VYflVH

<
>

<

>

u
v

<

VV A

>,>

-v
;

llllii
iiiitii?iii

PRI NCI PLES

OF THE

ALGEBRA OF LOGIC
WITH EXAMPLES

ALEXANDER MACFARLANE
M.A., D.Sc. EDIN.), F.R.S.E.

READ BEFORE THE ROYAL SOCIETY OF EDINBURGH

rit/i DECEMBER 1878 AXD 2of/i JANUARY 1879

EDINBURGH

DAVID DOUGLAS
1879

[All rights reserved.}

Edinburgh

Printed by Thomas and Archibald Constable

FOR

DAVID DOUGLAS.
LONDON

HAMILTON, ADAMS, AND

CAMBRIDGE

MACMILLAN AND

GLASGOW

JAMES MACLEHOSE.

PWlleS

631004

CO.

CO.

PRINCIPLES
OF THE

ALGEBRA OF

LOGIC.

generation will arise in which the leaders of education will

know

the

value of logic, the value of mathematics, the value of logic in mathematics,

DE MORGAN, Syllabus, p. 44.

in logic.
we then err in regarding that as the true, science of Logic,
laying down certain elementary laws, confirmed by the very testi

and the value of mathematics

Shall

which,

mony

of the mind, permits us thence to deduce, by uniform processes, the

entire chain of its secondary consequences,

application
in

any

methods of

science,

and furnishes
Let

perfect generality.

it

viewed either as a system of truth or as the foundation of a

practical art, there can properly be

any other

test

of the completeness and

laws, than the completeness of

its

system of

and the generality of the methods which

BOOLE, Laws of Thought, p. 5.

it

serves to

fundamental characters of

its

derived truths,
establish.
It

is

for its practical

be considered whether

curious to

compare the properties of these quaternion sym

bols with those of the Elective

wonderful treatise on the

Symbols of Logic, as given

Laws of Thought ; and

in

Boole

to think that the

same

grand science of mathematical analysis, by processes remarkably similar to

each other, reveals to us truths in the science of position far beyond the
powers of the geometer, and truths of deductive reasoning to which unaided
thought could never have led the logician.
p. 50.

PROFESSOR TAIT, Quaternions,

PREFACE.
THESE

Principles

were originally contributed to

the Royal Society of Edinburgh in a

Memoir

received

the Secretary Qth October 1878, and in a supple

by

mentary paper received 5th November. I had the

honour of reading an Abstract before the Society at
the meetings of i6th December and 2Oth January.
In the interval between the 5th
present time

November and

the

have improved several of the demon

strations, introduced illustrative matter, and prepared

The work,

the collection of examples.

state,

in its present

forms an elementary treatise on the science of

Formal Reasoning.
I

consider

it

proper to state that the theory of the

operation of the mind in reasoning about Quality,

which

is

advanced

years ago

in this

and that

me

five

have directed towards

its

work, occurred to

development the whole of

my

subsequent study of

the Mathematical, Physical, and Natural Sciences,

which are embraced in the curriculum for the degree
of Doctor of Science (Mathematics) at the University
of Edinburgh.

ALEXANDER MACFARLANE.
EDINBURGH, i$d January

1879.

TO THE

REV. PHILIP

KELLAND,

M.A., F.R.S.

PROFESSOR OF MATHEMATICS IN EDINBURGH UNIVERSITY

of

iftogal Socfetg of OEoinfourgfj

tfje

THIS

WORK

AS A

IS

DEDICATED

MARK OF RESPECT
BY

FORMER

PUPIL.

CONTENTS.
PAGE
I,

II.

The Science of Formal Logic an Algebra,

Universe and Character,

III.

The

sign

IV.

The

signs

V.

The

signs x

+ and -

and

-f-

VI. Rule of Signs,

VII. Integral Symbols,

VIII.

On

.15
17

.....
.....
.

20

25

26
28

the Equation as expressing a general proposition,

....

IX. The principle of Identity and the Axioms of

mediate Inference,

X. Axioms of Mediate Inference,

XI. Conditions for a Character being Single,

36

40

42

......

XII. The signs of inequivalence

XIII. Division,

Im

XIV. Expansion of a

and

>

function of a

<

53

54

number of independent

symbols in terms of the primary parts into which

the universe

XV.
XVI.

Definition,

is

...

divided by the symbols,

61

69

Inference from one or more equations of the form

x=m

XVII. Inference from one or more

xy=m
XVIII. On

(Categorical),

81

(Hypothetical),

certain forms of the disjunctive equation,

XIX. The

equations of the form

Aristotelian forms of inference,

-106

.113

Contents.

XX. On

Probability,

XXI. Fundamental
of a

?AGE
118

relations

between the single functions

number of independent

XXII. General Method of deducing

form from given data,

XXIII. On Boole
Example s,

General Method,
.

characters,

122

a conclusion of a required
.

.124
.

131

135

ALGEBRA OF

LOGIC.

I.

THE SCIENCE OF FORMAL LOGIC AN

ALGEBRA.

1. THOUGH it is evident a priori to one who reflects on

the matter, that the theory of Necessity and the theory of
Probability are the complementary parts of one whole, it

is

nevertheless true that the foundations of the

science, of

which they form the

parts,

were not

general

laid until

The merit of conceiving and under

quite recent times.
taking this important unification is due in some measure
to De Morgan, but principally to Boole.
That the science of inference is capable of being
may be inferred from the fact that the
rules
about
Conversion and Syllogism are estab
ordinary
2.

treated analytically,
lished

by a comparison of

circles,

taken to represent the

In one of the best

terms of the propositions considered.

modern manuals of Logic, it is stated that the testing

whether a given combination of premises leads to a valid
inference, and the proof of the validity or invalidity, must
depend on the comparison of the spheres, within which,
according to the premises, the notions under consideration
find application ; and that these spheres are made apparent
to the senses by geometrical figures (especially by circles)
whose reciprocal relations agree with the* relations of the

spheres of the notions to each other in

for demonstration.

all relations essential

Ueberweg s Logic^ translated by Pro

fessor Lindsay, p. 379.)

The introduction
grams is commonly attributed to Euler.

of these dia

The Science of

which con

this graphical method,

the use of diagrams, there is an analytical method,
which consists in the use of symbols. The relative advan
the two, when applied to Quality,
tages and disadvantages of
The
are precisely the same as when applied to Quantity.
exhibits an individual case of the given data with
3.

Corresponding to

sists in

diagram

all

the clearness of the concrete

on the other hand, the

the essential relations from

analytical expression separates
the accidental, with which they must be mixed up in any
individual example.
4.

The reason why

the operations of Boole

calculus

that the
appear mysterious and its employment difficult, is,
a
sufficient
calculus is not founded upon
theory of the
That it
in reasoning about Quality.
mind
the
of
operation

not all that a Logical organon ought to be, is evident

The dis
from what Venn says in Mind, vol. i. p. 484
is
the
boldness,
tinctive characteristic of Boole s system
not to say audacity, with which he carries on his processes
is

through stages which have no logical or other signification

whatever, that is, which admit of no possible interpretation
result.
provided only they terminate in an interpretable
Boole himself claims nothing higher for his calculus. He
which
would, however, have objected to the statement

makes (Principles of Science, p. 71), that

Boole imported the conditions of number into the science
of Logic, and produced a system which, though wonderful
in its results, was not a system of logic at all.
Professor Jevons

5.

