of significance of

y) is the class of

the ordered couples which can be formed by combining any


itself

member
of

of

the class with


class be
i,

or with

any

other.

Thus

if

the

members

such a

G2

>

10 as, etc., the class of couples in question will be

(ai, ffi),

(fli,
(
2,

a 2 ),
a,),

(i,

a-s),

(a z

ai),

(2, a 3 ),
,

...

(a 3 ,
.
.

aO,

3,

a 2 ), (a 3 a 3 ), ...

Etc., etc.
is

But
y",

if

<p(x,

y) represent

"x

a citizen of
y",

?/",

or

"a:

is

a proposition about

or

"x

is

member
is

of the class

the class of x

and the

class of y s

for

which
9

<p(x,

y) is significant will
"A

be mutually exclusive.
is

Presuming that
Schroder treats
first

brother of

A"

significant

i.

e., false.

10

all relatives

as derived

Alg. Log., in,


relatives.

three chapters.)

But

this

from such a class of ordered couples. (See is an unnecessary restriction of the logic of

Systems Based on Material Implication

247
does not neces

Although

<p(x,

y) represents

some

relation of x

and

y, it

sarily represent

any

relation of the algebra, such as x

cy

or x

y;

and

it

cannot represent relations


<P(XI,

which are not assertable.


prepositional functions of one variable, y.
<p(x

y),

<p(x*,

y), etc., are

Hence H v v(xi,
and properties

y),

Uv

<p(x,

y),

^v

lt

y), etc., are propositions,

the meaning

of

And U y

<p(x,y),

2 v v(x,y),

which follow from preceding definitions and theorems. and 2 t Il x <p(x,y), y) are propositional
<t>(x,

functions of one variable.

We

can, then, define propositions involving


follows:

two variables and two operators, as


11-01

11-02
11

UxUy 2xUy
n*Z Sx S

<f>(x,

y)

==

ll t

{n y

<t>(x,

y)}.

Def
Def.
Def. Def.

==
<p(x,y)

2 x lU y <p(x,y)}.

03

tf <p(x,

y)

==

nx

sy

<p(x,

y)

} .

11-04

tf

?(.r, y)

--=

2,{S^(.r, y)}.

from these definitions that our explanation of the range of ^significance of functions of two variables was not strictly required; it The same con follows from the explanation for functions of one variable.
It will be seen

vention regarding the


the propositions
is

number

of values of variables

and interpretation

of

also

extended from the theory of functions of one

variable to the theory of functions of two.

(Where the
will

first

variable has a subscript, the


is

comma between
.r,

the two

be omitted:
Since

<p(x_y)

p(.r 2

y), etc.)

Uy

<f>(x,

y)

is

a propositional function of one variable,

the defini

tion. 10-02, gives us

IUIX*,
And

y)

==

U^Il^x,

y)}

--=

Tl y

<p(x,y)

xll y

<f>(x

z y)

the expansion of this last expression, again by 10-02,


{

is

<p(xiyi)

x
x
{

<x

x ^Oi*/ 2 ) x x 2) x
<>x?

<p(.ri?/s)

x
x

^(.r 2 z/ 3 )

<p(x*yi)

<p(x

3 y) x

x ... Etc., etc.

And

similarly,
y)

by 10-01,

SJlXz,
And

= S x {n^(;r,

y)}

--=

U
is

the expansion of the last expression, by 10-02,

x
x

x +
. .

Etc., etc.

248
Or, in general,

Survey of Symbolic Logic


is

any prepositional function with two operators

expanded

into a two-dimensional array of propositions as follows:


(1)

The operator The

nearest the function indicates the relation (+


line.

or

x)

between the constituents in each


(2)

subscript of the operator nearest the function indicates the


the lines.
(

letter
(3)

which varies within

The operator
The
from

to the left indicates the relation

+ or

between

each two lines.


(4)

subscript of the operator to the left indicates the letter which


line to line.

varies

Some
2xUv
<p(x,

caution must be exercised in interpreting such propositions as


y}, etc.

It is usually sufficient to
it

read ZJI y
x,

"For

some
such

x.

ami.

every

y",

but
<p(x,

strictly

should be
"For

"For

some

every y
is

is

that".

Thus 2 x H y
true".

y} should be
<p(x,

some

x,

every y every

such that

<p(x,

y) is

And H v 2 x
is
true".

y) should be

"For

y,

some x

is

such that

v(x, y)

which

ma

The two here chosen be made clear as follows:


2 x ll y
<p(x

illustrate the necessity of caution,

y)

=
"Either

That

is,

2x

IIy<p(x,

y)

means
<p(x,

for x\

or for # 2

...

or for

and every y, y) is true, or for some other particular x and every

and every y, .T 3 and every y,


y,
<p(x,

<p(x,

y) is true, y)
is

<p(x,

true,

y)

is

true".

On

the

other hand,
Tl y

2x

<p(x,

y)

= 2x

<p(x,

i/i)

xS x ^(.r,

*/ 2 )

xS x
yi,

<^(.T,

y
y) is true,
<p(x,

That

Uy^ x (p(x, some x and y z


is,
,

y)

means

"For

some x and
for

<p(x,

and

for

<p{x,

y) is true,

and
is

some x and y 3

y) is true, y)

and

.";

or

"Given

any y, there

one x (at

least) such that

<p(x,

is true".

The

following illustration of the difference of these two is given in Principia Mathematical n Let y) be the propositipnal function y is a proper
<p(x,

"If

fraction, then x of y,

is

a proper fraction greater than


?/), so

y".

Then

for all values

we have 2 x

<p(x,

that
"If

Ii y

2x

<p(x,

y)

is

satisfied.

In fact,

U^
is

<p(x,y)

expresses the proposition: y a proper fraction greater than


sition:
tion",

is

a proper fraction, then there

always

?/".

But 2 x li y
is

<p(x,

y) expresses the

propo
frac

"There is

a proper fraction which

greater than

any proper

which

is false.
if

In this example,
11

we should read S x n y

"For

some x and every

#";

See

i,

p. 161.

Systems Eased on Material Implication

249
these two

IlySz

"For

every y and some

x",

we should make equivalent


this caution
is

very different propositions.


infrequent, as

But cases where

required are

we
r

shall see.

Where both operators

are

II

or both 2, the two-dimensional array of

since every rela propositions can be turned into a one-dimensional array, tion throughout will be in the one case x in the other + and both of
,
,

these are associative

and commutative.

It follows

from our discussion of

the range of significance of a function of two variables that any such func the ordered tion, y), may be treated as a function of the single variable,
<p(x,

couple, (x, y).

Hence we can make the further conventions:

11-05

SsSy^Gc, y)
tt x ll y
<p(x,

= S (x

V )<p(x,

y)
y)

==

S Xf

<f>(x,

y).
y).

11-06

y)

= n (x

V )<p(x,

==

nx

<?(x,

The second
1 1

half of each of these serves merely to simplify notation.


<p(x,

is

x r and y a be any values of x and y, respectively, in == r y s ). such that a value of (x, y) y) n say, (.T, y) n

07

If

y)

there

<f>(x,

<p(x

11-05 and 11-06 could be derived from 11-07, but the process is tedious, and since our interest in such a derivation would be purely incidental,

we

prefer to set
If

down

all

three as assumptions.
y
<p(x,

we wish
2x 2y

to identify a given constituent of S*,


<p(x,

y) in

with a con

stituent of
is

y),
if

some convention

of the order of

terms
<p(x,

Sz

<p(x,

y)

required, because

the order of constituents in 2 x 2 v

y) be unaltered,

this identification will

determined

be impossible unless the number of values of y is we which, by our convention, need not be the case. Hence
of

make, concerning the order


tion:
if
<p(x

terms

in

S Xf
<

<p(x,

y),

the following conven


in

m yn) precedes

<p(x

y s)

if

m+

r
y

s,

and where

n =

+ *,

<

s.

Thus the order


+
<p(x

of

terms

in S*,

<p(x,

y) will be

<p(xiyj

2 yi)

This arrangement determines an order independent of the number of values r y e ) can always y) n in terms of of x, or of y, so that the equivalent of
<p(x,
<p(x

to govern the be specified. exactly similar convention is supposed with the terms y) and their identification arrangement of terms in n*, y
12

An

<p(x,

of

U x ll y
12

<p(x,

y).

These conventions

of order are

never required in the

proof of theorems:

we note them

here only to obviate

any

theoretical

familiar device for denumerating the rationals i. corner. agonals, beginning with the upper left-hand

the This arrangement turns the two-dimensional array into a one-dimensional by successive e., by proceeding along

250

A
The
y)

Survey of Symbolic Logic


of
is it

objection.

identification

2x2y
will

<p(x,

y)

with

Sx

<p(x,

y),

and

of

HJIy^Gr,

with

nx
is

<p(x,

y),

of little

consequence for the theory of


of functions of

propositional functions
of relations

itself,

but

be of some importance in the theory

which

to be derived

from the theory

two or

more
to

variables.

Having now somewhat tediously cleared the ground, we may proceed the proof of theorems. Since 2 X y y) may be re y) and H x y
,

<p(x,

<p(x,

garded as involving only one variable,


at once from

(x, y),

many theorems

here follow

those of the preceding section.


y)

11-1

sx

<p(x,

= ^X^ V

<P(X,

y}

= -H z

y -<p(x,

y)

= -[U x U v

-<t>(x,

y)}.

[11-05-06, 10-05]

11-12

Xf y

<p(x,

y)

= U x ll y

<p(x,

y)

= -S Xf

v -<p(x,

y)

= -{S^-pfo

y)}.

[11-05-06, 10-04]

11-2

Hx

<p(x,

y)

<p(x,

y) n

[10-2]

11-21

<p(x,

y) n

c2 Xry
y)

<p(x,

y).

[10-21]

11-22

nx
nx

<p(x,

c s*

<p(x,

y)

[10-22]

11-23
(x, y) n

<f>(x,

y) is

equivalent to
".

"

Whatever value

of (x, y), in

<p(x,

y),

may

be,

<p(x,

y) n

[10-23]

11-24

U x Tl y

<p(x,

y)

is

equivalent to
be,
<p(x

"Whatever

values of x and

y,

in

v(x, y), x,

and

y,

may
r

r y^".

[10-23]
IL y
<p(x,y),

HJI^O, y) is x may be, H y


value of
y, in
<p(x

equivalent to
<p(x

"Whatever
<f>(x

value of

.r,

in

yY\
y
s

And H y

y)

is

equivalent to

"Whatever

(f>(x,y),

may

be, v(x

r y^)".

But
y).

[11-01]

the values of x in

Hy

y) are the values of

x in

<p(x,

Hence

Q.E.D.
"

11-25
is

"Whatever

value of (x,y), in
values of
.r

<p(x,y),

(x,y) n

may

be,

<p(x,

y) n

equivalent to
"

"Whatever

and

y, in (p(x, y),

x r and y,

may

be,

<p(x

ys]
[11-06-23-24]

11-26

n x n,

<p(x,

y)

tf

<f>(x

n y)

[11-01, 10-2]

Systems Based on Material Implication

251

11-27

n x n^(.r,
S x S^Gr,

y)
is

= U v TL x

<p(x,

?/).

Since x

associative

and commutative, Q.E.D.


y).

11-28

y)

==

S S x ^(a;,
tf

Since +
11

is

associative
.

and commutative, Q.E.D.

29

IIJI,

<p(x,

y)

c n x ^(jr, y n )

[11-2(3-27]

11-291

lUI^C*,
[2-2,

y)

c<p(av#,).

11-24]
y)

11-3

IIJI^O,

c SJI^Gc,

y).

[11-01, 10-21]

11-31

SJI^(a;,
[11-03]

y)

cU

2x

<p(x,

y).

ny s x

<p(;e,

y)

vfayi) + vfayd + + ^22/2) + x


I {

<p(x

yi)

+ + +

<K#i2/2)

^(^3^/2)
^(.TSZ/S)

<p(xiyz)

^(.r 2 z/ 3 )

x ... Etc., etc.


Since x
to the
is

distributive with reference to +


of the products of each
is,

this expression is equal

sum

column separately, plus the sum

of all the cross-products, that

to
z)

+ +

p(.ri2/i)
<f>(x

x x

<f>(xiy

<p(.ri#

3)

x x

z yi) x

<p(x

yz)

<f>(x

y3)

Etc., etc.

where

/I is

the

sum
is

of all cross-products.

But

[11-02] this

SJI^O,
==

y)

+ A.
/).

Hence SJI^Or, i/) +^4 Hence [5-21] S x n y ^(.r,

n^S^Gr,

y)

c I^Z^Gr,

y).

We
is

have already called attention to the fact that the implication of 11-31 not reversible that S x n^(.r, y) and I^S^Cr, y) are not equivalent.

11-32

nx

2y<t>(x,

y)

c Sx S

tf

^(.r, ?/).

[11-03] n,S,

[11-04]

S x S^(a-,
S
c2

And

[5-992]

We

have

also the propositions concerning formal implication

where

252
functions
of

Survey of Symbolic Logic


of

two variables are concerned.


<f>(x,y)

$(x,y)

by
f

The formal implication y) c \//(x y)] may be written either U Xt v


[<p(x t

or

UsUy
11-4

[<f>(x

y) Ci//(x, y)].

By

11-06, these two are equivalent.


first of

We

shall

give the theorems only in the


IT X
,

these forms.
y) y)

y [<p(x,

y)

c t(x,

y)]

= H = Ux

x>

v [-<p(x, v -[<p(x,

+ j(x,

y)]
y)].

x-^(z,

[10-6]

11-41

nx

y [<p(x,

y)

c t(x,

y)]

c[<p(x,

y) n

c f(x,

y) n ].

[10-61]

11-411

( (

nx

y [<p(x,

y)

c t( x

y)]

<p(x,

y) n }

c j(x, y).

[10-611]

11-42

nx

y [<p(x,

y)

c$(x,

y)]

c Sx

y [<p(x,

y)

cf(x,

y)].

[10-62]

11-43

n*,

[<f>(x,

y)

^(.r, y)]

c [n,

^>(a;,

?/)

Xf y

$(x, y)].

[10-63]

11-431

{n x

,[^(.r,

y)cf(x,

y)]

xH x

^(x,

?/)}

nx

y }(x,

y).

[10-631]

11-44

nr

y [<p(x,

y)

^(.T, ?/)]

c [2 Xt

y<p

(x, y)

c S,

y ^(.r,

?/)].

[10-64]

11-441

{n x

tf

[^(a;, y)

^(ic, y)]

x Sx

,^(.T, y)}

2,,

y $(x,

y).

[10-641]

11-45

{n x

y [<p(x,

y)

cj(x,

y)]

xH

If v [^(a;,

y)

cf(a

[10-65]

11-451

n Xftf [^(ic,2/)c^(a:,y)]
c {n x
[10-651]
,

[^fe

2/)

cr(.T, y)]

cn

x>

J^GT,

y)

c{(x, y)}}.