It is the object of this little

work

to investigate the

foundations of the analytical method of reasoning about

laid down
Quality, with special reference to the principles

by Boole as the basis of his calculus, and to the observations

which have been published by various philosophers con
I bring forward a new theory of
cerning these principles.
the operation of the mind in reasoning about Quality, which
enables me to correct Boole s principles, and place them
on a clear rational basis. I endeavour to show that the

Formal Logic an Algebra.

analytical method of reasoning about Quality is an Algebra,

which coincides with the Algebra of
Quantity when the
symbols are integral, but is a generalised form of the
latter when the symbols are fractional.
The rest of the
work is taken up with the investigation of problems
by
means of this algebraic organon, especially such

problems

as are suggested by the

ordinary Logic.
6. Logic, as the
Algebra of Quality, is a formal science.

It

investigates the general properties of the

symbol of Quality,
and by means of these properties deduces equations which
are true generally, or combines such equations with data

of given forms.

It

is

not

its

province to consider

how

particular form of datum can in any case be asserted to be

true
that subject of investigation being left to the Transcend

ental -Logic ; it is sufficient that

examples of such a form
occur, or may occur, in the practical or theoretical activities

of mankind.
7. The properties of the
symbol of Quality are not laws
of thought in the common acceptation of that term.
For
the properties of the symbol of
Quantity, on which the
ordinary algebra is founded, are held not to be laws of
thought, but to refer to the actual constitution of

things

and there

no

difference

the

two

methods, when
developed, which indicates the existence of such a dis
is

in

If the basis of the science of

Quality is subjec
so only in the same sense in which the basis of
the science of Quantity is subjective.
There is
for

tinction.
tive, it

is

ground

believing that the true reason why the former science has
remained so stationary is, that there has been too much
introspection into the individual mind in the hope of
rinding laws of thought there, and too little contemplation
of the form and nature of the truths of Science.
The

men reason equally well about

Quality, fallacies being possible only by a momentary lapse
of attention ; but the mathematician never assumes that all

logician assumes that all

men

reason equally well about Quantity.

The Science of

4
8.

Boole entitled

his

An

great work on reasoning

of Thought, on which

are
Investigation
founded the Mathematical Theories of Logic and Prob
he says that the Laws in
abilities, and in several places
in a sense in which the Laws of
question are subjective
He considers
are not.
of

the

Laws

Quantity

I have endea
in particular to be a subjective law ; but
voured to show (Art. 118) that it is a special condition,
which the symbol of this Algebra must satisfy in order to

be of a particular kind.
the properties of the
one, I suppose, contends that
laws
the
of
or
Chemical Symbol,
Quaternionic Symbols, are
Since the corresponding properties of the
of thought.
it is
different symbols differ greatly among one another,
constitu
actual
the
to consider
surely better in every case
tion of things as suggesting rules for thought to the mind,
9.

No

rather than the

10.

mind imposing laws of thought upon

Logic, as the Algebra of Quality,

is

itself.

a true organon.

can determine whether a conclusion of a required form

can be deduced from data of given forms ; and if so, what
It can manipulate complex data, as is
that conclusion is.
It

shown

in*

the

examples appended.

Syllogismus ad

media

principia

axiomata frustra

Bacon

scientiarum
adhibetur,

judgment

non adhibetur, ad

quum

sit

subtihtati

true of the scholastic ex

however
naturae longe impar
does not apply to the Algebra of
position of the syllogism,
be made to discover principles,
can
latter
the
for
Quality ;
and to imitate to some extent the subtlety of Nature. It
that in
adapt a remark of De Moivre)
and
of
pro
necessary
numerable questions in the theory
bable reasoning can be solved without any manner of
the mere force of the notation
trouble to the

may be

said (to

imagination,

by

supplied by this Algebra.

The Algebra of Quantity is acknowledged to be the
the Experimental
weapon for the philosopher who attacks

Formal Logic an Algebra.

Sciences; the Algebra of Quality is the weapon for the

who attacks the Sciences of Observation.

philosopher

Thus viewed, Formal Logic is not the short and dry

which even Kant held it to be. Any one who has

11.
science

studied Boole

s Calculus, may well imagine that the theory

of reasoning was not completed by Aristotle ; and that so
far from any
System of Logic having ever been written,

there

is still

need to consider the foundations.

UNIVERSE AND CHARACTER.

II.

Boole in his analysis of language draws no distinc

and Adjective ; he considers their

12.

tion between Substantive

He says (Laws of
function in reasoning to be the same.
The
substantive
Thought, p. 27),
proper and the adjective
may indeed be regarded

as differing only in this respect,

that the former expresses the substantive existence of the
individual thing or things to which it refers ; the latter im
plies

that

existence.

If

we

attach to the

universally understood subject

"being"

or

adjective
"

thing/

it

the

be

and may for all the essential

purposes of reasoning be replaced by the substantive.
Accordingly, he uses the symbol x to denote men or good
things or white things or horned things, as the case
may be. For instance he says, if x alone stands for
white things and y for sheep, let xy stand for white
sheep ; and in like manner if z stands for horned things
and x and y retain their previous interpretations, let zxy
represent horned white sheep.

comes

virtually a substantive,

13.

Again

when

investigating

the

operations

of the

mind, which are implied in the use of language as an in

strument of reasoning, he finds no difference in the opera-

Universe

6
tion expressed

and

Character.

by a substantive from

that expressed

by an

He

says that there is a universe of discourse;

but this universe is not one described by a substantive.
adjective.

In every discourse, he says, whether of the mind con

in his
versing with its own thoughts, or of the individual
intercourse with others, there is an assumed or expressed

which the subjects of its operations are con

The most unfettered discourse is that -in which the
fined.
words we use are understood in the widest possible applica
tion, and for them the limits of discourse are co-extensive
But more usually we
with those of the universe itself.

limit within

confine ourselves to a less spacious field.

Sometimes, in
we
imply (without expressing the limita
discoursing of men,
that
it is of men only under certain circumstances and
tion)

conditions that
the vigour of
relation.

Now, whatever may be the extent

may

of the field

the objects of our discourse are found, that

properly be termed the universe of discourse.

within which
field

speak, as of civilised men, or of men in

or of men under some other condition or

we

life,

all

Laws

of Thought, p. 42.
From the passage just quoted, as well as from many
others, it appears that what Boole means by the universe of
14.

is not the objects denoted by a Universal Sub

a definite part of the whole realm of things
but
stantive,
a limited portion of the physical universe, with all the

discourse

entities which are or can be imagined to be in it, whether

mental or physical, ponderable or imponderable, atomic or

complex.
15.

The

substantive

men

expresses

an operation of

election from the universe of all the beings to which the

term men is applicable; the adjective good in com
bination, as

good men,

mentally from the class of

directs us

men

all

still

those

further to elect

who

possess the

further quality of good ; and if another adjective were pre

fixed to the combination, it would direct a similar operation
In short, he supposes that the mind
upon good men.

Universe and Character.

always proceeds along the predlcamental line ; whereas that

is only one mode of its procedure.

In

16.

consequence of

this

analysis,

the

subjects

of

thought in Logic and in Arithmetic are said to be perfectly

distinct ; and it is not of any importance to compare the
symbols of logic with the symbols of quantity generally.
Attention

is

the symbols

directed so exclusively to an Algebra in which

x y, z, etc., admit indifferently of the values
t

and of these alone, that some logicians have sup

the symbols can have no other value.
that
posed
17. Another consequence of this -analysis is, that Boole
o and

i,

obliged to make a new and independent investigation of

Secondary Propositions. In the case of Secondary Pro
of the symbol i is held
positions, the proper interpretation
The question is sug
to be eternity, or a part of eternity.
of
case
in
the
whether
Propositions i does
Primary
gested,
He thinks not; because the
not really represent space.
is

sign

of identity

connecting the members of the corre

sponding equation, implies that the things which they repre

sent are identical, not simply that they are found in the
same portion of space. The reason why the symbol i in

Secondary Propositions represents not the universe of events,

but the eternity in whose successive moments and periods
they are evolved, is, that the same sign of identity connect

im
ing the logical members of the corresponding equations
which
those
members
the
events
that
not
represent
plies,
are identical, but that the times of their occurrence are the

same.
18.

much
taking

Laws of Thought, p. 176.

The principles of the Calculus

of Identity

become

and

their application greatly facilitated, by

clearer,
into consideration the difference of the functions of the

and of the Adjective used in an attributive sense.

objects expressed by the common noun, or rather
universal term of a proposition, constitute the universe
Substantive,

The

the actual whole considered by the

of the proposition
mind in forming the judgment. The attributive adjectives,

Universe

and

Character.

whether one or more, which appear in the proposition, refer

to that subject, and not to things in general.
In thinking
sheep that are white and horned, I do not consider
white things or horned things.
It is even questionable
whether the mind can consider some adjectives as denoting
of

classes of things.
Can we consider small things or wise
things or primary things ? Boole remarks, with reference
to this very attribute wise, that, before denoting it by a
symbol, we must consider whether it is to be used in its

absolute

sense

or

absolute sense.

The

Categories,

is

But

only relatively.

Nothing by

described as great or small.

"large,"

down

in

A moun

may be said to be little," and a millet

from the fact of the one being greater, and

tain, for instance,

seed

has no

small

itself/ Aristotle lays

"

the other

It
less, in respect of things of the same nature.
reference to things of the same nature that I wish to
draw attention to.

is this

19.

As

quantities

have a certain abstract meaning in

themselves, but no definite meaning unless with reference

to a given unit ; so qualities have a certain abstract meaning
in themselves, but no definite meaning unless when referred
to a given universe.
20. Let the particular kind or collection of objects
considered in any judgment or series of judgments be
denoted by a capital letter
a symbol used in an analo

gous sense by -De Morgan.

tion

of objects

When

the

same kind

or collec

the judgments con

need not be expressed, but is to be understood.
sidered,
Let an attribute, character, or quality, be denoted by a
is

small
21.

the subject of

letter, as x.

This modification of Boole

contrast between the Substantive

is

all

notation brings out the

and the Adjective; which

indeed only one form of the general contrast between

which is the subject of the operations of thought and

that

the operations themselves.

Another common form of the
contrast is that, made prominent in the Theory of Prob-

Universe and Character.

between the

ability,

event

and

the

9
in

way

which

it

can

happen.
If

U denote the

Members of the House of Commons at

x may denote Liberal or Conservative.
denote triangles/ x may denote isosceles or
c

the present time/

Or

if

equilateral.

22. Boole, in his investigation, generally considers a com

bination of a Substantive with an Adjective prefixed.
Some

languages, however, show, by a difference of position

before or after the Substantive
that the Adjective may be

used in two senses,

viz., as forming part of the Substantive

In the
equivalent to a relative phrase.
English language, where the adjective is commonly placed
before the substantive, the distinction referred to is brought

or as

merely,

The prefixed adjective, when emphasised,

to
a
relative phrase ; when not emphasised, it
equivalent

out by emphasis.
is

is

part of the subject of thought.

23. For example
I find on looking

up a Polyglot Bible
the proposition of Proverbs xv. 20 is expressed as
follows
wise son maketh a glad father.
:

that

Ytos cro^os

ev(/>paivei

Trarepa.

Filius sapiens laetificat patrem.

Ein weiser Sohn erfreuet den Vater.

L enfant
II figliuol

sage rejouit son pere.

savio rallegra

El hijo sabio alegra

Here the subject of discourse

al
is

il

padre.

padre.

Sons

and

it

is

to be

observed that while the two Teutonic languages place the

conditioning attribute before the subject, the others put it
after.

24.

Boole holds that Primary Propositions refer to things,

facts; and that the idea of time is in

and Secondary to

volved in the Secondary.

Now there are propositions
to
which
facts
do
not
involve time, or a collection
relating
of portions of time, as the underlying subject; for example,

Universe

io

and

Character.

those which refer to place or a collection of places. We

have not only the relative when, but the relatives where
*

who.
Hence if fact-propositions, which relate to the
identity of portions of time, required a special investigation,
those which relate to the identity of portions of space would

and

also require a special investigation.

If, however, we draw
a contrast between the subject and its characters, one

investigation suffices for all the different kinds of subject.

Instead of two I s, of which the one means the actually
existent universe, and the other eternity, there is an infinite

number of 7 s, any one of which may be the subject of

discursive thought.
25. This view of an essential difference in the functions

Common Noun and

of the
results

Adjective

&f philological research.

is

supported by the

According to

Max

Miiller

(Lectures on the Science of Language, vol. i. p. 291 ), the com

ponent elements of language, which remain at the end of a

complete grammatical analysis, are of two kinds, namely,

In such a
roots predicative and roots demonstrative.
as
the
where
the
Chinese,
language
predicative root may
as
itself
be
used
a
noun
or
a
or an adjective, the
verb
by

noun

distinguished from the verb by its collocation

In the Aryan languages no predicative
root can by itself form a word ; in order to have a substan
tive it is necessary to add a demonstrative root, this forming
is still

in the sentence.

the general subject of which the meaning contained in the

root is to be predicated.
If Boole s view of the operation
of the mind were correct, we should have only predicative
roots.

26. Aristotle, in

his discussion of the Categories,

a strong contrast between Substance

(ovo-ta)

draws

and Quality

and between Primary and Secondary Substances.

By Substance is meant a particular thing (root rt) ; by
(TTCHO V)

Quality that which

is

in a subject.

Primary Substance, the thing

in

number

In the case of the

and one

signified is individual

in the case of the Secondary, the thing signified

Universe and Character.