11-452

nx

[*(x,y)ct(x,y)] c {n^,

y [<^(^

y) c^(.r,

?/)]

cn

tf

[^(a;, y)

c rfe

y)]}.

[10-652]

11-46

n,,

y [<p(x,

y) c^(;r,

?/)]

= nx

,[-^(.1%

i/)

c-^(.r,

y)].

[10-66]
Similarly,

we have the theorems concerning the formal equivalence

of

functions of two variables.

Systems Based on Material Implication

253

11-47

n*,

v [v(x, ?/)

}(x, y)]

=--

(n x

v [<p(x,

y)cf(x,

y)]
y)c<p(x,

xn,, [*(*,
[10-67]

y)]}.

11-48

{n*.

y [*>(ar,

y)

}(x, y)} xll x

v [t(x,

y)

(x, y)}}

cU
[10-68]

y [<p(x,y)

(x, y)}.

11-481

II X

[*>(*,

?/)

iKz,

?/)]

{II X

,[^0, y)

f(z, y)]
f

c n,
[10-681]

[*>(*,

y)

f(x,y)]|.

11-482

n,

y [^(o;,

y)

- rfe

2/)]

c fn x

tf

[^(ar, y)

t(x, y)]
v [<p(x,y)

cU,,
[10-682]

= rfe

11-49

Hx

J^GT, y)

t(x, y)}

[^(a;, y) n

t(x, y) n ]

c[U c[2 x
[10-69]

x>l

,<p(x,y)

<p(x,

y)

= Ht = S Xf
y)

t(x,y)]
t

*(x

y)].

11-491

n,

tf

[^(.r,

i/)

- ^fe

?/)]

= nx

y [-^(ar,

-j(x,

y)].

[10-691]

Further propositions concerning functions of two variables are simple


consequences of the above.

The method by which such functions are treated readily extends to those of three or more variables. y, z) may be treated as a function
<p(x,

of three variables, or. as a function of one variable, the ordered triad

(a:,

y, z)

of x and y, or of the ordered just as \l/(x, y) can be treated as a function with each extension of Strictly, new definitions are required pair (x, y). number of variables, but the method of such extension our theory to a larger
will

be entirely obvious.

For three
y, 2)

variables,
f

we should have
t

njiji^O,
SJIyll
,>(&,

y, z)

= u x {u v u = S x {lI n
1/

tf>(x

y, z)}

^(.T, y, z)\

Etc., etc.
It is interesting to note that the

most general form

for the analogues of

11-05 and 11-06

will

be

n
and
Since

*)*>(.r,

y, z) y, z)

==

S (I
y, z)

==

,)<t>(x,

njl^, S x S (y
,

g )<p(x t

y, z)

f)

^(ar, y, z)
z y,

njl (v

,)<p(x,

= n

(j/

,)<p(xiy,

z)

xn

(y ,

g}

<p(x

z)

xn

(l/

,)^(x 8 y, z)

254
x

A
.

Survey of Symbolic Logic

.,

and

n (y
y
,

Z )<p(x n y,

2)

= Uy

llz<p(x

n y,

2), etc.,

we

shall be able to deduce

n
And

( x,

Z)<P(X,

y, 2)

= nji = n n
tf

(I ,,

2)

^(.r,

?/,

2)

(I>

Z)<P(X,

y, 2)

= n ^iwo, #, 2) = njiji^O, y, 2)
(X|
<p(x,

similarly for

S (I

2)

This

calls

our attention to the fact that

as a function of three variables or as a function y, 2) can be treated not only of one, but also as a function of two, x and (y, 2) or (x, y) and 2 or (x, 2)

and

?/.

In general, the conventions of notation being extended to functions of any number of variables, in the obvious way, the analogues of preceding

theorems for functions of two

will follow.

We

failed to treat of

such expressions as

U<px

xUfy,

2<px

U\f/y, etc.,

under the head of functions of one variable.

The reason

for this omission

was that such expressions


of the type H. x l\. y of functions of
(<f>x

find their significant equivalents in propositions

x^z/),

2 x Hy

(<px

+ ty),

etc.,

and these are

special cases
of the

two

variables.

We may
<p

also remind the reader


<px

+ U\l/x and n tpx + U\{/y. between two such expressions as II need not be identical; there and The ranges of the two functions,
difference
\f/,

may

be values of

.r

in

<px

which are not values


,

of y in $y.

But

in

any

expression of the form

<px

x fa n x n as a value of x in #x must be identical

with x n as a value of

.r

in ^.r.

For

this reason,
is

we have adopted the con


Hence the case where

vention that where the same letter

used for the variable in two related

functions, these functions have the


\[/y is

same range.

the more general case, in which the functions are not we have px and Theorems involving functions of this type restricted to the same range.

not always be significant for every choice of v and \J/. There may even be cases in which an implication is not significant though its hypothesis is
will

significant.

But
true;

for

whatever functions such theorems are

significant,

they

they will never be false for any functions, however chosen. + \f/y) follows from the The meaning of an expression such as 2 x U y (
will

be

<px

definition of

2x

Tly<f>(x,

y).
(
i

S x IIj,(

<px

\j/y)

= II, =
+ +

<px

i/^)
z

Il y (

<px

.r

<xi

+ $y) + U v ( + \/ z x

<px

+
\/

^)
3

+
.
.

<xi

{(VP.TS
. .
.

^I) x( ^3 +

^2/2)

x( ^3 +

^2/3)

Etc., etc. of

And

for

any such expression with two operators we have the same type

Systems Based on Material Implication


two-dimensional array as for a function of two variables
only difference x \{/y, etc. or
<px

255
in general.
<^.r

The
+
\fry

is

that here the function

itself

has a special form,

12

(1)

[1-3]
II
<>.r

n^.r

(2)

II\l/y

= =

<pX]_

</?.r

x
(<p.r

(<f>X!

x n^i/) x

(>.r 2

x n^z/) x

x Ityy) x

[5-98]

xn
(3)

tf

(^r 8 x^y)x...

[10-371]

[ii-oi]

By

(2)

and

1-3,

tpxi)

x (n^y x

<^.r

2)

(H^
3)

(p.r 3 )

x ...

n y (^
xll
.

<^.r

2)

tf

(^x^.r
^.r)

x ...

[10- 37]

[11-01]

(4)
"

Similarly, ll^y
.r

xH^r = U y U x (\f^y
is

=
is

11^(^.1- x^y).
"For

<px

is
a;

true for every

and

\f/y

true for every


true",

z/"

equivalent to

every
12-2

and every

y,

<px

and

^z/

are both

etc.

S^.r + S^z/

= 2^z/+2^.r = ^^^(^.r
S^i- +

(1) (2)

[4-3]
2<p.i-

S^

- S^+S^r.
i

S^

= (^

^2 + ^3 +...)

->/

[5-981]

[10-31]

SxZ^ +

ifc/).

[11-04]

(3)

By

(2)

and 4 -3,

[10-3]

S x S,(^+^r).
tf

[11-04]

(4)

Similarly,

S^+2^.r = S

256
for

A
some
x,
<px,

Survey of Symbolic Logic


or for
\f/y",

"Either

some
etc.

y,

tyy"

is

equivalent to

"For

some

x and some
12-3

y, either

<px

or

S^r

=
[1-3]

S5

(1) (2)

S^c S#e S^y =


x X S^y) +

O?2
(

X S^y) + (^T 3 X
[5-94]

+ S

tf

x $y) + S y ( +

x ^y)
[10-361]

...

S^C^cx^).
S
x
i

[11-04]

(3)

By

(2)

and
x

1-3,

Si =

S^
y

x ^cj +

<^x

+ Sy
=

(^x^ )+
8
tf

..-

[10-36]

SA(^x^).
#,
^?/"

[H-04]
#r)

(4)
"For

Similarly,
<^.r,

S^y xS^o; = S y S,(^ x


some
is

- S S x (^c x^).
"For

some
y,
<px

x,

and

for

equivalent to

some x and

some
12-4

and
+

^y",

etc.

v?.r

n^
H<px

n^ + n^
IL<f>x

= nji^z
=

^) = n*n y (^ +
(<px

<?x)

UyUx
3

== Il v

U x (frj+

<px).

(1) (2)

[4-3]

+ TIty

= U^y +
x
<^x

U^px.

+ n^?/

=
=

(^aji
(<pxi

<px

+ ttty) x

(^2

+ n^z/ ) + n^i/) x (^T 3 + n^y) x


. .

...

[5-941]

+
=

M
x

x...

[10-371]

n s n (^ + ^).
tf

[ii-oi]

(3)

By

(2)

and 4 -3,
+ n^?/

= (n^y +

^o x (n^ +

^2)

n
.T 3 )

x...

[10-37]

(4)
"Either

Similarly,
x,
<px,

Uty

H<f>x

= U v U x (^y +
\f/y"

<px)

= U y tt x

(<px

for

every

or for every y,

is

equivalent to

"For

every x

Systems Based on Material Implication

257

and every
invalid.

y, either

<px

or

\f/y",

etc.

At
<px

first

glance this theorem


If
a-

may seem
it is

One may
If

say:

"Suppose
it is

be
.

is

a number,

odd

and $y be

is

a number,

even

Then
even
,

U<px

U^y

will

be
\f/y)

Either
will

every number is odd or every number Every number is either odd or even
illustration lies in misreading TL x

is
".

but

U x ll v

(<px

be

The mistake
\^y).

of this supposed

Uv

(<px

It
is

is

legitimate to choose,

as in this case,

<px

and

\f/y

such that their range


+ $y) as
if

identical:

but

it is

not

legitimate to read

UxUy Ux

(<px

with

a corresponding value of y.
lly(<px

To put

each given value of x were connected nzllx (#c + #c), it another way:

as a special case of
either
of x
<px

or

\l/x",

but would be
<px

+ ty), would not be "For every value of x, "For any two values of x, or for any value
$x"
.

and

itself,

either

or

Thus HJIj/^T +
is

1/^)

in the

supposed

illustration would not be as above, but

in fact
is

or for

any number and


+ TLfy and

itself,

either one

any odd or the other


"For

pair of

numbers,
so

is even"

that
lent.

U<px

Il x lly(<px

+ $y) would here both be

false,

and are equiva

A
12-5

somewhat

similar caution applies to the interpretation of the next

two theorems.
2<f>x

The analogues

of these, in

<p(x,

y),

do not hold.
^x

Wy
By

= n^y + 2

<px

= 2 X II V

<f>x

+ ty)

==

(1)
(2)

[4-3]

proof similar to

(2) in 12-2,

2<px

And by (3) By proof similar to And by proof similar to


"Either

proof similar to (3) in 12 2,


(2) in 12-4,

<px

Ufy = S x n v (^r + + Ufy = S,n (^ +


tf

ty)
<px)

(3) in 12-4, Uifry


is

H^y + S + S^r

==

Uy ^x
"For

for

some

or,

^r, or for every y,


<?x

^"

equivalent to

some x

and every
12-6

y, either

or

^?/",

etc.

s<^.r

S^r By proof similar to And by proof similar to


(1)

[1-3]

(2)

(2) in 12-3,
(3) in 12-3,

S^c xltyy
2<px

==
==

S*n (^c x
tf

xH^z/

S,H y (^ x

By proof similar to And by proof similar to


(3)
"For

(2) in 12-1,
(3) in 12-1,
is

Ityy

Ityy

xS^c xS^r

==

n H

tf

2 2

some
y,
<px

x,

<px,

and

for every y,
etc.

fy"

equivalent to

"For

some x and

every
18

and

^?/",

258

A
We may

Survey of Symbolic Logic

generalize theorems 12-1-12-6

by saying that

for functions

of the type

(<px

\f/y)

and

(px

\j/y)

the order of operators and of

members

in the function is indifferent;

and

for propositions of the type

n
l 2J
the operators

may

be combined, and the functions combined in the relation

between the propositions. It will be unnecessary to give here the numerous theorems which follow from 10-5-12-6 by the principles pqcp, pcp + q, and Il^xc Spx, etc.

For example, 10-51,


IlipX

gives at once
(i)

nx

(<px

(2) (3)
(4) (5) (6)

Etc., etc.

And

12-2,

2<px

2\j/y

= 2x 2v
(1) (2) (3) (4) (5) (6)

+
(<f>x

\fry),

etc.,

gives

S,

2<px

2^z
from

Etc., etc.

Another large group


the combination of

of theorems, only a little less obvious, follow


cH.<px,

H<px

or

S#ccS#r,

with

Ii\f/y

c S^y, giving

by 5 -3,

(3) (4)

<px

+
x

Tl\{/y
TL\f/y

<px

c 2 (px + ^^y x 2^?/ c2


<px

Etc., etc.

Each

of these has a

whole

set of derivatives in

which

n (px + U\I/y is replaced

Systems Based on Material Implication

259
etc.

by

H. x lly(<px

1/

?/),

etc.,

H<px

xZi/ Z/ by
(2),

n x S (^x^),
J/

We

give, in

summary form,

the derivatives of

by way
any one

of illustration:

Any one

of

of

xU\l/y

x
x

<px)

<px)

S n x (^ x
tf

U y Ux

9x x ^)

S vnz (

<px

x
<?x

n ipx

zty

xn

etc., etc.

etc., etc.

^),
This table summarizes one hundred
fifty-six

etc., etc.

theorems, and these are only a

portion of those to be got by such procedures. and ((px + Functions of the type of
(<f>xx$y)

g ive four different


(2)

kinds
(3)

of

implication
,

relation:
(4)

(1)
tf

njl^rc^);
$y).
"implications",

n x S v (^c^);
of the
of
all

2 x n y (<pxcty)

and

S x S (^c c
field

With the exception


though

first,

these relations are unfamiliar as


of

them

could be illustrated from the


ticularly useful:

mathematics.

Xor

are they par

the results to be obtained by their use can always be got


material
implications
or

by means

of

formal

implications.

Perhaps

UJl y
12-7

(<px

c $y)
tf

is

of sufficient interest for us to give its elementary properties.

n x n (#cc
[11-01]

y^
i

c Ufy) x

(^

c
...

[10-42]

And

this last expression is equivalent to the set


<pxi

Uifry,

<pxz

c H^y,

<px

c nty,

etc.

12-71

{IIJI

J/

Ozci/

?/)

x^r B

[9-4, 12-7]

260
If for

A
every x and every
is

Survey of Symbolic Logic


implies

y,

<px

\l/y,

and

for

some given

x,

<px

is

true

then $y
12-72

true for every y.


<px

Il t Tl v (

c }y) = S ^.r c Ilty =

II

- #c
<p

(1)

If

IIJI^Oo; c^y), then [12-7]

Hence

[5-991]

And

if

S^cltyi/. S^ccltyy, then [10-42]

tt y (2

<px

c ty)

and hence [10-41]

(2)

[9-3]
is

S^c
"If

HxHy

(<px

c $y)
is

equivalent to
y".

there

is

some x for which px

is

true,

then $y
12-73

true for every

{nji w (#rc^)
[12-72]
If

xn y n,(^cf2)} cn x n,(^c^).

IlJIyC^rc^) and n v n(^i/cfz), then


Hence
[5-1]

and S^y c Hfz. But [10-21] Il^yc^y.

S*ccII3, and

[12 72]
:

This implication relation is here demonstrated to be transitive. In fact, it is, so to speak, more than transitive, as the next theorem shows.

12-74

{(S^ccS^) xn n,(^cfz)} cIIJI.C^c cfz). = s^z/cn^. [12-72] n v n (^crz) And [5-1] if S^c cZfy and S^y c Hfz, then S^.T
tf

c Hfz, and [12-72]

12-75

(n x
[12-72]

ns
if 2<pxcttty

And

[5-1]

and H^y c Hfz, then S^c cHfz, and [12-72]

IV.

DERIVATION OF THE LOGIC OF CLASSES FROM THE CALCULUS OF


PROPOSITIONAL FUNCTIONS

The

logic of classes

and the

logic of relations

can both be derived from

the logic of propositional functions.

In the present chapter, we have

begun with a calculus of propositions, the Two-Valued Algebra, which includes all the theorems of the Boole-Schroder Algebra, giving these
theorems the propositional interpretation.
sidered as belonging to the calculus of

have proved that, con propositions, these theorems can


"If
.

We

validly be given the completely symbolic form:

then

."

Systems Based on Material Implication


being replaced by
"...

261

...","...

is

equivalent to

..."

by

".

.",

etc.

The
it

Two- Valued Algebra does not presuppose the Boole-

Schroder Algebra;

simply includes it. the calculus of propositions the Two-Valued Suppose, then, Algebra our fundamental branch of symbolic logic. We derive from it

we make

the calculus of prepositional functions by the methods of the last


sections.

We may

then further derive the calculus of

logical classes,

two and a

calculus of relations,

by methods which are to be outlined

in this section-

and the next.

The present section will not develop the logic of classes, but will present the method of this development, and prove the possibility and adequacy At the same time, certain differences will be pointed out between of it. the calculus of classes as derived from that of propositional functions and
the Boole-Schroder Algebra considered as a logic of classes.
distinguish class-symbols from the variables, x,
tions,
y, z, in

In order to

propositional func

we

shall here represent classes

by

a, 0, y, etc.

For the derivation

of the logic of classes


is

from that of propositional

conceived as the aggregate of individuals for functions, a given class is a man", n represent which some propositional function is true. If
"a:
<px

then the aggregate of x

s for

which

<px

is

true will constitute the class of

men.

If,

then,

propositional

represent the aggregate of individuals for which the will be "the class determined by is true, function
z(<?z)
<pz

z(<pz)

the function
acter
etc.
13
<>".

<pz",

or

"the

class

determined by the possession of the char

We

can use

a, 0, 7, as
"a

an abbreviation

for

z(<pz),

2(^2), z(&),
<pz",

z(tpz) will

(In this connection,

is the class determined by mean and we should remember that


<px

the function
<pz

are the

same

function.)

The

relation of

an individual member
"

of a class to the class itself will


is

be symbolized by e. x n e a represents x n This relation can be defined. n is an


"x

member

of

a "-or

briefly

a".

here borrow the notation of Principia. The corresponding notation of Peirce and Schroder involves the use of S, which is most confusing, because this S has a meaning But in of a propositional function. entirely different from the S which is an operator such does not represent an aggregate of individuals; it represents Principia, And z(tpz) is not a primitive idea but a notation supported by an elaborate that Our procedure above is inelegant and theoretically objectionable: we adopt it theory. because our purpose here is expository only, and the working out of an elaborate technique would impede the exposition and very likely confuse the reader. As a fact, a more satis
13

We

"

z(<pz)

<pz".

factory theory on this point

makes no important

difference.

262
13-01
"x

A
xn
a
z(<pz)

Survey of Symbolic Logic


Def.

=
<px

n is

member

of the class

determined by

<pz"

is

equivalent to

"

<px

n is

true".

(For convenience of reference,

we continue
is

to give each definition

and

theorem a number.)

The
class
"a

relation

"a

is

contained in
of

ft"

the relation of the class a to the

ft

is

when every contained in


and

member
ft"

a
ft.

is

by a c
"is

a member of (3 also. We shall symbolize The sign c between a and ft, or between
in";

z(vz)

2(^2), will be
",

contained
e

c between propositions
e

will

be

"implies

as before.

xn

is,

of course, a proposition; x

a, a prepo

sitional function.

13-02

acft =
is

U x (xeacxeft)
"For

Def.
c

a c

ft

equivalent to
e o:

every

x,

is

an a implies x

is

ft

".

nx

(>

cx

e (3)

is

a formal implication.
is

that the logic of classes

It will appear, as we proceed the logic of the formal implications and formal

equivalences which obtain between the propositional functions which deter

mine the
13-03
(

classes.

ft)

= U x (x e a =
"For

e (3)
{

Def.

a
to

=
x

ft

is

equivalent to

every

is

member
a c

of

/3

".

a =
e.,
ft

ft

x is a member of a is equivalent x, thus represents the fact that a and ft

have the same extension


xn
e

i.

consist of identical

members.
propositions.

a,

ft,

and a

are assertable relations

But

the logical product of two classes, and the logical sum, are not assertable
relations.

They

are, consequently, defined

not by means of propositions

but by means of functions.


13-04

axft

= x[(xea) x(xe
of

ft)}
ft,

Def.
is

The product
which
this

two

classes,
"x

a and

the class of x
is

propositional function
is

is

an a and x

ft".

determined by the The class of the x s for


s

true constitute

a*

ft,

the class of those things which are

both as and

/3 s.

The relation x between a and ft is, of course, a different relation from x between propositions or between propositional functions. A similar remark applies to the use of + which will represent the logical sum of two classes,
,

two propositions or propositional functions. This double use of symbols will cause no confusion if it be remembered that a and ft, and 2(^2), etc., are classes, while x e a is a propositional function,
as well as of
(<pz)

and x n

a,

a c

ft,

and a =

ft

are propositions.

Systems Based on Material Implication


13-05

263

a+
of

/3

= x{(xea)
classes,

+ (x6(3)}

Def.
is

The sum

two
x
is

a and

/?,

the class of z
,

such that at least one

of the two,

an
is

and x
an a or x

is is

a
/3.

is

true, or loosely, the class of z s

such that either Z

The negative
13-06

of a class

can be similarly defined:


Def.
is false.

-a = x-(xta)

The negative of a is the class of or s for which x is an a The "universe of discourse", 1, may be defined by the

device of selecting

some prepositional function which is true for all values of the variable. Such a function is (f.r cfz), whatever prepositional function x may be.
13-07
1 is

= x({xcx)
xs
x
s.

Def.

the class of
all

for

which x implies
"null-class",

x.

Since this

is

always true,
1.

1 is

the class of

The
Def.

0, will

be the negative of

13-08

=
is,

-1

That

by

13-06,,

= x-(txcx), and
s will
/3

since

-(fac^)

is

false for all

values of x, the class of such x

be a class with no members.


2(^2).

This will be tt x #r). c$x), and a = assertable relations of establishes at once the connection between the

Suppose that a Hence a c ft will be

2(^2)

and

Then, by 13-01, x n
==
(<px

==
<px

n.

Ux

(<f>x

classes

and formal implication and equivalence.

To

illustrate the

way

in

which

this connection enables us to derive the logic of classes

from that

of

propositional functions,
It will

prove a number of typical theorems. be convenient to assume for the whole set of theorems:

we

shall

a
13-1

= 2O),

]8

2(*z),

&(&)

= x-({xc{x). = -1. Hence [13-06] = x -(x * 1). = f.r c c. Hence = x -(s*.r [13 01 07] x e 1

f .r)

13-2

n x (xel).
[13-01-06] x n el

Hence

U x (x
-

1)

== (x n cxn). = U x (xc{x).

But

[2

2]

fr n c

r-r n .

Hence
a

[10 23] of the

n,(^ c

x)

Every individual thing

is

member
.

"universe

of

discourse".

the range of discourse", 14 This defines, not the universe of discourse, but "universe With 1 so defined, propositions which invol of significance of the chosen function, f whenever t, etc., a the classes (**), (*), etc., and 1, will be significant
<p,

the same range, and true

if

significant.

264
13-3
II x

A
-(.eO).
[13 -01 -06 -07]
a- n

Survey of Symbolic Logic

eO =

-(f.r n cf.r n ).

Hence

[3-2] ~(x
f.r.

n e

0)

(r n

c far)

But $x n c
For every
class.

Hence
that x

[10-23]
e

n x (fx

fr),

and
is

H x -(a;

0).

x, it is false

no individual

member

of the null-

13-4

cl.
[13-01-06] X n tl
Since a

frr n Cf.r n ).
e

%(<pz),

[13-01] x n

ipx n -

c [*c n c (fz n c &)]. [9-33] r n ). Hence since f.r n c f.r n n c ($x n c c ({x c{x)], and [13-2] a cl. Hence [10-23] tt x
cf.r n )
,
<p.T

On

[<px

Any

class,

a, is

contained in the universe of discourse.

It will

be noted

(13-2 and 13-4) that individuals are members

of 1, classes are contained in 1.


"A

In the proof of 13-4, we

make
c $x n

use of 9-33,
is

true proposition
it is

is

implied
n.

by any

proposition".

$x n

true.
x,

Hence

implied by

<px

And

since this holds,

whatever value of

xn

may

be, therefore,

But

(px is

the function which determines the class


1.

fa:

cfa:, the function

which determines

Hence

<px

n is

xn

a,

and

xn

cx

n is

xn

e 1.

Conse

quently we have 11^ (a:caca:cl). contained in", this is a c 1.


"is

And by

the definition of the relation

13-5

Oca.
[9-1]

-O n e 0)

is
e

equivalent to (x n

0)

0.

Hence

[13-3] (x n

0)

0,

and
a,

[9-32]

(x n e 0)

<px

n.

Hence
Hence

[13-01] x n
[13-02]
is

eOcx n e
a.

and [10-23]

U x (xeQcxe

a).

The
"A

null-class

contained in every

class, a.

In this proof,
-(fa: n

we

use 9-32,

false proposition implies


.
<px

hence implies

and

<px,

and any proposition". But -(fxcfx) is the function which determines 0; n the function which determines a. Hence c a.
cfa: n )
is false,

The

proofs of the five theorems just given are fairly typical of those

which involve

and

1.

But the great body


This connection

of propositions
of classes

make more
and formal

direct use of the connection

between the relations

implications or equivalences.
following:

may

be illustrated by the

Systems Based on Material Implication


13-6
z(

265

= n x ( (px c \j/x) [13-02] z((pz) cz(^z) = U x [xe z((pz) c.ce = (px n and x n e z(\f/z) = $ [13-01] x n e z((pz} ex e (^z)] = U x Hence [2-1] U t [x e
(pz)

z(^z)

z(<pz)

(<px

c
\f/z"

"The

class

determined by
"For

(pz is

contained in the class determined by


\frx".

is

equivalent to

every

x, (px implies

[13-03]
[13-01] x n
e

=
[z(<pz)

z(^z)]
<px

= Hx

[xez(<pz)

a-ez(^z)].
.

z((pz)
e

==

n,
e

and x n
z(^z)]

z(^z)
(<px

Hence
"The

H x [x

=
z(<pz)

= Hx

= \l/x n = fa)^z"

class

determined by
"For

<pz

is

equivalent to the class determined by


is

is

equivalent to

every

x,

<px

equivalent to

$x".

[10-66]

Il x [x e

z(<pz)

Hence
Hence
13-9

II x (a: e

ex
e

e /3)

ex e z(tz)] = n x [-(.i- e

==

E x (-[.r
c-(x
e
e

z(^z)] c-[.r

z(^z)]).

/3)

a)].
.T e

[13-01-06]

-(a;

a)

a; e

-a, and -(x

0)

-/3.

[13-02]

(c/3) - (-/3c-).

[(aC|8) x(/3c T )]
[13-6] (a c
j8)

c(aC7).

= Ux
c

(<px

c ^ r),
t

(|8

7)

= H x (tx c

^.i-),

and (a c

7)

And
The

[10-65] [U x
"is

(<f>x

\f/x)

x
is

n x (^.r

.r)]

n x (^a;
is

c ^).
the
first

relation

contained

in"

transitive.
is:

13-9

form

of the

syllogism in Barbara.

The second form


c j8).

13-91

[(aC]8) x(.Tnea)] c(ar B


[13-6] (aC]8)

= n x (^c^.r).
<px

[13-01] (x n

a)

==

n,

and

(x n

0)

^
.

n.

And
If
is

[10-611] [n x (^.r c ^r) x


is

^.r n ] c\j/x n

the class a
a

contained in the class


/3.

&

and

.r n

is

member

of a, then

.r n

member
[(a

of

13-92

]8)

x(/3
\<2

==

7)]

c (a

=:

7)~

[lo-/J

Py

-iix^cp.c

/>

\fj

And
The

[10-68] [H x (^,r

^.r)

xH x (^.r =

^.r)]

cH x (^ =

^r)-

last three

theorems

illustrate particularly well the direct connection

266

Survey of Symbolic Logic

between formal implications and the relations of classes. 13-6 and 13-7 = 0. Similar alternative defini are alternative definitions of a c /3 and a
tions of the other relations

would be

l5
:

-[z(<pz*)]

z(-<pz)

z(<pz)

We may

give one theorem especially to exemplify the

way
by

in

which

every proposition of the

Two-Valued Algebra,

since

it

gives,

10-23, a

formal implication or equivalence, gives a corresponding proposition con cerning classes. We choose for this example the Law of Absorption.

13-92

[a+(aX]8)]

a.

[13-04-05] [a+(aX|8)]

= x{(xea)
x/3)]}

+ [(xea) x(.re/3)]}.

Hence

[13-01]

{x n e

[a + (a

=
But
But
[13-03]

{(x n

ea) +

[(x n

ea)x(x n
0)]}
==

e(3)}}.

(1)

{[+(
=

x/3)]

=
e

a}
a) + [(x
e

n,[{(.T

a) X (x
,

(x

)J.

(2)

[13-01]

(x n e

a)

=
<px

n,

(x n e 0)

= $x n and by
+
(<px

(2),

{[a+(a*{3)]

=
+

a}

But

[5-4]

[<pX

+(<pX n X\l/X n )]

= Ux =
xifrx)]

{[<px

xjx)]

=
<px}

V%n>

Hence

[10-23]

TL x

=
<px}.