1 1

involves a quality, but so as to denote a particular kind of

It is the characteristic of Substance that being
substance.

one and the same in number

it

can receive contraries

the characteristic of Quality to be that with

respect to which things are said to be like or unlike.
27. The Primitive judgment is so called because it does

while

it

is

not refer to an exact subject, but to the whole external

universe as one substance having all the physical changes
which occur for accidents. For example, the judgment
It rains

refers to the states of a portion of the Physical Universe,

and of these equates the

present with raining.

ordinary Eulerian Diagrams do not represent the
whole of thought, but leave it indefinite ; unless we suppose
28.

The

to be represented by the finite sheet, on which the

butes are represented by circles.
it

attri

made up of a type and certain finite

type corresponds to the predicative part
of the Noun, of which Professor Max Miiller speaks, and
The most
the limitations to the demonstrative part.
29.

is

limitations.

in general

The

Space and Time ; which,

looked
be
upon as logical variables.
may
We may suppose the Time to be constant, and consider all
the U s throughout a given region ; or we may suppose the
individual to be constant, and consider its successive states
The zoologist, when he
within a given portion of time.

common

limitations are those of

in this aspect,

compares the members of a genus, takes them in the adult

state ; when he considers the life-history of a particular
form, he follows an individual through its cycle of states.
30. It is with the notion of the type that questions about
Essence and Abstract Ideas are more properly concerned.
Berkeley draws a distinction between two kinds of abstract
ideas.

As

qualities or

the

mind frames

mental separation attain

to

itself abstract

ideas of

by the same prescission or

abstract ideas of the more com

modes, so does

it

pounded beings which include

several co-existent qualities.

Universe

Professor Fraser

and

s Selections

Character.

from

Berkeley, p. 16.

He says

two kinds of abstraction to correspond.

The distinction considered is correlative to that between
Universe and Character.

also that there are

31. Arithmetical value of U.

Since 7 signifies a definite
collection of individuals of a given type, its arithmetical

value must be an integer.

but may be singular or

The

integer

infinite.

is

It is

in general fihiral,
infinite

individual parts are not discrete but continuous.

recognises two of these cases.

when

the

Grammar

It is interesting to consider how the

subject of thought
has naturally an integral value, while the operation of
thought has naturally a fractional value, how the relation

between the symbols

is

their kinds of quantity.

32.
also may be

instance, the

mirrored in the relation between

real

either

or

imaginary.

For

judgment

The

goat- stag

is

white

an imaginary universe of goat-stags. When

is
it
to
indicate
that
fact
imaginary,
appears proper
by saying
that the arithmetical value of
is o ; which is therefore, in
this aspect, an integer.
refers to

33. The Universe holds the same position in the Algebra

of Quality that the Unit does in the Algebra of Quantity.
It may be said to be a generalised tmit.
34.

and

When

x, y,

x, y, etc.,

meaning

etc.,

is

have a
is

used to denote the subject of thought,

to denote

operations on it, the symbols

value ; and as their

definite arithmetical

supposed to be fixed throughout a discourse,

must also be supposed fixed. If

their arithmetical value

denotes a single positive attribute, its value is a fraction

lying between o and i ; but if it is negative, its value lies
between o and
i.
Suppose that we have a complex
character

x and

y,

as xy ; being compounded of two characters

which are in their statement independent of one

Universe
another.

It is

and

Character.

then necessary to suppose that the arith

x and y are preserved independently of

metical values of
the combination

But

if

refers to

only.
symbols depend on
but to Ux, then the
to
not independent of x and y may have
for these

U, and y not

;
meaning of y is
several arithmetical values according to the several orders

it

has in combination.

or the whole ; while o denotes

all
i denotes
and o are to be considered as operating symbols
Some is an indefinite operating
of the same kind as x.
carries the additional meaning of
it
but
generally
symbol;
than nought
having an arithmetical value which is greater
35.

Thus

none.

<

36. It is very frequently necessary to express the arith

convenient notation is x. Boole
metical value of x.
the circumlocution Prob. x
uses for this

purpose

37.

The mind, when reasoning on

matters such as are

discussed in the Theory of Probability, considers a particular

class or kind of things ; as has been well shown by Venn in
his Logic

signification

lying

The /s and

^ s are in their first

arithmetical values
with
symbols,
A dependent event involves
and i.

of Chance.

selective

between o

another event as a presupposition; and its arithmetical

value depends on that connection a circumstance which

shows us that attributes which are independent in their

statement must be conceived as operating upon the uni

also

verse directly.

of Quality is the more general method.

the relations of the characters of a Universe,
whether that universe comprise one, several, or an infinite
and whether the characters change or are
number of
38.

The Algebra

It discusses

parts,

independent of time

whereas the Theory of Probability as

5),
(see Venn s Logic of Chance, p.
to comprise a very large or infinite
universe
the
supposes
number of individuals, and the proper arithmetical value
of to be a certain limiting ratio, to which the actual value

commonly

p
/

stated

is continually approaching the greater the

individuals in the universe.

of

number of

and

Universe

4
39.

There

is

an important

Character.
distinction

among

adjectives

The contrary of
according as they involve being or having.
the former is formed by prefixing not, that of the latter
by
Let us consider a collection of balls.
prefixing without.
attributes which, in the Theory of
Probability, such
balls are commonly supposed to have are such as
white,

The

These

black.

red,

are

attributes

they apply to the whole of the

whole of

surface.

its

ball,

of

or at

since

being,

least to

the

But we may consider a ball, which

colours upon its surface ; such attri

has stripes of different

having a white stripe or having, a red stripe
are adjectives of having.
The adjective with a white
white applies
stripe applies to the ball as a whole, but
the
only to the stripe. The former denote qualities

butes as

proper,

latter

As

what are called characters

in the Natural Sciences.

the term character

is the more
general of the two, I
instead of quality.
40. The ordinary doctrine of
Subject and Predicate is
often departed from in Manuals of
Logic without the writer,

prefer to use

it

apparently, being aware of

contraposition, converts

Every regular

figure

it.

For instance, Ueberweg, by

may be

inscribed in a circle,

into

Every

figure

which cannot be inscribed

in a circle

is

not

regular,

thus leaving

the subject of thought unchanged, and

the
characters regular and inscribable
manipulating only
in a circle.
41.

<

figure

De Morgan

prominence.

brought the idea of the Universe into

In his Syllabus (section
122) he defines it as

meaning the whole extent of matter of thought under con

sideration.
I do not consider this definition as
sufficiently
exact

as indicating precisely that which in the

Algebra of
Quality corresponds to the Unit of the Algebra of Quantity.
With him the term does not denote a definite
of

subject

The Sign

15

but rather limits within which

operations,

selective, operations

Hence

may be made.

not

symbols are

U and x,

Instead of separating between

denotes
the result by X.
he
Ux,

all integral.

as in

THE SIGN =

III.

42.

elective,

his

U{x=y}

By

meant, that those Z7 s which have the character x are

identical with those which have the character y ; or, in
is

U the character x

another aspect, that within the universe

is equivalent to the character y.

It follows that the equation is true arithmetically.

If

*sajr,

x=y;

then

the equivalence of two characters involves the equality of

their arithmetical values.
43.