{[<px

(<?x

All but the last

two

lines of this proof are


/3)]

the connection between [a + (a x

= a and

concerned with establishing the formal equivalence


<px]

n a {[#c+(#ex#r)] =
Once
this connection
is

made, we take that theorem


[a + (a x
if/x n

of the

Two- Valued

Algebra which corresponds to substitute in it n for p and


<px

/3)]

namely

5-4, (p + .p q)

p,

for

g,

and then

generalize,

by 10-23,

to

the formal equivalence which gives the proof.

An

exactly similar pro

cedure will give, for most theorems of the Two-Valued Algebra, a corre sponding theorem of the calculus of classes. The exceptions are such
propositions as

p =
its

(p

or an equivalence.

which unite an element p with an implication In other words, every theorem concerning classes can
1),

be derived from

analogue in the Two- Valued Algebra.

We may
15

conclude our discussion of the derivation of the logic of classes

As a fact, these definitions would be much more convenient for us, but we have chosen to give them in a form exactly analogous to the corresponding definitions of Principia (see
i,

p. 217).

Systems Based on Material Implication

267
set of postulates
will

from the

logic of prepositional functions

by deriving the
II.

for the Boole-Schroder Algebra given in

prove that, beginning with the Two- Valued Algebra, as a calculus of propositions, the This procedure may have the appear calculus of classes may be derived. ance of circularity, since in Section I of this chapter we presumed the
propositions of the Boole-Schroder Algebra without repeating them.

Chapter

This

But

the circularity

is

apparent only, since the Two- Valued Algebra

is

a distinct

system.

The

postulates of Chapter II, in a form consonant with our present

notation, can be proved so far as these postulates express symbolic laws.

The postulates of the existence, in the system, of -a when a exists, of a x when a and exist, and of the class 0, must be supposed satisfied by the fact that w e have exhibited, in their definitions, the logical functions which
r

determine a x
14-2
(a Xa)

(3,

-a, and
a.
e

O.

16

[13-01] x n

=
<fX

n.

Hence
Hence

[13-04] x n
[13-03]

e(a*a) =
a]

[(x n

a) x (x n

a)]

[(aXa) =
x

= H x {[.re

(a x a)]

= = x = Ht

(<px

^.r n ).

a]

[(<px

==
<px)

<&].

But

[1-2]

(<f>x

=
<pxn)
<px

n.

Hence
14-3

[10-23]
(/3

Ut
a).

[(<px

==
<px)

<px].

(ax0) =
[13-03]

[(ax0)

= U x {[x e (a x 0)] = = II x {[(.r6a)x(.r60)] = [(.r e /3) x(.r =


(0 x a)]

[x e (ft

x a)]}

a)]}.

[13-01-04]

Hence

= (0 x a)] = nj(#rx#r) = [13-01] [(ax/3) = (#r x ^r n ). But [1-3] n xtx n = (^ x ^.r)]. Hence [10-23] U x x^.r)
(<px

(*x x

<px)}.

[(<f>x

14-4

(aX0) X T
[13-03]

==

ax(0x 7 ). [(X0)X 7 =
==

X(0X T )]
x
(.r

= n x [{.T6[(aX/3)x 7 ]l = n x [{[(.T e a) x (x e (3)]


Hence
[13-01]

{a?e[ax(0x 7 )]}J
e

7)

==
J {

(.r e

a) x

[(.r

0)

(.r e

7 )] }J.

[13-01-04]

[(x0)

x 7

- ax(0x 7 )]

is possible when the satisfactory derivation of these existence postulates functions is treated in greater detail. See Principia, I, pp. 217-18. theory of propositional
16

more

268

A
But
[1-04]
(<px

Survey of Symbolic Logic

xtx n )
{[(<px

xx

=
<px

n x(\f/x n

Hence
14-5

[10-23]
0.

Hx

x^x) x &]

=
[<px

x^x n ). x(+x x &)]}.

axO =

[13-1-01] .TncO

= (x n eO)} = -(fzcMOj. n.u^x-o-zc^)] But [2-2, 9-01] (r.rcfar B ) = 1, and [3-2] -({x n c{x n ) = 0. = = -(fr cc). Hence [1-5] n x-(x n c x n )] Hence [10-23] n s {[^r x-frr c^)] = -(fa; cfr)}.
[13-03-04]
eQ)]
=
[<px

= -(frcr B ). [axO = 0] = U x l[(x n ea)x(x n

[13-01]

0, in

the fourth and fifth lines of the above proof,


of the calculus of classes.

is

the

of the

Two-Valued

Algebra, not the


these proofs will

Since the general

method

of

now

be clear, the remaining demonstrations can be some

what abbreviated.
14-61

[(x-|8)

0]

c[( a x/3)

a].
is

[13 -01 -02 -04 -06]

The theorem

equivalent to

But

[13-3]

Il x

-(xeO), and hence [9-1]


is

IL x [(x e 0)

0].

Hence the theorem


IT*
{

equivalent to
(xeQ)]c[(<pxxifrx)

[Or x -i/a*) =
is

But

[13-3]

H x -(xeQ), and
Ux
{[(<px

hence [9-1]

TL x [(x e 0)

= =

<px]}

0].

Hence the theorem

equivalent to

x-fr)

=
0]

0]

But

[1-61]

[(<f>X

n X-\j/Xn)

C[(<px

c[(^ x^) = = n X$X n )


=
0).

<px]}

<pX

n }.

Hence
14-62

[10-23]

Q.E.D.

{[(Xj8)

a]

x[(x-/3) =
is

a]}

c (a

The theorem
But

equivalent to
==

n,[{[(#rxr)
[13-3, 9-1]

^]x[(^x-^) =
0)

<px]}

c[<px

(x

0)]]

n x [(.r
is ==

0].

Hence the theorem

equivalent

to,

n,[{[(#cx^.T)

^i ]x[(^x-^.r)
==
<px

^r]}
==

c(<px

0)]

But

[1-62]

{[(?.rxr n )
Q.E.D.

n]

x[(^x n x-^x n }

<px

n }}

c(^.r n

0).

Hence

[10-23]

The

definition, 1

-0, follows readily

from the definition given

of

in

this section.

The other two

definitions of

Chapter II are derived as follows:

14-8

O+/3) - -(- ax -j8).

The theorem

is

equivalent to

Il x [(<px

\f/x)

= -(-$x

x-\f/x)].

Systems Based on Material Implication

269

But

[1-8]

(<px

+tXn) =
Q.E.D.

-(-<?x

x-\l/x n ).

Hence
14-9
(aCjS)

[10-23]

[(ax/3)
is

a].
V/.r)

The theorem
But
[1-9]
(<f>x

equivalent to IIjOz c
is

= Uz
tt )

[(<px

x$x)

==
<px].

c^x n )

equivalent to

[(^cx^ic

==

v*c n ].

Hence [10-23-69] Q.E.D.


Since the postulates and definitions of the calculus of classes can be deduced from the theorems of the calculus of prepositional functions, it
follows that the whole system of the logic of classes can be so deduced. differences between the calculus of classes so derived and The

important

the Boole-Schroder Algebra, as a logic of classes, are two: (1) The BooleSchroder Algebra lacks the e-relation, and is thus defective in application,
since
it it

which

cannot distinguish the relation of an individual to the class of of which is con is a member from the relation of two classes one
(2)

tained in the other;

The theorems

of the

Boole-Schroder Algebra

of the cannot validly be given the completely symbolic form, while those functions can calculus of classes derived from the calculus of prepositional
17 be given this form.

V.

THE
is

LOGIC OF RELATIONS

The
tions of

logic of relations

two or more

variables, just

derived from the theory of propositional func as the logic of classes may be based

of one variable. upon the theory of propositional functions in extension when we logically exhibit the A relation, R, is determined
If to y. y) class of all the couples (x, y) such that x has the relation then x $[<p(x, y)} is the relation "parent of of is
<p(x,
"

represent

"x

parent

y",

is This defines the relation in extension: just as the extension of so the the class of all those things which have the property of being red,
"red"

extension of the relation

"parent

of"

is

the class of

all

couples in the universe.

a property that is the extension of the relation is, couples (or triads, etc.) of a certain class; The calculus of relations, like the calculus thus, the class of couples itself.
relation
is

the parent-child common to all the

of propositions,
"Oftentimes,

and

of classes,

is

a calculus of extensions.
i,

as in Schroder, Alg. Log.,

the propositional of classes have been represented in the symbols of the Book-Schroder Algebra. calculus has been treated otherwise than as an interpretation of to the system, tl in such a case, if these symbols are regarded as belonging But
is

the relations of propositions in the algebra calculus befo

invalid.

270

A We

Survey of Symbolic Logic


the relation

assume, then, the idea of relation

R meaning
etc.

the class

of couples (x, y) such that x has the relation

R to

y.

R =
This notation
is

x y(x Ry),

S = w z(w S

z),

simpler

S = w

z[\[/(iv, z)],

but

it

and more suggestive than means exactly the same thing.

R = x A triadic

y[<p(x,

y)],

relation,

T, will be

such that

T = x
or

y z[T(x y
}

z)}

is

the class of triads


is

(x, y, z)

for

which the prepositional function

T(x, y, z)

true.

But

all

relations can be defined as dyadic relations.

triadic relation can be interpreted as a relation of a

dyad to an individual
y, z),

that

is

to say,

any function of three variables, T(x,

can be treated as a
.

two variables, the couple (x, y) and z, or x and the couple (y, z) This follows from the considerations presented in concluding discussion of the theorems numbered 11-, in section III. 18 Similarly, a tetradic relation
function of

can be treated as a dyadic relation of dyads, and so on.


of dyadic relations
is

Hence the theory

a perfectly general theory.

Definitions exactly analogous to those for classes can be given.

15 -01

(x, y) n e z

w[R(z,

ic)]

R(x, y) n

Def.

It is exactly at this point that

our theoretical considerations of the equiva

lence of

<p(x,

y) n

and

<p(x

y a ) becomes important.

For

this allows us to

treat R(x, y), or (x

Ry), and by 11-07, we can give our


15-01
"The

as a function of one or of

two

variables, at will;

definition the alternative form:

(xm

y n )ezw(zRw)

xm

R yn R yn

Def.
or extension, of the relation deter
true.

couple (x m y n ) belongs to the


(z

field,

mined by
15-02

R w)
is

"

means that x m

is

RcS = U

x, v

[(xRy) c(xSy)].

Def.

This definition

strictly parallel to 13-02,

(ac/3)
because, by 15-01, (x

= Hx
y)
e

(xe

acxe
(x

/3)

R y)
,

is (x,

and

y) is (x, y)

S.

similar

remark applies to the remaining


15-03

definitions.
y)].

(R

S)

= nx

y [(x

Ry) = (xS

Def.

and

and S are equivalent in extension when, (x S y) are equivalent assertions.


18

for every

x and every

y, (x

R y)

See above, pp. 253

ff.

Systems Based on Material Implication

271

15-04

R xS =

x y

[(x

Ry)* (x

y)].

Def.

product of two relations, R and S, is the class of couples (x, y. such that x has the relation R to y and x has the relation S to y. If R is

The

logical

"friend

of",

and S

is "colleague

of",

R x S will be
Def.
S,
is

"friend

and colleague

of")

15-05

R + S = xy[(xRy)
sum
of
y

+ (xSy)].

The
will

logical

tw o relations,

R and

the class of couples

(x, y)

such

that either x has the relation

to y or x has the relation

S to

y.

R+S

be

"

Either

fl of or

of".

15-06

-R = xy-(xRy).
the relation of x to y

Def.

-R

is

important to + S)y, and # -E y are significant assertions. z(#


It is
}

when x does not have the relation R to y. note that R x S, R + S and -# are relations: x(R x *$)#,
and the
"null-relation"

The

"universal-relation"

are also definable

after the analogy to classes.

15-07

x y

[?(x,

y)c{(x,

y)].

Def.
is

x has the universal-relation to y in case there


$(x, y)

a function,

f,

such that

c(x,

y),

i.

e.,

in case x

and y have any

relation.

15-08

-1.

Def.
for relations different

Of course, 0,1, + and x have different meanings

from their

meanings for classes or for propositions. 0, + etc., are strictly analogous.


,

But these

meanings ot

As was pointed out

in Section III of this chapter, for

every theorem

involving functions of one variable, there is functions of two variables, due to the fact that a function

a similar theorem involving


<p(x

y)

may

be
for

regarded as a function of the single variable (x, y). Consequently, each theorem of the calculus of classes, there is an exactly corresponding

theorem

in the calculus of relations.

We

may, then,

cite as illustrations of

this calculus the analogues of


classes;

the theorems demonstrated

to hold

for

and no proofs

will here

be necessary.

These proofs follow from the

theorems of Section
are given

III,

numbered

11-, exactly as the proofs for classes


in Section II,

by the corresponding theorems


y)].

numbered 10

15-1

= x$-[t(x,y)c(x,
null-relation
i.

The

is

the relation of x to y

when

it is

false that $(x, y) implies


is

(x, y), (x, y)


19

e.,

when

x has no relation to #. 19

Of course, there

no such

couple which can significantly be called a couple. As in the case of the 1 and of the class calculus, the 1 and

of relations, defined as

272
15-2

A
nx
,

Surrey of Symbolic Logic

y [(x,

y)

e 1].

Every couple

is

member

of the universe of couples, or has the universal

(dyadic) relation.

15-3

nx

-[(x, 2/)eO].

No

couple has the null-relation.

15-4
15-5

jRcl. Ocfl.
relation, R, is implied

Every

by the

null-relation

and implies the universal

relation; or, whatever couple (.r, y) has the null-relation has also the relation 72, and whatever couple has any relation, R, has also the universalrelation.

15-6

(RcS) =
S".

tt x

For

relations,

RcS
By

is

[(xRy)c(xSy)}. more naturally read


"

"

implies

S"

than

"

is

con

tained in
if

15-6,

x has the relation

R to

y,

means "For every x and every y, implies then x has the relation S to Or R implies
S"
"

y".

S"

means

"Every (x,

y) couple related
y [(x

by

are also related

by

S".

15-7

(R =

S)

= Ux

Ry) =

(x

y)].

Two

relations,

R
S,

and

S, are equivalent

when the

couples related by

JR.

are

also related

by

and

vice versa (remembering that

is

always a reciprocal

relation

c).

15-8

(RcS) = (-Sc-R).

If the relation

implies the relation S, then

when S

is

absent

also will

be absent.

15-9

[(RcS) *(ScT)]c(RcT).
implication of one relation

The

by another

is

a transitive relation.

15-91
If

[(RcS)x(x m

Ry

)]c(x m

Sy n ).
this couple are

implies

related also

S and a given couple are related by R, then by S.


S)

15-92

[(R

x(S

==

T)]c(R =
is

T).