Though the 7

s,

which have the respective characters,

one another, the characters
Their relation appears to be best

are asserted to be identical with

may

not be identical.

expressed by the term Equivalence.

If the proposition is
the characters derive their equivalence from nature ;
but if it is verbal, they derive their equivalence from con
real,

Two

vention.

characters

may be

or by definition, or by reality.
triangles

x, equilateral

equivalent in themselves,

For example

y, equiangular

let

U denote

then

U{*=y}
does not assert the identity, that is, the undistinguishable
sameness of the characters equilateral and equiangular,
*

but only their logical equivalence.

44. We may also have an equation of the form

U{x = x}
where x denotes the arithmetical value of x, and is sup
posed known. This equation also expresses the logical

The Sign

equivalence of the members; but the second

member

is

indefinite as regards identity.

For example let
denote
an infinite series of throws with a die/ and x denote
ace,
:

U{x =

then

This

is

J}.

an example of a form of equation which bulks

largely in the Theory of Probability.

45. Since x and y, when
regarded as operating on a
definite universe, acquire an
unambiguous and exact arith

metical value, no valid objection can be raised

against ex
pressing a judgment of the kind considered as an equation.
It has, indeed, been
rightly maintained by the opponents
of the doctrine of the Quantification of the
Predicate, that it
is by their attributes that we think of
things ; but I hope to
show that the supposed conclusion does not follow

that,

therefore, the proposition cannot

be expressed by means of
The theory of reasoning about quality, ad

an equation.
vanced in these Principles
alone

is

denotative

the antecedent

and

is,

that the basis of the judgment

and

that the members of the

judgment,
the consequent, are both attributive.

viz.,

The

contrast comes out in the fact that the arithmetical value of

the subject is integral, while that of the antecedent and of
the consequent is generally fractional.
46.

It

members

appears to be of the essence of an Equation, that its

This holds of the Chemical

be equal
arithmetically.

Equation.

For example,

in

Alcohol becomes Water and Olefiant Gas.

the arithmetical value of the

right-hand
the arithmetical value of the left-hand

member

is

equal to

member.

Also in

the Quaternion Equation

+ P = 7,
the arithmetical value of a -f

ft is

equal to that of

y.

+ and

The Signs

IV.

47.
is

THE

SIGNS

+ AND -

U{x+y]

By

meant
7 s which are x together with

that

which are

jy/

the result of the addition of the selective operation

x upon
to the selective operation y upon U.
Read with
respect to a single U, the expression means
is,

A U which

48.

that

Suppose

that

some of them

x or y.
collection of cubes,
have a dot on their upper face,
is

either

we have a

and
and

some have a

dash.
Let the operation
the separation of those having
and the operation y of those

x denote
a dot

Suppose x first per

formed
the collection is broken up
into two parts, I. and II., viz., those
which have a dot, and those which are
having a dash.
j

without a dot.

The subsequent opera

tion y consists of

two

parts, viz., a selec

tion of those having a dash,

first

among

those having a dot, and secondly,

among
those without a dot ; and therefore gives as a result four
x y denotes the addition
parts, i, 2, 3, 4, of the Figure,
of those having a dot to those
having a dash, whether the
first division, which consists of those
having a dot and hav
ing a dash, vanish or does not vanish.

Since the sign

is by definition the
sign of the
operation which is the reverse of that denoted by +; and
since the sign -f- has its
ordinary algebraic meaning, the
only difference consisting in the fact that the symbols joined
involve identity ; the sign
must have its ordinary
49.

algebraic

meaning, but modified by the above-mentioned circum

stance.

+ and

The Signs

U{x-y]
means
which are x minus Z7 s which are y.
destroy one another, so far as they coincide ; and
in general consists of a positive and a negative

Z/

x and y
the result

part.

50. Consider the collection of cubes represented by fig. i.

denotes that those having a dot are to be separated

x y

and those having a dash negatively. Those

cubes which are both x and
(division i) will be manipu
lated, but left out finally ; the result of the operation being
that division 2 of the figure is taken positively, and division
positively,

j>

3 negatively.
51. If the operations x and y are formally independent,
rs in the above example, the arithmetical value of

x+y

cannot be affected by any real relation existing between

and y.
x and y.

simply the

It is

sum

The

applies to the

operation y (fig. i), though subsequent to

whole of the universe ; hence

=1+1 in
The above

is

not the view

of the arithmetical values of

x,

the case represented.

commonly taken. De Morgan

The more classes aggregated,

says (Syllabus, section 131),

so long as each class has something not contained in any
of the others, the greater the extension of the aggregate

term.
52. If it were necessary for x and y in x+y to be
mutually exclusive, a restriction would be placed upon -f
which would prove fatal to the development of an Algebra of
But the imposing of the restriction would be a
Quality.

mistake, for the basis of our operations is not a Unit but a

Universe. x-\-y does not necessarily denote a single char

but what it in general denotes is equally intelligible,

;
a summation of two characters.
Similarly, it is not necessary for x in x y to be inclusive

acter

of y, or for

to be inclusive of x.

In the Algebra of Quan-

+ and

The Signs

19

these conditions hold ; they do not hold in the Algebra

of Quality, because the Universe is a generalised Unit. The
Universe becomes a Unit when it contains only one indi

tity

vidual.

53. Definition of Single.

when

does not select any

it

symbol x

member

is

said to be single

of the universe

more

than once, and always selects with the same sign.

Observation.
This embraces exclusion, and both positive

and negative
54. If

and z

is

inclusion.

we have

z=x-{-y,

a single character, then

exclusive.

It is in this

when

restriction

it is

way

x and y must be mutually

that the Algebra imposes the

Similarly, if

requisite.

z=xy,
and

be single, then the x must be inclusive of the y.

55. To express not x in terms of x.
z

By

not

is

meant the whole excepting x; hence

it

is

ix.

expressed by

The

arithmetical value of not x is i

x.
In the example
of Art. 48, i
x denotes not having a dot, and forms part
II. of the figure ; while i
y denotes not having a dash/
and forms parts 2 and 4.
56.

If x

is

positive

and single ;

then

ix

is

positive

and

single.

For x being included

positive, the result of

in
is

and i being single and

For an
single and positive.
i,

analytical proof, see Art. 120.

57. When x and y are in their statement
independent of
one another, their combination by
and
is
subject to

the formal laws

(i)
(2)

x-y=-y+x.

(3 )

58. The truth of the formal law

(i) may be seen by a
consideration of the particular instance of the cubes with

The Signs + and

2O

If we first separate those having a dot,

different marks.
then separate those having a dash, and add the two results
same cubes in the final result as
together we shall get the
when we commence by separating those having a dash, then
;

and add the two

having a dot,

those

separate

results

together.

The formal law (2) shows that independent characters,

which are not necessarily single, are subject to the same
law of combination as independent characters, which are
single.

The

formal law (3) shows that

tion of the

mind

V.
59.
is

as a

symbol of opera

quite as independent as -f

is

THE SIGNS X AND

-^

Uxy

By

meant
C/ s

which are both x and

y.

both is frequently dropped. Ac

language
expresses the signification of and ;
cording to Boole
It is, however, an
while X properly expresses sequence.
additional argument in favour of a difference of notation for
J

common

In

Universe and its characters, and generally between

and operations which are
operations which are subordinate

the

co-ordinate

expresses

that,

when

it

is

done,

expresses

or

and

and.

60. Consider, as before, a collection of cubes, Art. 48 ;

but we shall now suppose that we have

a great number. As before, let x denote

having a dot, and y having a dash.
*

Let x separate to the left hand (fig. 2) ;

and then let y separate to the bottom.

The
I

FIG.

2.

operation xy is a consequence,
being the separation of the cubes, which
both have a dot and also have a dash.

The Signs x and

61.

When

when each

x andjy

the symbols

refers to

21

-r-

are independent

that

is,

U simply the compound symbol xy

U a definite arithmetical value. This

has for any given

is not determinable from those of

value, however,

x and y

but they give limits to the value.

Consider the example of
the cubes (fig. 2).
It is evident that xy cannot be greater
(include more) than x, or greater than y ; and that it
less than (must include) x-\-yi, or o.
Hence
arithmetical value cannot be greater than x or y, and
cannot be less than
i, or o.

cannot be
its

x+y

important to draw a distinction between the

independence of the operations x and y, and the relation of
62. It

is

these to one another as characters.

The operation y
independent of x, when it is not confined by its expression
In the Theory of Probability it
Ux, but applies to U.
assumed
But

that,

this is

ratio of the

when x and y

is

to
is

are independent,

xy = x Xy.
an assumption resting on the supposition that the

Uy s

in the

ratio of the t/y s in the

x part is the same as the

It is
the not x part.
to
part

part to the

not

assumed that the two straight lines, which together separate

the y in fig. 2, form one straight line.
It is, no doubt, the
best assumption to make, when we do not know how the two
portions of the line lie ; and is of the same nature as the
doctrine that the values of two characters not being known,
We have supposed
they are to be assumed to be equal.
that the Universe contains a great number of cubes, in
order that each portion of the y line may be supposed to be
straight.

63.

Venn shows

Logic of Chance,
rule

the true nature of the assumption in his

For the establishment of the
157

p.

under discussion,

both necessary and sufficient

by each of the above
should
each
sub-divide
of
the classes created
distinctions,
by the other distinction in the same ratio in which it sub
divides the whole.
If the independence be granted and so
it

is

that the division into classes caused

The Signs x and

22
defined as to

mean

this,

-r-

the rule of course will stand, but,

without especial attention being drawn to the point, it does

not seem that the word would naturally be so understood.

The

distinction

maybe

called that

between formal and real

independence.
64.

But one character may

in

its

statement involve

another character, so as to be formally dependent on the

Let
denote the universe of objects, a any
latter.
character, and x a character which is formally dependent

on

a.

Then Uax denotes

s which are a and of these such as are x

7 s which have a which have x.

or

The symbol x operates upon

The arithmetical value of x
Hence
to Ua, not to U.

Ua,
is

not upon U.

measured with reference

65. In this case, and this case only, does the mind pro
ceed in the predicamental line. Boole supposes that it
In Uxy he considers jy to
always proceeds in that manner.
in fact, to mean Uxy and supposes that it
operate on Ux,
can preserve the same signification in Uy. But in fact its
meaning and arithmetical value may both be changed.
66. In the ordinary Algebra \ \ means one-half of a
,

The one operation is formally

dependent on the other. But we have shown, Art. 61, that
in the Algebra of Quality
third or one-third of a half.

is
does not mean one-half of a third or a third of a half, and
can be considered to be equivalent to such an expres
sion

only

when the characters are really independent of

Thus the Algebra of Quality is more general

one another.

than the Algebra of Quantity.

67. Suppose that we consider the

Then an example

of ax

Mammalia.

is

having red blood corpuscles

which are oval

The Signs x and

and an example of a^

-r-

is

having red blood corpuscles

which are
It is

circular

the function of the Relative Pronoun

and

discoid.

to denote

depend

ence.
68.

When x and y

are

of one

formally independent

another

xyssyx.
Also,

when

x,

y and

z are formally independent of one

another,
69. Since ax is equivalent to an independent character
of arithmetical value a x, the laws that are true of inde

pendent characters apply to ax as a whole.

same laws apply to characters which are

dependent upon a common character. For

70. Also, the

co-ordinately

example,

and

U{axy =ayx }.

71. Special laws.

U{ax =xa }.
Also,

if

0*+*+^+

then
Observation.

be freed of

its

refers to a

+*=*

supposed that x
dependence on a ; which
It is

noun contained

72. Deduction of the

is
is

such that

in a.

meaning of

-r

from

that of

X-

Let xy=ixz,
4being the reverse of
have such a signification that

then

is

to

xz

It is

evident from

fig.

3,

it

not possible

which repre

sents a case of
FIG.

3.

can

when

The Signs x and-^

24
that --

Hence

indefinite.

is

the

00

proper

reading

of

_xz
is

y
It is true in the

is

equivalent to an

case of

3 that

fig.

yz=x,

z=

hence

which means
z

is

equivalent to an

y
For there are evidently several z

which can

satisfy the

condition

yz=x,
although

COR.
For

x and y be kept
is

if
00

constant.

not in general

were necessarily

i,

the truth of the equation

xy=xz
would necessarily involve

that of

But an inspection of the

figure

y=z.
shows that

this is

not the

case.

yz

73. The condition

is
always satisfied by the fractions
of ordinary Algebra ; but it is not satisfied
by the fractions of
the Algebra of Quality, unless as a
This pecu
special case.
liarity

f^-

not necessarily equal to i\ was

clearly pointed

out by Boole.
It is commonly
supposed, and I think that
Boole himself supposed, that it is the
only peculiarity of the
symbols of this Algebra. For example, Professor Robertson
Smith says
There is one limitation only to our right to
:

The Signs x and--

25

manipulate logical and mathematical identities by the same

From xy=xz we cannot infer
In Transla

yz.

rules.

Now the fact

of Ueberweg, p. 569.
has a generalised meaning.
tion

74.

Boole

interprets

symbol; but

this

is

mean an

to

that each sign

indefinite

class

interpretation,

he

says,

cannot except

upon the ground of analogy be deduced from its arith

metical properties, but must be established experimentally.
Laws of Thought, p. 92. In Art. 166 I show that under a
must be

certain condition

single within a certain part of

When no

the universe.
is

meaning

condition

is

imposed upon

it, its

quite indeterminate.

75. Definition of the

x m where

Index

sign.

an integral symbol, is defined to mean

the selective operation x repeated m times.
For instance,
>

is

x2
It is to

= xx.

be noted that x may be single or complex, and

positive or negative or both.

76.

Law

of Indices.

x mx n

= xm+n

Also

I
ocr

But

xn

is

not in general

X
would be
x
2

For

if it

were,

xm~ n

=x

which

is

not necessarilv

true.

VI.

RULE OF

SIGNS.