The equivalence
If it

of relations

transitive.

be supposed that the postulates concerning the existence of rela tions are satisfied by exhibiting the functions which determine them, then,

we have
nificant,

defined them, are such that propositions involving them are true whenever sig and significant whenever the prepositional functions determining the functions in question have the same range.

Systems Based on Material Implication


as in the case of classes,

273

we can

derive the postulates (or remaining postu

lates) for a calculus of relations

sitional functions.

from the theorems of the calculus of prepo The demonstrations would be simply the analogues of

those already given for classes, and

may

be omitted.

16-2
16-3

(RxR) = (RxS) =

R.
(S xR).

16-4
16-5

(RxS) xT = Rx(SxT). RxO = 0.


[(Rx-S) = 0}c[(RxS) =
R] H(RxS) = -(-Rx-S). (R + S) (RcS) = [(RxS) = R].
----

16-61

R].

16-62 16-8 16-9

x[(Rx-S) =

R}}

c(R =

0).

These theorems

may

also be taken as confirmation of the fact that

the Boole-Schroder Algebra holds for relations.


relations"

In

fact,

"calculus

of

most frequently means


a, b, c,

just that

the Boole-Schroder Algebra

with the elements,

etc., interpreted as relations taken in extension.


is

So

far,

the logic of relations


are

a simple analogue of the logic of classes.

But there
analogies,

many

properties of relations for

which

classes present no

h. Fact, and these peculiar properties are mosl important, of mathematics, worked out by Peirce, Schroder, the logistic development Peano and his collaborators, and Whitehead and Russell, has de

Frege,

While pended very largely upon a further study of the logic of relations. manner we can do no more, within reasonable limits, than to suggest the of this development, it seems best that the most important of these proper
ties of relations

should be given in outline.

But even

this outline

cannot

be complete, because the theoretical basis provided by our previous dis


cussion
is

not sufficient for completeness.


relation, R, has a converse,

Every
17-01
If

*R, which can be defined as follows:

*R =

y x (x Rij).

Def.

x has the relation

to y, then y has the converse relation,


e

R, to

x.

It follows at

once from the definition of (x m y n )

that

xm
because
(x m

R yn = R

yn

*R x m
e
"R

Ry

n)

(x m y n )

R =

(y n x m )
is

Vn

"R

xm

>

The converse

of the converse of

R.

w(wfl)

= R

19

274
since

A
=
(*>R)

Survey of Symbolic Logic

x y (x R y) proof would require that we demonstrate


x y (y

R x) =

R.

(This

is

not

proof:

n,

[(*, y) e

~(~R)

(x, y) e

R]

But

it

is

obvious that such a demonstration

may

be given.

In general,

we shall not pause for proofs here, but merely indicate the method of proof.) The properties of symmetrical relations follow from the theorems con
cerning converses.
versal relation,
(x 1 y)
1,

The uni For any symmetrical relation T, T = and the null-relation, 0, are both symmetrical:
*->T.

[rO, y) cf(s,

y)]

==

==

[f(y, x)

c{(y,

.r)]

(y I x)

(The

"1"

in the

middle of this

proof

is

obviously that of the calculus of

propositions.
(

Similarly for

in the next.)

y)

-[r(.r, y)
if

$(*, y)]

-[f (y, x)

{(y, x)]

(y

a-)

It is

obvious that

two

relations are equivalent, their converses will be

equivalent:

(#

(R

$)
in

Not

quite so obvious

is

the equivalent of

(RcS),

terms of
Instead

*R and

^S.

We

might expect that (R c S) would give (/S

c^).

we have

(RcS) = (*R c
for
(fl

S)
y [(y
"R

,S)

= H Xf

y [(.i-

E y)

(x

y)]

= Ux
is

x)

(y

^S

x)]

= (vRc^S)
"

Parent of implies
of".

ancestor

of"

equivalent to

"Child

of

implies

descendant

The converses

of

compound

relations

is

as follows:

(/JxS)
for

= vRx^S
(x

(R x S)y

= y(R

x S)x

(y

R x) x(yS x) =

y ) x (x ^S y)

=
If

x(*Rx"S)y

is

employer and
by".

exploiter of y, the relation of y to

.r

is

"employee

of

and exploited

Similarly

If

is

either

employer or benefactor of

y,

the relation of y to x

is "either

by". employee Other important properties

of or benefitted

of relations

concern

"relative

sums"

and

Systems Based on Material Implication


"relative

275

These must be distinguished from the non-relative sum and product of relations, symbolized by + and x The non-relative of "friend and "colleague is "friend and product colleague
products".
.

of"

of"

of":

their relative product

is

"friend

of a colleague
of":

of".

Their non-relative

sum
by

is "either

friend of or colleague
of".

their relative

sum

is "friend

of

every non-colleague

We

shall

denote the relative product of

and S

S, their relative

sum by

R t S.
Def.
relation
z

17-02

S = xz{2 y [(xRy) x(ySz)]}.

R\S

to y

for

(x, z) when for some y, x has the and y has the relation S to z. x is friend of a colleague of some y, x is friend of y and y is colleague of z.
is

the relation of the couple

when,

17-03

R-tS = xz{n y [(xRy)


relation of x to z

+ (ySz)}}.

Def.

R t S is the R to y or y
It is

has the relation S to


y, either

when, for every y, either x has the relation z. x is friend of all non-colleagues of z
is

when, for every

is

friend of y or y

colleague of

z.

noteworthy that neither relative products nor relative sums are commutative. "Friend of a colleague is not "colleague of a friend
of"

of".

Nor

is "friend
of".

of all non-colleagues

of"

the same as

"colleague

of all non-

friends

But both

relations are associative.

R\(S\T) = (R\S)\T
for

2 t {(wR

x)

x X (S T)z]
[

2 X (w R x) x z v [( x Sy)x(yTz)]} = 2 y 2 x {(wRx)x[(xSy)x(yTz)]}
--=
{

= 2 v 2,{[(wRx)x(xSy)]x(yTz)} = 2 v (2 x [(wRx)x(xSz)]x(yTz)} = 2 v {[w(RS)y]x(yTz)}


"Friend

of a (colleague of a neighbor
of".

of)"

is "(friend

of a colleague) of a

neighbor

Similarly,
"Friend

R
all

t (S t T)

= (R

t S) t

T
of)"

of all (non-colleagues of all non-neighbors

is

"(friend

of all

non-colleagues) of

non-neighbors

of".

De Morgan
ucts.

Theorem holds

for the negation of relative

sums and prod

-(R\S)
for

= -flt-S
--=
}

-{

2 t [(x

Ry)*(yS z)]

tf

-[(*

R y)

x (y S

z)}

276

A
of
"friend

Survey of Symbolic Logic


is "non-friend

The negative
Similarly,

of a colleague

of"

of all colleagues

(non-non-colleagues)

of".

-(R t S) = -R -S
of
"friend

The negative
colleague
of".

of all non-colleagues

of"

is "non-friend

of a

non-

Converses of relative sums and products are as follows:

for

x *(R S)z
|

= z(R

S)x

= ? y [(z Ry)*(yS x)] = 2 v ((y S x) x(zR y)]

If

is

fitted

employer of a benefactor of by an employee


of".

z,

then the relation of

z to

"

is

bene-

Similarly,
If

^(R t
all

/S)

= ^S

^R
a;

is

hater of

non-helpers of
by".

z,

the relation of z to

is "helped

by

all

who are not hated The relation of


relative addition.

relative product

is

distributive with reference to non-

for

x[R (S + T)]z
\

R\(S+T) = (RS) + (R T) = ? y {(x Ry)x [y(S + T)z]


}

X ^^2/

f 1

(T v**

7? tl/

ii\ c//

(11 L\c/
F

Si

?^
/

+ +

T?y
\"

y 2^1 /J

1
)

= 2 y {[(x Ry)*(yS
Similarly,
"Either

z)]

[(x

Ry)x(y T z)}}

(R + S) T = (R T) + (S T)
\ \
\

friend or colleague of a teacher

of"

is

the same as

"either

friend

of a teacher of or colleague of a teacher

of".

somewhat curious formula

is

the following:

It holds since

x[R (S x T)]z
\

R\(S*T)c(R\S)x(R\T) = ? y {(x Ry)x [y(S x


=

T)z]

Z y {(xRy)x[(ySz)x(yTz)}}

and

since a x (b xc)

(a

x6) x (a xc),
=

z y {[(x
2,[(.T

Ry)*(yS z)] x [(x Ry)x(yT z)] Ry)*(yS z)] x ? y [(x Ry)*(yT z)]

And

this last expression

is

[x(R\S)z] x[x(R\ T)z].

Systems Based on Material Implication


If

277

is

student of a friend and colleague of


z.

z,

then x

is

student of a friend and

student of a colleague of
cause
"student

The converse
colleague"

of a friend

and

implication does not hold, be requires that the friend and the

friend and student of a col colleague be identical, while "student of a where S y is repeated, last step in the proof does not. (Note the league"
,

and observe that


Similarly,

this step carries exactly that significance.)

(flxS) T c (R\ T) x (S T)
\

The corresponding formulae with t


and seldom
useful; they are omitted.

instead of

are

more complicated

is

no particular importance, but the relative product In terms of this idea, "powers" of a relation are a very useful concept.

The

relative

sum

is

of

definable

# = R
2

JR,

# = R
3

2
|

R,

etc.

transitive relation, S,

S n cS.

The

predecessors

S 2 c 8, and hence distinguished by the fact that of predecessors ... of x are predecessors of predecessors of the powers of a relation plays a This of x.
is

conception

prominent
ties of

part in the analysis of serial order,

and

of the

fundamental proper

the
of

number
20

series.

By

use of this and certain other concepts, the

method

"mathematical

induction"

can be demonstrated to be com

pletely deductive.

were not given were letters by which relations were symbolized separate treatment. The Any also interpreted as relative terms by a sort of systematic ambiguity. class of entities which have that relation also stood for the
In the work of

De Morgan and

Peirce,

"relative terms"

relation

symbol

to something.

But

in the logistic

development

of

mathematics, since that

time, notably in Principia

Mathematical

relative

terms are given the

The "domain" of a given separate treatment which they -really require. relation R to some that is, the class of entities which have the relation, R be defined as follows: fl, which can thing or other-may be symbolized by

17-04

DR =
of

&[2y(xRy)].

Def.
"For

The domain
x has the

R is the class of x s determined by the function D R will If R be "employer of relation R to


",

some

y,

be the class

y".

of employers.

The
20
21

"converse
i,

domain"

of fl

that

is,

the class of things

See Principia,
See
i,

Bk. n, Sect. E.

*33.

The notation we

use for domains and converse domains

cipia.

278

Survey of Symbolic Logic

which something or other has the relation and similarly defined:


17-05
(I

R may

be symbolized by

CL

R =
x,

y[2,(xRy)].

Def.
is

The
will

converse domain of

the class of y

"For

some

a-

has the relation

to

y".

determined by the function dR If R be "employer


of",

be the class of employees.

The domain and converse domain


the
"field "of

of a relation, R, together constitute

R,

R. + (yRx)]}.
all

17-06

C R = xi2 y [(xRy)

The

field of

will
If

be the class of be
"employer

terms which stand in either place in

the relation.

of",

CR

is

the class of

all

those

who

are

either employers or employees.

The elementary

properties of such

"relative terms"

are

all

obvious:

xn

eC R = 2 y [(x n Ry) C R = D R+a R

(yRx n )}

However,
are of
"

development of mathematics, these properties 22 the highest importance. We quote from Prirwipia Mathematical
for the logistic

Let us ... suppose that

is

the sort of relation that generates a series,

say the relation of less to greater among integers. Then that are less than some other integer = all integers, Q
that are greater than some other integer
case,

DR =
R =

all

integers integers

all

all

integers except

0.

In this

C R = all integers that are either greater or less than some other = all integers .... Thus when R generates a series, C R becomes integer
important.
..."

have now surveyed the most fundamental and important characters of the logic of relations, and we could not well proceed further without
elaboration of a kind which
is

We

here inadmissible.

But the reader

is

warned

that

we have no more than

scratched the surface of this important topic.


"What

About

1890, Schroder could write

a pity!
it".

developed instrument and nothing to do with

To have a highly And he proceeded to

make

beginning

in the bettering of this situation

of relatives to the logistic

development
in the

of certain

by applying the logic portions of Dedekind s


of

theory

of

number.

Since that time, the significance of symbolic logic has

been completely demonstrated


22
1,

development

Peano

Formulaire

p. 261.

Systems Based on Material Implication

279

and

of Principia Mathematica.
is

And the very head and front of this develop

ment

a theory of relations far

previously given.

We

more extended and complete than any can here adapt the prophetic words which Leibniz
"I

puts into the

begin to get a very different opinion I had regarded it as a scholar s of logic from that which I formerly had. in the way you understand it, it is a kind of diversion, but I now see that,
of Philalethes:

mouth

universal mathematics."

VI.

THE LOGIC OF

Principia Mathematica
of the Boole-Schroder Algebra

We

have now presented the extensions

the Two-Valued Algebra, prepositional functions and the propositions derived from them, and the application to these of the laws of the Two-

Valued Algebra, giving the calculus of propositional functions. Beyond with the Twothis, we have shown in outline how it is possible, beginning Valued Algebra as a calculus of propositions, to derive the logic of classes in a form somewhat more satisfactory than the Boole-Schroder Algebra, In so doing, we have presented of relations and relative terms. and the
logic

as

much
work

of that
of

the

being followed the exact forms which that development took


considerably modified
it

But, our purpose

development which begins with Boole and passes through Peirce to Schroder as is likely to be permanently significant. we have not here expository rather than historical,
Instead,

we have

in the light of

since the publication of the

work

of

what symbolic logicians have learned Peirce and Schroder.


in detail our divergence

Those who are interested to note


historical

from the

development
of

will

and VIII

Chapter
II

I.

be able to do so by reference to Sections VII But it seems best here to point out briefly what

these alterations are that


interpreted S^.r,
^r,

we have made.

In the

first

place,

we have

S^(.r, y), etc.,


,

propositions of the form ^r n iK-r m y n ), this, in consideration of the serious


<px

sums or products of Peirce and Schroder avoided etc. But while theoretical difficulties.
explicitly as

still as an actual product, S#c as an actual sum, they did not treat II which result the laws which they give for propositions of this type are those method by from such a treatment. There is no slightest doubt that the and formulated these laws is substantially the one

which Peirce discovered

which we have exhibited.

symbol for a a sum, makes demonstration possible where product, S#r as the symbol of made and, for further otherwise a large number of assumptions must be resorted to. a much more difficult and less obvious style of proof

And

this explicit use of n^.r as the

principles,

280

Survey of Symbolic Logic

In this part of their work, Peirce and Schroder can hardly be said to have formulated the assumptions or given the proofs.
In the second place, the Boole-Schroder Algebra
of

the general outline

which

is

already present in Peirce s

work

probably seemed to Peirce


(though there are indications

and Schroder an adequate calculus

of classes

With this system before in the paper of 1880 that Peirce felt its defects). them, they neglected the possibility of a better procedure, by beginning
with the calculus of propositions and deriving the logic of classes from the
laws which govern propositional functions. And although the principles which they formulate for propositional functions are as applicable to func
tions of one as of
their interest
relatives

two

variables,

and are given

for

one as well as for two,


of

was almost

entirely in functions of

two and the calculus

which

may

be derived from such functions.


is,

The

logic of classes

which we have outlined


for,

then, something which they

laid the

foundation

but did not develop.


exposition thus far in the chapter have been
relation of this earlier treatment of symbolic

The main purposes of our two: first, to make clear the

logic with the later and better treatment to be discussed in this section;

and second, to present the logic of propositional functions and their deriva tives in a form somewhat simpler and more easily intelligible than it might
otherwise be.