77. If + be defined to be equivalent to +,

by the usual proof, that
2

and

it

follows

Rule of Signs.

26
Observation.

nection

_j_

is,

of course, immaterial in what con

x\ or
y), or
(
x(

-|

occurs; whether in

The

It

superposition of the two mental operations can

be considered apart from the accidents with which the

operations are mixed up in use.

INTEGRAL SYMBOLS.

VII.

To find

78.

the

meaning of

where

and n are

each

TL-

integral.

a/=

Let

nw=m.

Then

72);

(Art.

for

every

definition applies conversely.

Since

and n are each

peated a

number of

equation

cannot

m.

Then

w=

be

integral (that
it

times),

satisfied

quotient.

is

unless

Thus

is,

the whole re

evident that the latter

is

is

a divisor of

impossible unless

divides m.

COR. i.
-|=2.
\ = i.
is impossible.
COR. 2.
-|
But \x will be possible, when x = 2y. For then we have
\y ; which is = y.
Observation.
\ does not mean the same as J. For \
means the whole upon two wholes ; whereas J simply means
one-half.
79.

To find

the

meaning of

integral symbol.

First ; for

and
o

where n

is

any

27

Integral Symbols.

w= no

Let

nw=o

Then
From
be=o.

this latter

it

equation

(by definition).
is evident that

must

=o.

COR.
i

Second; for

T
Let4

w=.

n
o

&w=n

Then
Suppose

(by definition).

be integral; the

to

latter

equation

is

then

evidently impossible.

Suppose

to

it

be fractional and single

So also, when w

then also impossible.

tional

is

the equation

is

supposed frac

and complex.

Hence

is

COR.

an impossible selective operation,

Jfx=

y,

and x

is

For
but

y must

be

o.

ox=y,

x being a

80.

possible ; then

To find

possible selective symbol,

0^=0

meaning of --

the

w=

Let

*L.

Then
The

ow
latter

o.

w is an integral
any fractional symbol, whether

equation will be satisfied when

symbol; also when

single or complex.
81. Since the

is

Hence -

is

quite indeterminate.

symbols and the signs of the Algebra of

Quality have a meaning, which is a generalised form of

their meaning in the Algebra of Quantity, every theorem in

On

28

the

Equation

the latter

is true in the former,

provided that any special
conditions which have been introduced are removed.
For

example

The former equation can be deduced from the latter in

the Algebra of Quantity, because in it
abc
i always.
abc

The propositions contained in the Theory of Deter

when so generalised, are of great interest and
importance in the analysis of Quality. The ordinary de
82.

minants^

monstration of

these propositions does not necessarily

the
depend upon
peculiarities of the symbol of Quantity.
It, however, proceeds on the assumption that the symbols
are independent of one another.
83. In the preceding investigation of the

fundamental
Algebra and their immediate conse
quences, I have endeavoured to imitate the clear logical
manner in which the fundamental definitions of the Calculus
of

definitions

this

of Quaternions are laid

Treatise on Quaternions.
exact,

and

down by Professor
The definitions given

in this respect stand in

indefinite statements about

vague
be found

VIII.

84.

in

Tait in his

are perfectly
contrast to the

marked
and X which

are to

even the best manuals of Logic.

ON THE EQUATION AS EXPRESSING

A GENERAL PROPOSITION.

Any

relation

between the characters of a definite

means of an equation between

universe can be expressed by

as expressing a General Proposition.

29

The logical
the symbols which represent the characters.
equation is an analytical statement of a truth, which may
expression in language one or more general
a
statement which, from its brevity, clear
propositions
is of the greatest aid to the mind when
and
exactness,
ness,

require for

its

Whether other logical rela

reasoning from complex data.
tions besides those of Quality can be stated more fully and
concisely by means of symbols and the equation than by
means of words and the proposition, is a different question.
85. Every general proposition refers to a definite universe ;

which

is

series of

the subject of the judgment, and,

men

All

No men
86.

It

is

same

are mortal,

are perfect,

said, for

iv.),

instance by Boole

(Laws of

that

men

All
is

be, of a

universe.

often

Thought, chap,

may

men.

refers to the universe

refers to the

it

For example;

judgments.

are mortal beings,

a more complete statement of the judgment

All men are mortal,
events, expresses the same judgment of the mind.
in the one case, the subj-ect of discursive thought is

or, at all

Now,
the

class

in the

beings;

judgments, which apply to

other, the

class

men.

Can

different universes of objects,

be

In considering whether, in the universe of

men, the attribute mortal is equivalent to the whole,
I do not require to consider beings in general.

the

same?

87.

Let

U denote men,
All

is

and x mortal; then

men

are mortal

expressed by

U{i= X }.
But

if

the judgment

is

Beings which are

let

U denote beings,

human are
y human

mortal,

U{y=x-w}

mortal
;

then

On

30

the

Equation

denotes an unknown character, the complement

(where
of y to x) expresses the judgment considered.
Another form of the equation is

Again

to express

No men

are perfect

U{o=p}.

And

to express

Some men

are wise

U{v=w],
where v
88.

is

a proper fraction, and greater than o.

write these equations, dispensing

We may

abbreviations, which is indeed a very good

soundness of a notation.

Men
Men
Men

test

with

of the

= mortal}.
=
{none perfect}.
{all

{some=wise}.
These are precise and undistorted representations of the
judgments actually expressed by
All

men

No men

are mortal.
are perfect.
are wise.

Some men
89.

We may

consider the analytical expression of those

examples of general judgments, which Aristotle discusses in

his

book on

Interpretation.
7TOS

ov

(TTIV

TTOLS

ecrriv

7TOS 4(TTiV

ov Tras ecrrtv

ai/0/9t07ros

ov

The common

universe of these propositions is CCI

and the quality considered is St /ccuos. The equations are
respectively

1=3.

(!)

iv=S.

(2)

= i-S.

(3)

as expressing a General Proposition.

iv denotes all minus some, some being greater than nought

which

equivalent to not all.

90. It is evident that the grammatical subject of the pro
The
position does not form one member of the equation.
is

members of the equation may be the quantity and the

predicate ; as in the above examples, and in judgments
such as

Men {^-= shortsighted}.

and consequent as in
which
are
equilateral are equiangular,
Triangles

Or, they

may

be the antecedent

T{x=y

w}.

may contain one, several, or an

infinite number of individual parts, and since a judgment
91. Since

a universe

always refers to a universe, it is evident that the judgi.ient

be modified in an important manner according to the

will

number of

individual parts in its universe.

The strictly
singular judgment refers to a universe, which has only one
individual part \ as, for example,

A given man
It is to

be observed

latter to

make

at

a given time

is

sleeping.

that, in order to

produce a singular
universe, it is not enough that the substance be fixed ; the
time must also be fixed.
92. When the judgment refers to a singular universe, the
only values which any character can assume are i and o.
The former is required to make an affirmation, and the

The

given

a negation.
at the given time

man

,,

is

sleeping.

is

not sleeping.

is

not-sleeping.

is

not not-sleeping,

are expressed respectively thus

i=..
o=j.
1
1-*.

0=1

j.