The

theoretically sounder

and more adequate

logic of Prin-

cipia Mathematica

is

given a form which


is

so far as propositional functions


its

and

their derivatives

concerned

seems to us to obscure, by

notation,

the obvious and helpful mathematical analogies, and requires a style of


proof which
is

much

disclaim any idea

With regard to this second purpose, we that the development we have given is theoretically
less

obvious.

adequate

its

chief value should be that of


difficult

tory to the

more complex and

an introductory study, prepara treatment which obviates the the

oretical shortcomings.

Incidentally, the exposition

which has been given

will serve to indicate

how much we

are indebted, for the recent development of our subject, to

the earlier work of Peirce and Schroder.

The Peirce-Schroder symbolic


Peano
s

logic

is

closely related to the logic of

connection

Formulaire de Mathematiques and of Principia Mathematica. This is easily overlooked by the student, with the result that the sub
the Boole-Schroder Algebra and
its

ject of his first studies


likely to
logistic

applications

is

seem quite unrelated to the topic which later interests him the development of mathematics. Both the connections of these two

Systems Based on Material Implication

281

and

their differences are important.

We

shall

attempt to point out both.

And

because, for one reason, clearness requires that

we

stick to a single

illustration,

our comparison

will

be between the content of preceding


logic of

sections of this chapter

and the mathematical

Book

I,

Principia

Mathematical

The Two- Valued Algebra

is

a calculus produced by adding to and re

interpreting an algebra intended primarily to deal with the relations of And it has several defects which reflect this origin. In the first classes.
place, the

same

logical relation is expressed, in this system, in


"If

two
q

different

ways.
per",

We

have, for example, the proposition


p, q,

peg
. .

and
.
,

r,

then
..."

where

and
"

are propositions.
relation
"

But

"if

then
q,

is

supposed to be the
Also,

same
in

which

is
r

expressed by c in p c
is

r,

and per.
system
"either

"and

pc

q and q c

"

the relation which

is

other

wise expressed

by

and
use

so on, for the other logical


of
"if
.
. .

relations.

The
. .
.",

involves
.

the
. .
.",

then
to

.
",

"...

and

or
.
.

"...

is

equivalent

.
.",

and

"...

is

not

equivalent to

.
",

just as
,

exactly the relations c

any mathematical system may; yet these are + = and =H whose properties are supposed to
,
,

be investigated in the system.

Thus

the

system takes the

laics of the logical

relations of propositions for granted in order to prove them.

Xor

is

this

paradox removed by the fact then ..." and c, ability of


"if
.
.

that we can demonstrate the


of
".

interchange-

and

..."

and x,

etc.

For the very demonstration of this interchangeability takes for granted the logic of propositions; and furthermore, in the system as developed,
it is

form impossible in most cases to give a law the completely symbolic until it has first been proved in the form which involves the non-symbolic
So that there
demonstration of the laws of

expression of relations.
in the

which the circularity propositions can be removed in this


is

no way

in

system.

Another defect

of the
"p

Two- Valued Algebra


is
true"

is

the redundance of forms.


p,
s

The
23

proposition p or

is

symbolized by
method
of

by p =

1,

by p
is

4=

Logically, as well as historically, the

Peano

Formulaire

a sort of

The general intermediary between the Peirce-Schroder mode of procedure and Principia. method of analysis and much of the notation follows that of the Formulaire. But the Formulaire is somewhat less concerned with the extreme of logical rigor, and somewhat more concerned with the detail of the various branches of mathematics. Perhaps for this fundamentals which is the dis reason, it lacks that detailed examination and analysis of
tinguishing characteristic of Principia. of the relation D (in our notation, c
):

For example, the Formulaire retains the ambiguity p?q may be either the class p is contained in
q".

the class
laire

or "the proposition p implies the proposition contains no specific theory of propositions.


q",

In consequence, the Formu

282
the negation of

A
p
or

Survey of Symbolic Logic


is
false"

etc.,

"p

by

-p,

p =

0,

-p =

1,

=}=

1,

etc.

These various forms may, it is true, be reduced in number; p and -p may be made to do service for all their various equivalents. But these equivalents
cannot be banished, for in the proofs
that
it is

necessary to
0)

make
is

use of the fact

p =

(p

1)

(p 4= 0),

-p =

(p

(-p

1), etc., in

order to

demonstrate the theorems.


avoidable.

Hence

this

redundance

not

altogether

calculus of propositions in Principia


etc.,

Both these defects are removed by the procedure adopted for the Mathematical Here p = 1, p = 0,
are not used;

instead

we have simply p and


it

its

negative, symbolized

by ~p. may every mathematical system has always taken


And, impossible as
ideas
are:

seem, the logic of propositions which


for granted
is

not presumed.
(2)

The

primitive

(1)

elementary propositions,

elementary
the

propositional functions, (3) assertion, (4) assertion of a prepositional func


tion,
(5)

negation, (6) disjunction, or the logical sum;

and

finally,

idea of

but

is

which does not belong in the system a notation to indicate that one symbol or complex of symbols merely
"equivalent

by

definition",

be replaced by another. An elementary proposition is one which does not involve any variables, and an elementary propositional function is such as "not-p" where p is an undetermined elementary proposition.

may

The

what would be supposed a proposition may be asserted or merely considered. The sign h prefaces all propositions
idea of assertion
is

just

which are asserted.


where

An
q.

asserted propositional function

is

such as

"A

is

"

is

undetermined.

p v q, instead of p + p and q, is true".

pv

The disjunction of p and q is symbolized by At least one of the two propositions, q means
"

The
#1-01
"p

postulates and definitions are as follows:

poq. =
(

~pvq.
q"

Df.
is
is is

(materially) implies

the defined equivalent of


true
,

"At

least

one of

the two,

is

false

and

is

a true
is

proposition".

(The explana

tion of propositions here

ours.)
its

p^q

the same relation which

we

have symbolized by p c

q,

not

converse.

Principia.

(The propositions quoted will be given the number which they have in The asterisk which precedes the number will distinguish
in earlier chapters or earlier sections of this

them from our propositions


chapter.)

The
24

logical
i,

product of p and q

is

symbolized by p

q,

or p

q>

See Bk.

Sect. A.

Systems Based on Material Implication

283

*3-01
"p

p.q. =
true

~(~p

v~ry).
is

I)f.
"It

is

and

is

true"

the defined equivalent of


</,

is

false that at

least

This is false". one of the two, p and Morgan s Theorem in our notation, (p q) =

is,

of course, a

form of

De
.

-(-/>

-(/).

The
#4-01
"p

(material) equivalence of

p and

q
.

is

symbolized by p

q or

r/.

= ,p?q q^p.
.
q"

Df
is

is

(materially) equivalent to

the defined equivalent of


p".

"p

(ma

terially)

implies q
q)

and

g (materially) implies

In our notation, this


==

would be (p =

= (pcq)(qcp).

Note that

...

...

and

Df

are different relations in Principia.

The dots

in these definitions serve as punctuation in place of parentheses

and brackets.

takes precedence over one, as a bracket over a In #4-01 we have only one dot after =, parenthesis, three over two, etc. because the dot between p D q and q D p indicates a product a dot, or two
dots,
:,
:

Two

dots, indicating a product

is

always inferior to a stop indicated

by the

same number

of dots

but not indicating a product.

The

follows: postulates of the system in question are as

*!!
*1-11

Anything implied by a true elementary proposition


stands for
<px

is

true.

Pp.

("Pp."

"Primitive proposition".)

can be asserted, where x is a real variable, and <pxo\f/x can be asserted, where x is a real variable, then #r can be asserted, where x

When

is

a real variable.

Pp.

A
#1-2

"real

variable" is

such as p in -p.
Pp.
p.

\-:pvp.?.p.

In our notation, (p + p} c

*l-3

\-lq*-D.pvq.

Pp.
q).

In our notation, q c (p +

*l-4

\-mpvqmOmqvp.
q)

Pp.

In our notation, (p +

c(q +

p}.

*l-5

\-:p

v(q vr) .3.

v (p vr).
r)]

Pp.
r)].

In our notation, [p +

(q

+
.

[q

+ (p +
r.

#1-6

h!

^
<7

pv
c

pv
+

Pp.
r)].

In our notation,

(</

r)

[Q;

(/)

c Q; +

Note that the

the sign of assertion in each of

above

is

followed by a

284
sufficient

A
number

Survey of Symbolic Logic

of dots to indicate that the

whole of what follows

is

asserted.
$K-l-7

If

is

an elementary proposition, ~p

is

an elementary proposition.

PP.

*1-71

If

p and
Pp.
<pp

q are

elementary propositions,

pvq

is

an elementary

proposition.

#1-72

If

and

\I/p

are elementary prepositional functions

which take

elementary propositions as arguments,


tional function.

v
<pp

\f/p

is

an elementary proposi-

Pp.

list of assumptions. The last three have to do with the method by which the system is developed. By *l-7, directly any proposition which is assumed or proved for p may also be asserted to

This completes the

hold for ~p, that

is

any proposition
or q or
r,

of

~p may be substituted for p or q or r, etc., in the system. By #1-71, p v q may be substituted for p
to say,
72,
if

etc.

And by *1
"

any two complexes

of the foregoing

symbols
in the

which make sense as

statements"

can be treated in a certain

way
all

system, their disjunction can be similarly treated.


of these,
.

By
q,

the use of

three
q

any combination such as p v

q,

q,

p3

q,

i>

p,

~p v . p v q, ~p v ~ry, etc., etc., may be substituted for p or q or r in any assumed proposition or any theorem. Such substitution, for which no
postulates would ordinarily be stated,
is

one of the fundamental operations

by which the system


Another kind
for

is

developed.

of substitution of

which

is

fundamental

is

the substitution
exists.
".

any complex
is

symbols

of its defined equivalent,

where such

This operation

covered by the meaning assigned to "... = ... Df Only one other operation is used in the development of this calculus

elementary propositions the operation for which *!! and ^1-11 are assumed. If by such substitutions as have just been explained there
of

results a

complex

of

and

if

that part of

which the main, or asserted, relation is o the expression which precedes this sign is identical with a
symbols
in
,

postulate or previous theorem, then that part of the expression which


follows this sign

may
. . .

be asserted as a

lemma

or

new theorem.

In other

words, a main, or asserted, sign D has, by *!! and *1-11, the significant then This property is explicitly assumed property of
"If
,

.".

in the postulates.

The main thing


not so

to be noted

about

this operation of

inference
strictly

much a piece of reasoning as a mechanical, or mathematical, operation for which a rule has been given. No
is

that

it is

Systems Based on Material Implication


"mental"

285

involved except that required to recognize a previous proposition followed by the main implication sign, and to set off what
operation
is

follows that sign as a

then,

assertion. The use of this operation does not, that the processes and principles of ordinary logic are tacitly presupposed as warrant for the operations which give proof.

new

mean

What is the significance of this assumption of the obvious in # 1


*l-7,

1,

#1

1 1,

#1-71, and *l-72?


so

Precisely

this:

these
is

postulates explicitly

assume

much

of the logical operations as

necessary to develop the


is

system, and beyond

this the logic of propositions simply


it

not assumed.

To

illustrate this fact,

will

be well to consider carefully an exemplary

proof or two.

#2-01

\-ipD~p.o.~p
Dem.
Taut M-: ~p v ~p
01)]
.

~p

(1)

Vip?~p.?.~p
#1-2 above.
}

"Taut"

is

the abbreviation for the Principle of Tautology,

~plp indicates that

~p

is

substituted in this postulate for p

giving (1).

This operation is valid by #1-7. Then by the definition #1-01, above, p D ~/j is substituted for its defined equivalent, ~p v ~p, and the proof is
complete.

*2-05

\-mmqDrm3ipoqm3mp3r
Dem.

Sum [(1)
.

M-:

=>

~p v

~p v
.

(1)

O1-01)]

\-: .

qor .D: pDq .? p?r


And
(1) is

Here ~p
is

"Sum"

refers to

*l-6, above.

what

*!(>

becomes when

Then, by #1-01, p^q and pir are substituted for their defined equivalents, ~pvq and ~pvr, in (1), and the resulting
substituted for p.

expression

is

the theorem to be proved.


illustrates the use of
m"3l

The next proof

*1

and #1-11.

*2 06

\-l ."3

-zr

."3

Dr

Dem.
p,

Dr.D:pDg.D.pDr:.3:p3g
[#2 -05] hs
[(1)
.

.3Sg

DrD.p3r
por

(1) (2)

3r

D:p3
-

s. par

(2)

*1-11] h:

/>3r/.D:

q?r .0.

286
"

A
is

Survey of Symbolic Logic

"Comm

When,
it

in

*2-04, previously proved, which is p . D . q D r : D : q D . p D this theorem, q D r is substituted for p, p D q for q, and p D r for

r.

r,

becomes the long expression (1). Such substitutions are valid by *l-7, #1-71, and the definition ^1-01: if p is a proposition, ~p is a proposition; if ~p and q are propositions, ~p v q is a proposition; and p D is the defined If we replace equivalent of ~p v q. Thus p D q can be substituted for p.
ry

the dots by parentheses,


h
{

etc., (1)

becomes
r)]
}

(q

r)

[(p

D
</)

D
(/>

(p

ry)

[(ry

r)

D (p D

r)]

But, as (2) states, what here precedes the main implication sign is identical with a previous theorem, *2-05. Hence, by #1-11, what follows this
the theorem to be proved can be asserted. sign Further proofs would, naturally, be more complicated, but they involve no principle not exemplified in the above. These three operations sub
stitutions according to *l-7,

main implication

*1-71, and *l-72;

substitution of defined
are the only

equivalents;

and

"inference"

according to #1-1 and #1-11

processes which ever enter into any demonstration

in the logic of Principia.

The

result

is

of propositions for

that this development avoids the paradox of taking the logic granted in order to prove it. Nothing of the sort is
explicitly stated postulates

assumed except these


served.

whose use we have ob


is

And

it

results

from

this

mode

of

development that the system

completely symbolic, except for a few postulates, #1 1, #1-7, etc., involving no further use of then "either ... or ...","... and
"if
. .

.",

"

etc M
\^

We have now seen that the calculus of propositions in Principia Mathematica avoids both the defects of the Two-Valued Algebra. The further comparison of the two systems can be made in a sentence Except for the
:

absence, in the logic of Principia, of the

=|=

0, etc., etc.,

redundance of forms, p, p = 1, and the absence of the entities and 1, the two systems
of this part of Principia

are identical.
into a valid

Any theorem
of the

can be translated

Two-Valued Algebra, and any theorem of the Two-Valued Algebra not involving and 1 otherwise than as {=0} or
can be translated into a valid theorem of Principia. In fact, the qualification is not particularly significant, because any use of and 1 in the Two-Valued Algebra reduces to their use as = and = 1 For
{

theorem

as a term of a sum,

and

as a factor, immediately disappear, while the


1 in

presence of

as a factor

wise expressed.

and the presence of But p = is -p, and p =

sum can always be

other

1 is p.

Hence the two systems

Systems Based on Material Implication

287

are simply identical so far as the logical significance of the propositions

they contain

is

concerned. 25
of our treatment of propositional functions with the

The comparison
same topic
In the

in Principia is
first place,

not quite so simple. 28


is,

there

in Principia, the

"theory

of

types,"

which

concerns the range of significance of functions.


sideration of this.

But we
of

shall

omit con

Then, there are the differences

notation.

Where

we
or

write

U<px,

or

H x (px,
by

Principia has
.
<px.

2x

<px,

Principia has (Kx)

and where we write S#r, (or) further and more important difference
.
<px\

may

be

made
-{(.?)

clear

citing the
.

assumptions of Principia.
~#r.
~<px.

*9-01

.?&}. =

(3.r)
.

Df. Df.
Df.

*9-02
*9-03
25

~{(3a:)
(x)
.
<px

.?&}. =

fa)
(x)

.vp

<px

vp.

This

may be proved by

are contained

amongst the propositions

noting that, properly translated, the postulates of each system of the other. Of the postulates in Principia,

rendered in our notation:

^1-01

is

(peg)

(-p +

q),
is is

which

is

contained in our theorem 9-3.

^1 -2
sfcl

is
is,

(p + p) cp, which

a consequence of our theorems 2-2 and 5-33.

-3

pc
p+

(p + q),

which

our theorem 5-21.

^1 -4
^1-5
by
^el-6

is
is,

(p + q) c(q + p), which follows from our theorem 4-3,

by

2-2.

+
(</

r)

cq

+ (p +

?),

which

is

a consequence of our theorems 4-3 and 4-4,

2-2.

is (q cr) c [(p + q) c(p + r)], which is a consequence of our theorem The remaining (non-symbolic) postulates are tacitly assumed in our system.

5-31,

by

2-2.

Of our postulates, 1-1-1-9


1 is

in

Chap,

n and

9-01 in Chap, iv:

a consequence of

^1-7 and ^1-71

in Principia.

2
3

is
is

-^4 -24 in Principia.

^-4-3 in Principia.

4
5
of

is

^2-3

in Principia.
"If

is

equivalent to

0,

then a x

=
0",

hence to -x c -(a

x),

which

is

a consequence

*3-27

in Principia,

by *2-16.

= x), is a consequence of -#-4-71 and ^-4-61 in Principia, 1-61, in the form -(x -a) c (x a by ^4-01 and *3-26. = y)]c-y, is a consequence of #4-71, *5-16, and 1-62, in the form [(y a = y)(y -a ^2-21 in Principia. to [(a; = 1)Q/ = 0)] c (x = -?/), hence to (x -y) c (x = -y), which is an 1-7 is
equivalent

immediate consequence
1-8
1-9
is is

of ^-5-1 in Principia.

^K-4-57 in Principia.

^4-71
is

in Principia.
(q

9-01

equivalent to

1)

[p

(p

q)},

hence to q c

[p

(p

q)],

which

is

an

immediate consequence
26

of -^5-501 in Principia.

See Principia,

i,

15-21.

288
In this

A
last,

Survey of Symbolic Logic

two

sides.

note the difference in the scope of the "quantifier" (x) on the If the dots be replaced by parentheses, *9 03 will be
{[(x)
.

^r]

vp] =

{(x)

[tpxyp]}

in the scope of each of the further definitions. 27

A similar difference
*9-04
p
.

(x) or (3.r)

on the two sides characterizes

(x)
<px

<px

=
:

(x)
.

pv
.

<px.

Df.
p.

*9 05
#9 -OG
*9-07

(3.r)

v
.

= =
.

(3.r)

<px

Df

p ***
(.r)
.

("3.x)

<px

(3z)
:

pv
(#)
(.T)
:

<px.

Df.
.

<px

mv
.

(3#)
.
(.I-)

^
^r

= =

(3#)

^ v ^.
$y v
^.r
.

Df.

*9 OS

(3y)

^
.

(3y)

Df

Besides these definitions, there are four postulates (in addition to those

which underlie the calculus of elementary propositions).

*9 *9

1 1 1

h: hs

<?x

(3
.

z)

<pz.

Pp.
.
<p

<f>x

^?/

(3

2)

Pp.
is

*9- 12

What

is

implied by a true premiss

true.

Pp.

#9-13 In any assertion containing a real variable, this real variable may be turned into an apparent variable for which all possible values are asserted
to satisfy the function in question.

Pp.
is

By
#9-01
for
<px.

proved proposition.
is
-H<px

our method, every one of these assumptions, except ^9-12, In our notation,

= S
=

-<px,

which

is

our theorem 10-1, with

-<px

substituted

*9-02

is

-2^.i<px.

n-^.r, which

is

our theorem 10-12, with

-<?x

substi

tuted for

*9-03

is

tt<px

P = n x (^c + P),
II X

which which

is
is

our theorem 10-32.


10-33.
10-3.

*9-04
*9-05 *9-06

is
is

is
is is

P + n^r = S^ + P = P+ S^r U<px

(P+

<px),

2 x (^-c + P), which

is is

2 X (P+

<px),

which
$y),
^.r),

10-31.
is is

*9-07
*9-08

+ Z^y = U x 2 y = U x 2 y (ty + 2ty+tt<px +


(<px

which which

contained in 12-5,
also contained in 12-5.

The
<py,
<pz,

etc.,

The postulates require explanation. to represent values of the function


<px

<px.

authors of Principia use In other words, where

we have written
"Ibid.,
i,

they simply change the

letter.

This

is

a valid con-

135-38.

Systems Based on Material Implication


vention (though
is

289

it

often renders proofs confusing) because the range of


<p,

<px

determined by not by x, and x is conventions aside indifferent. z in where we should write n is called a "real variable", x in (x) . and (3or) . an "apparent variable". With this it is clear
<pz,
<px

<px

<px,

explanation,

that:

*9-l
#9-11

is

<px

c2<px,

which
c2<px,

is

10-21.
is

is

<px

m+

<px

which

an immediate consequence

of 10-21,

by 5-33.

*9-13

is

"If

whatever value of
is

x, in

<px,

xn

may

be,

<px

n,

then

n^.r,"

and

this implication

contained in the equivalence stated by 10-23.


in Principia

These principles which are assumed


ficient to give all further propositions

Mathematica are suf

concerning functions of one variable,


<px^
<px

without assuming
etc.), (3a-)
.
<px

(x)

.
<*r

to be the product of
of
<px

2,

etc. (or

<py,

<pz,

to be the

sum
.

lf

<f>x

2,
"

etc.
<px

These are simply assumed


x",

as

new

primitive ideas, (x)


"

<?x

meaning
a:".

for all values of

(3or)

.
<px

meaning

<px

for

some values

of

This procedure obviates


<px

all

questions

about the number of values of x

in which troubled us and secures the universality of theorems involving prepositional functions without any discussion or convention covering the cases in which the values of the vari

able are infinite in number.


of

The

proofs in Principia reflect this difference

method.
all

They
+

are, in general,

what ours might have been


etc.,

if

we had

based

further proofs directly upon 10-23 and the propositions con

necting

2<f>x

with 2 x

(<px

+ P),

not making any use, after 10-23,


2<^.r

of the properties of TLtpx as a product, or of

as a sum.

The theory
*11 *11

of functions of

two

variables, in Principia Mathematica,

requires two further assumptions:


-01
(x, y)
.
<p(x,

y)

=
.

(x)
:

(y)
:

.
<p(

y).
<p(

Df.
,

-03

(3z, y)

<p(x,

y)

(3.x)

(30)

y).

Df.

These are identically our assumptions:


11-06
11-05

nx

<p(x,

y)

2 XlV

<p(x,y)

H x ll v y), and = S x S^fey).


==
<f>(x,

The

difference

between the treatment

of propositional functions

which

we have given and

the treatment in Principia is not necessarily correlated with the difference between our treatment of propositions and theirs. The

method by which we have developed the theory


20

of propositional functions

290

Survey of Symbolic Logic

might exactly as well have been based upon the calculus of elementary A few minor propositions in Principia as upon the Two-Valued Algebra.
alterations

would be

sufficient for this change.


is

The

different procedure

for propositional functions, in the twr o cases,

a difference to be adjudged

independently, without necessary reference to the defects of the

Twothere
to the

Valued Algebra which have been pointed out.

Beyond the important


different use of notation.

differences

w hich have been mentioned,


r

are minor and trivial divergences between the

two systems, due

Neglecting these, we may say that the two methods give the same results, with the following exceptions: 1. There are certain complexities in Principia due to the theory of
types.
2.
3.

In Principia the conditions of significance are explicitly investigated. Principia contains a theory of "descriptions", account of which is

here omitted.

But none

of these exceptions is a necessary difference.

They

are due to

the more elementary character of our presentation of the subject. may, then, say loosely that the two methods give identical results.

We

The

calculus of classes

and

of relations

which we have outlined

in the

preceding sections bear a similar relation to the logic of classes and of relations in Principia; that is to say, there is much more detail and com
plexity of theory in Principia, but so far as our exposition goes, the

two are

roughly the same.

And

here there

is

no important difference of method.

It should now be clear how the logic of Principia is related to the logic we have presented, following in the main the methods of Peirce and Schroder. There is much difference of method, and, especially in the case of the cal

culus of propositions, this difference

is

in favor of Principia.

And

in

Principia there

is

much more

of theoretical rigor

and consequent complexity:


"descrip

also there are important extensions, especially in the theory of


tions"

and the

logic of relatives.

But

so far as the logic

which we have

expounded goes, the two methods give roughly identical results- When we remember the date of the work of Peirce and Schroder, it becomes clear

what

is

our debt to them for the better developments which have since

been made.

CHAPTER
The systems
implication,
is

V
1

THE SYSTEM OF STRICT IMPLICATION


discussed in the last chapter were
"The

all
*

based upon material

a false
is

statement, p is true and q false/ p cq meaning exactly We have already called attention to the fact that statement".
"implies".

this

not the usual meaning of


of ordinary inference
is

Its

divergence from the


"A

"implies"

exhibited in such theorems as

false

proposition implies any


2

proposition",

and

"A

true proposition

is

implied

by any proposition". The present chapter intends


sitions

to present, in outline, a calculus of propo


"

based upon an entirely different meaning of implies one more in accord with the customary uses of that relation in inference
is
"-

which

and

proof.

We

shall call it the

system of Strict Implication.

And we

shall

refer to Material Implication,

meaning
it

either the

Two-Valued Algebra or

the calculus of propositions as

appears in Principia Mathematica, since

It will appear that the logical import of these two systems is identical. Strict Implication is neither a calculus of extensions, like Material Impli

and the Boole-Schroder Algebra, nor a calculus of intensions, like It includes relations the unsuccessful systems of Lambert and Castillon. Strict of both types, but distinguishes them and shows their connections.
cation

Implication contains Material Implication, as it appears in Principia Mathematica, as a partial-system, and it contains also a supplementary
partial-system the relations of which are those of intension. The numerous questions concerning the exact significance of implication,

and the ordinary or


Section V.
It will

"proper"

meaning

of

"implies",

will

be discussed in

be indicated

how

Strict Implication,

by an extension to proposi-

tional functions, gives a calculus of classes and class-concepts which exhibits In this, it provides the their relations both in extension and in intension.
Various studies toward this system have appeared in Mind and the Journal of Phi losophy (see Bibliography). But the complete system has not previously been printed. We here correct, also, certain errors of these earlier papers, most notably with reference to
1

triadic
2

"strict"

relations.
illustrations, see

For further
Logic,"

Chap, n, Sect,
etc.

i,

and Lewis,

"Interesting

Theorems in

Symbolic

Jour. Philos., Psych.,

(1913), p. 239.

291

292

Survey of Symbolic Logic

calculus of intensions, so often attempted before, so far as such a calculus


is

possible at

all.

I.

PRIMITIVE IDEAS, PRIMITIVE PROPOSITIONS, AND IMMEDIATE CONSE

QUENCES

The fundamental
Symbolic Logic and
1.

ideas of the system are similar to those of

MacColl

its

Applications.
q, r,

They

are as follows:

Propositions: p,

etc.
"p

2.
3.

Negation: -p, meaning


Impossibility: ~p,
3
true".

is

false".

meaning

"p

is

impossible",

or

"It

is

impossible

that p be
4.

The
q

logical product:
is
true".

p xg or p

q,

meaning

"p

and

both",

or

"p

is

true

and

5.

Equivalence:

p =

q,

the defining relation.

Systems previously developed, except MacColl s, have only two truthThe addition of the idea of impossibility and values,
"true"

"false".

gives us five truth- values,


(1)
(2)

all of

which are familiar

logical ideas:

p,

"pis

true".

-p,

"pis

false".

(3)
(4)

~p,

"p

is impossible". is is

p,

"It

false that

is impossible"

i.

e.,
i.

"p

is possible".
"p

(5)
true".

p,

"It

impossible that p be

false"

e.,

is

necessarily

Strictly, the last

two should be written -(~p) and -(-p): the parentheses


at

are regularly omitted for typographical reasons.

The reader need be


it is

no pains to grasp -~p and

as simple ideas:

understand -p and ~p, and to remember that each such It should prefix affects the letter as already modified by those nearer it. is equivalent be noted that there are also more complex truth-values. p as will be shown, but - ~ -p, to p, etc., are irreducible. p,
sufficient to
p,

We

shall

have occasion to make use of only one of these,

"

p,

It is
4
false".
,

false that it is impossible that

p be

true"

i.

e.,

"p

is

possibly

Each one
3

of these

complex truth-values

is

a distinct and recognizable idea,

though they are seldom needed

in logic or in

mathematics.

here use a symbol, ~, which appears in Prindpia Mathematica with a different meaning. The excuse for this is its typographical convenience. 4 is possibly is possibly true" and MacColl uses a single symbol for -~p, p,
"p "p

We

false".

The System of

Strict Implication

293

The dyadic
truth- values

relations of propositions can be defined in terms of these


logical product,

and the

p q?
Def.
true"

1-01
~(pq),
are

Consistency.
"It

poq = -~(pq}.
"It

is

impossible that p and q both be


is Hence -~(pq), and q are consistent".

would be

"p

and

inconsistent".

possible that

p and

q both be

true",

represents

"p

1-02

Strict Implication.

p-*q

~(p-q)-(p-q).

Def.

1-03 1-04

Material Implication.
Strict Logical

pcq =
p+
q

Def.
Def. Def.
p)
.

Sum.

p *q = ~(-p-r/).

1-05

Material Logical Sum.


Strict Equivalence,

= =

-(-p -q)(p
*

1-06

(p

q)

q) (q

-*

Def.

We

here define the defining relation

itself,

because by this procedure we

establish the connection between strict equivalence


Also, this definition

and

strict implication.

makes

it

possible to deduce expressions of the type,

p =

5 something which could not otherwise be done. q mains a primitive idea as the idea that one set of symbols

But p =

q re

may
Def.

be replaced

by another.
1 07

Material Equivalence,

(p

q)

(p c

q) (q

c p).

These eight relations the seven defined above and the primitive rela = q are the relations divide into two sets, p q, p c q, p + q, and p tion, p q

which

figure in

any calculus

of Material Implication.

We

shall refer to

them

as the

"material

relations",

p o

q,

-i

q,

p*q, and p =
".

q involve

the idea of impossibility, and do not belong to systems of Material Impli We may anticipate a These may be called the strict relations cation. of these two sets, which results from the exhibit the
"

little

and

analogy

theorem
~(pq)
shortly to be proved.
Strict relations
:

= -(pti
Material relations

piq =-(pO-q) pAq = _(_ p


(p

pcq =
P+<1

-(p -q)

o-<7)

q)

= -(po-q) x-(qo-p)

(p

=
"

q)

= -(-P-0) = ~(p -tf) *-( c

l~P">

hesitates to The "circularity" here belongs inevitably to logic. No mathematician If theorem A, then theorem B, of two propositions by showing that prove the equivalence know that a reciprc and if theorem B, then theorem A". But to do this he must already And the italicized relation is equivalent to an equivalence. then . .
"if
. .
.
."

to"

be assumed. represents a relation which must

294

Survey of Symbolic Logic

The reader will, very likely, have some difficulty in distinguishing in meaning p q from p c q, p A q from p + q. The above comparison may be of assist
-J

ance in this connection, since

it

translates these relations in terms of


q,
"p

p oq

and p
g,"

q.

We
q,

shall be in

no danger of confusing p o
true".

is

consistent with

with p

"p

and

q are both q

Both p A

and p +

would be read

"Either

p or

q".

But p A

q denotes
"

a necessary connection;

p + q a merely factual one.


q,
"2

day

is

Monday",

and

+
is
.

4:".

Let p represent To Then p+q is true but p A q is

false.

In point of

fact, at least one of the

two
is

propositions,

"Today is

Monday"

and

"2

4",

true;
.
.

but there
."

no necessary connection
in this respect.

between them.
the

"Either

or

is

ambiguous

Ask

members
or 2

of

Monday
"Either

+
or

any company whether the proposition "Either today is = is true, and they will disagree. Some will confine
4"

...

..."

to the

p*q

meaning, others

will

make
p =

it

include

the p + q meaning; few, or none, will


Similarly, the difference

make

the necessary distinction.

between p

and p

q is that

q denotes

an equivalence of logical import or meaning, while p = q denotes simply an equivalence of truth-value. As was shown in Chapter II, p = q may be and q are both true or both false". Here again, accurately rendered
"p

the strict relation, p


relation,

q,

symbolizes a necessary connection; the material

g,

a merely factual one.

The
1-1
If

postulates of the
q-i q

system are as follows

p
q are both true, then q

p and
q

and p are both

true.

1-2

-I

p
true, then

If q

and p are both


-*p
is

is

true.

1- 3
If

p
then p
is

true,

true and

is

true.

1-4
If

p(qr)4q(pr) p
is

true and q and

are both true, then q

is

true and

p and

are both

true.

1-5
If

p-i-(-p)

is

true, then

it is

false that

is false.

1-6
If

(p-tq)(q-lr)-l(p-lr)

strictly implies q

and

q strictly implies

r,

then p strictly implies

r.

The System of
1 7

Strict

Implication

295

~p-i-p
If it is

impossible that p be true, then p

is false.

1-8

p
"p

-J

~q

-J

~p
"

strictly implies
is

q"

is

equivalent to

is

impossible

strictly

im

plies

impossible

".

The
as

first six of

these present no novelty except the relation

-J

They
But,

do not, so

far, distinguish this

system from Material Implication.

we

shall see shortly, the postulates 1 7

and 1-8 are

principles of trans

formation;

and thus introduce the distinguishing characteristics of the system.


late 1 -7
is

they operate upon the other postulates, and on themselves, Postu


Postulate
If
1 -8 is
q,
.

(p-iq)

obvious enough. -i -J
(-

equivalent to the pair,

~p

~q)

p implies
is

then

is

possible

implies

possible
is
(

(~p

-J

~q)

-J

(-p

-J

-q)

If

impossible
is

implies
q
is

is

impossible
.

then

false

implies

false

These two

propositions are

more

"self-evident"

than the postulate, but

they express exactly the same relations.

(To eliminate parentheses, as far as possible, we make the convention


that the sign =, unless in parentheses, takes precedence over any other o and x that A that -1 and c take precedence over A + relation
;

and + take precedence over o and x c Thus


.

and that

-J

takes precedence over

and

is pq + -p -q -ipcq = (p c q)(p c r) pcqr


is

[(p q)
is

+ (-p
[pc.(q

-*
-</)]

(p

q)

r)]

[(p

c q)(p c
shall

r)]

However, where there


parentheses.)

a possibility of confusion,

we

put in the

The
1.

operations

by which theorems are to be derived from the postulates

are three:
Substitution.
If

Any

proposition

may

be substituted for p or q or
If

r,

etc.

is

a proposition, -p and ~p are propositions.

p and q are

propositions,

is

a proposition.

Also, of

any
is

pair of expressions related

by =
2.

either

may

be substituted for the other.


If

Inference.

is

asserted

and p
is

-J

asserted, then q

may

be

asserted.
cation,
3.

(Note that this operation


q.)

not assumed for material impli

pc

Production

It

p and

q are separately asserted,

may

be asserted.

These are the only operations made use

of in proof.

296
In order to

A
make

Survey of Symbolic Logic

clearer the nature of the strict relations,

and particu

larly strict implication,

we

shall

correlates in terms of strict relations.

wish to derive from the postulates their This can be done by the use of

postulate

8 and

its

consequences, for by 1-8 a relation of two material

relations can be transformed into a relation of the corresponding strict


relations.

But

as a preliminary to exhibiting this analogy,

we must prove

number

of simple but

fundamental theorems.

These working principles

will constitute the

remainder of this section.


will

The

first

theorem

be proved in

full

and the proof explained.

The

conventions exemplified in this proof are used throughout.

2-1

pq-tp
1-6 [pq/p; qp/q; p/r}:
1-1 xl-2-i (p
1-6,

q*p)
q
is

This proof

may

be read:
for
r,

"Proposition

when p

substituted for p,

qp
(P q
is

for
*

q,
"

and p

states that propositions 1-1


of the proposition

and 1-2 together imply


line of proof
is

P)

The number

which states any


is

given at the beginning of the line.


"p

Next, in braces,
indicates that

indication of

any

substitutions to be made,

q/p"

in the proposition cited, for p;

"p+ q/r"

p q would indicate that


proposition
1

to be substituted,

p+

was to

be substituted for

r,

etc.

Suppose we take

6,

which

is

and make the substitutions indicated by {p


get

q/p;

qp/p; p/r}.

We

then

(pq-lqp)(qp-lp)
This
since
is

(pqlp).
But

the expression which follows the brace in the above proof.


q
-l

is

1-1,

and

-J

is

1-2,

we

write 1-1 x

2 instead of

q-iqp)(qplp). This calls attention to the fact that what precedes the main implication sign is the product of two previous propositions.
(p

Since 1-1 and 1-2 are separately asserted, their product

may

be asserted;

and

since this product

may

be asserted, what

it

implies

the theorem to be

is its

proved may extreme brevity.

be asserted.

The advantage of this way of writing the proofs Yet anyone who wishes to reconstruct the demon

stration finds here everything essential.

2-11

(p

= q)4(p4q)
2-1 {p-lq/p; q-ip/q}: (p *
1-06: (p
q) (q *

p) 4 (p

* q)

q)

(p

The System of

Strict Implication

297

2-12

(p

= (/HfeHp)
Similar proof, 1-2 instead of 2-1.

2-2

(p-i q)

-J

(~g

-J

~p)

1-06: 1-8

[(p4q)4(~q4~p)}[(~q*~P}*(P*<l)}

l)

2-1: (1) H Q.E.D.

In this last proof,


(1),

we

introduce further abbreviations of proof as follows:

and placed after a lemma which has been established, thereafter in the same proof we write (1), or (2), etc., instead of that lemma. line of proof instead Also, we shall frequently write "Q.E.D." in the last In the first line of this proof, the of repeating the theorem to be proved. substitutions which it is necessary to make in order to get
or (2), etc.,
is

1 -8

[(p

q)

* (~q

-*

~p)][(~<7

H ~p)

-J

(P

<

</)]

are not indicated because they are obvious.

And

in the

second

line, state

ment

of the required substitutions

is

omitted for the same reason.

Such

abbreviations will be used frequently in later proofs. Theorem 2-2 is one of the implications contained in postulate 1-S. of a By the definition, 1-6, any strict equivalence may be replaced by pair
strict implications.

By

either of these postulate 1-2 and theorem 2-1,

implications

may
2

be taken separately.

2-21

(~q*~p)4(p4q)
1
:

[(1) in

proof of 2

2]

-J

Q.E.D.

This

is

the other implication contained in postulate 1-8.


q)

2-3

(-P *

* (~q * P)

1-1
2 .2

{-q/p

-p/q}:

-q-p-*-p-q
-q/q}
:

{-g -p/p;

-p

(1)

-J

-*
[~(-/>
-<?)

~H

"rfl

1-02: (2)

= Q.E.D.
pp-lp

2-4

pip
1-2 {p/q}:
1-G: l-3x(l)-{ Q.E.D.

2-5

-(-p)*p
2--i

{-p/p}:

-p*-p
(1)
-I

2-3 {-p/q}:

Q.E.D.

2-51

-(-p}

=p
2-5x1-5 = Q.E.D.

1-06:

298 2-6

A
(-p
-i

Survey of Symbolic Logic

-g)

-J

(g

-J

p)
-J

2-3 {-q/q}: (-p


2-51:
(1)
( fy
-j

-g)

-i

[-(-q) H p]

(1)

= Q.E.D.
-p)

2-61

(p

-j

-g)

-i

2-6 {-pip}:
2-51:
(1)

K-p) -gM = Q.E.D.


-*

(g

<

-P)

(1)

2-62

(p

q)

-I

(-g

-p)
-i

2-61 {-q/q}: [p
2-51:
(1)

-(-g)]

-J

(-g H -p)

(1)

= Q.E.D.

2-6J

(png) = H-t-p)
1-06:

2-62x2-6 = Q.E.D.
2-61x2-3 = Q.E.D.
and 2-62 are the four forms
is

2-64

(p*-q)

= (q*-p)
2-3, 2-6, 2-61,
of the familiar

1-06:

Theorems
terms.

principle that an implication

converted by changing the sign of both

2-7

(~p4~q}4(-p4-q)
2-21 [plq; q/p}: (~p -J ~g) -j (q * p ) 2-62 {p/q; q/p}: (q -J p) H (-p -t -q) 1-6: (1) x (2) H Q.E.D.
(1)

(2)

2-71

(-p

-J

-q)

-i

(~p
-i

-J

q)

2.6: (-p

-?) H (g

p)
-i

(1)

2-2 {^/p; p/gj:


1-6:

(g S p)

(~jM

~g)

(2)

(l)x(2)^Q.E.D.

(-p-t-q)

1-06: 2-71

= (~p*~q) x2-7 = Q.E.D.


:

2-7 {-p/p; -g/g}


2-51:
(1)

(~

-p

-J

-g)

-J

[-(-p)

-J

-(-g)]

(1)

= Q.E.D.
:

2-73

(p4q)-i(~-p-i~- q )
2-71 {-p/p -q/q} [-(-p) 2-51: (1) = Q.E.D.
,

-i

-(-g)] 4 (-

-p s -

-g)

(1)

2-7J1

(p-ig)

= (~-p^~-g) 1-06: 2-73x2-72 = Q.E.D.


:

2-74

(p^g)H(-^p^-^)
2-62 {-g/p; ~p/g}
1-6:

(~g ^ -p) H (-

^p

-i

^g)

(1)

2-2x(l)^ Q.E.D.

The System of

Strict Implication

299

2-75

(-

~p

-i

~g)

-i

(p H

ry)

2-6 {~p/p; ~q/q]

(-

~p

-J

~q)

-J

(~g

-I

~p)

(1)

1-6: (1) x2-21-jQ.E.D.

2-76

(p-Jg)

= (-T
=
(-

!--</)

1-06:

2-74x2-75 - Q.E.D.
-i

2-77

(p

-i

~p -i - ~g) - (~ -p ~ -g) = 2-76 x2-731 x2-G3 x2-712 = Q.E.D.


g)
"

(-g

-{

-p)

(~g

-I

~p)

"

"p

implies

g"

is
"

equivalent to

p
"

is

possible
q
is

implies

q
"

is is

possible

is

equivalent to
"q

is

necessary

implies
is

necessary
"

equivalent to
implies

is

false

implies
".

is

false

equivalent to

is

impossible

is

impossible

2-6-2-77 are various principles for transforming a strict implication. These are all summed up in 2 77. The importance of this theorem will be
illustrated shortly.

2-8

pq = qp
1-1
{g/p; pfq}: q

-I

(1)

1-06: 1-1 x(l)

- Q.E.D.
(1)

2-81

p =

pp

1-2 {plq}: pplp 1-06: 1-3 x(l) = Q.E.D.

2-9

p(qr) = q(p r) 1-4 {g/p; plq}: q(p


1-06: 1-4 x(l)

r)

-i

p(q

r)

(1)

= Q.E.D.
q)
q)

2-91

p(qr)

(p q)r

2-8: p(qr)

2-9: p(rq)
2-8: r(pq)

= = =

p(r
r(p

(p q)r

The above theorems


briefly

constitute a preliminary set, sufficient to give

most further

proofs.

II.

STRICT RELATIONS AND MATERIAL RELATIONS

We