(i)

(2)
(3)
(4)

On

32

the

Equation

remarks that here equation (3) follows

(2) ; but that in the more general case (Art.
is plural, equation (3) does not follow from
It is only judgments of the kind considered

93. Aristotle

from equation
89)

when

equation (2).
in the preceding article that are affirmative or negative in
the strict sense of these words.
94.

There are singular judgments of another kind.

example

For

Of all

the planets Jupiter

the analytical expression of which

Planets

{ Jupiter

is

the greatest

is

= the

greatest}.

a singular character, so is the greatest; that

Jupiter
can
is, they
apply to only one of the individual parts of the
Hence the singularity of this kind of judgments
universe.
is

does not consist in the universe of the judgment being an

individual, but in the characters being singular.
95. It

is

also important to consider the special nature of

the judgment which relates to a universe having an infinite

number of parts. For, questions arise as to how it is
possible in
characters.
96.

such a case to

As the relations of

among

the

the characters of the universe

may

assert

relations

be given in an infinite variety of ways ; the logical equation,

in order to be adequate, must be capable of assuming an
The forms treated in manuals of
infinite variety of forms.

Not only ought

Logic are but a few of the most common.
to be able to state explicitly what is thought implicitly,
but we ought to be able to state in one equation any one

we

number of characters satisfy; even though

number of sentences to express that con

condition, which a
it

take a great

dition.

97. The reason why Formal Logic has so long been

unable to cope with the subtlety of nature, is that too much
Arithmetic
attention has been given to pictorial notations.
could never be developed by means of the Roman system

as expressing a General Proposition.

of notation

33

and Formal Logic cannot be developed so

is represented by

long as Barbara

C,
or even

by the simpler spicular notation of

De

Morgan.

We

cannot manipulate data so crudely expressed ; because

the nature of the symbols has not been investigated, and
laws of manipulation derived from their general properties.
98.
consideration of the nature of propositions about

abstract numbers is of importance here, as throwing light

upon the nature of the logical equation. Kant appears to

be the only logician who

He

has

The

discussed

them

at

all

self-evident propositions as

says
thoroughly.
to the relation of numbers are certainly synthetical, but not
universal, like those of geometry, and for this reason cannot
:

That

be called axioms, but numerical formulas.

For neither in the repre

not an analytical proposition.
sentation of seven, nor of five, nor of the composition of
But
the two numbers do I cogitate the number twelve.
although the proposition is synthetical, it is nevertheless
is

In so far as regard is here had

only a singular proposition.
merely to the synthesis of the homogeneous (the units), it

cannot take place except in one manner, although our use

of these numbers is afterwards general.
Critique of Pure
Reason.

The

99.

where

full

C/"may

expression of the example discussed

denote any unit or any universe.

For

it

is

is

a logical as well as an arithmetical proposition.

^{?X5 =

Similarly

whatever
It is

The

35}>

7 be.

evident that the source of the nexus

peculiarity

is,

that

the

is

equivalence

not convention.
is

true

inde

pendently of the nature of U.

100. It

is

also of importance to consider the nature of the

On

34

the

Equation

chemical symbols and of the chemical equation.

considers a universe of material molecules ;
bols as

H and Cl denote a

The chemist
and such sym

distinctive character or set of

characters possessed by some of the molecules.

It is a
property of these characters, that they are exclusive of each
other.

HCl

H and

does not denote matter which has the pro

; but matter which has

also the property Cl

distinct
entirely
property

perty

HCL The meaning of HCl

formed by the combination of Zfand Cl? The symbols
have fixed fractional values that is, fixed proportions ; in
this respect differing from the symbols of Quality.
an

is

101.

The chemical equation

expresses

how

the distinc

tive characters of a particular set of molecules are

into other distinctive characters,

For example

and what the

changed

latter are.

Sulphate of potassium and chloride of barium become

sulphate of barium and chloride of potassium,
expresses that in a particular quantity of matter the
of distinctive characters

becomes changed

into the

sum

sum

of distinctive characters

The symbol = involves becoming. (Professor Crum Brown s

Manual of Chemistry, p. 46.) Hence the chemical equation
cannot be reversed without reversing the meaning of the
copula =. In De Morgan s language, the copula here is
inconvertible.

102.

The above equation expresses completely the change

it refers.
Its right-hand member is as definite as

to which
its

left-hand

member, and the two are equal

In particular cases we

may

arithmetically.
require only the incomplete

equation
just as instead of the complete logical equation

as expressing a General Proposition.

we may
There

in particular

35

circumstances consider only

an

important difference between what is requisite

to give a complete scientific expression of a
change or of a
relation, and what is requisite to give an expression of such
is

a part of that change or relation as

particular process of reasoning.
103.

The

may be

necessary in a

equations of Natural History are examples of

The dental equation of Man is written

logical equations.

pm 2

-Z

m^

-=7,2.
6

3
3
vJ
as a logical equation
<J

But

this truth

may be written
i=/ ~ c
(

Man]
(

irJ
i-i

pm 2 ~ 2 m
2-2

thus

3^3
3-3,

that is, considering the universe of men in the adult

state,
the whole is identical with those
having incisors which are
in

number and arrangement

are in

2
,

having canines which

number and arrangement -

which are
which are

in

in

number and arrangement

number and arrangement

characters

-2

!_!

_,

the character

having premolars

*~ 2
,

3-3

having molars

Each of the

must be equal to

etc.,

1? is

dependent with respect

to

i.

/,

In

and

composed of four co-ordinate characters.

104. If we consider the universe of
Mammals in the
adult state ; to what is the character
is itself

2-2
2

equivalent

We

have, I think,

The Principle of Identity and

36

v
Mammalia

bimana

quadmmana
catarhina

^2/1^* 2-2
22 33
"3^3

and, if so, we have a statement which

exact as any chemical equation.
105.

Professor

Huxley,

in

as complete

is

and

argument against the

his

assertion that Biology differs from the Physico-chemical

and Mathematical sciences in being inexact, holds that

the statements

A vertebrated

animal has jaws which never open side

ways, but always up and down,

and

An

annulose animal has jaws which always open side

ways, and never up and down,

are as exact as any of the propositions in Euclid.

Value

the

Natural History

Sciences.

of
Essays and Reviews, p. 79.)
The above statements are exact,

vertebrated

Animals I

annulose

IX.

being

thus

havin S J aws

havin

opening sideways

X opening up and down.

never opening up and down.

never opening sideways

the

Lay Sermons,

in the sense of

capable of expression by exact equations

Animals /

(On

aws

THE PRINCIPLE OF IDENTITY AND THE

AXIOMS OF IMMEDIATE INFERENCE.

106. Let

U denote

any universe; and x any character,

positive or negative, integral or fractional

U{x =

x}.

then

the

Axioms of Immediate

37

Inference.

The Principle of Identity, when put into the above form,

does not appear to merit the abuse which has been heaped
upon

is

and

Ens
Another form of

est Ens.

this principle is

which involves the meaning of the sign

107. An inference is said to be immediate, when it is
deduced from an equivalent equation and an identical
equation (or the former repeated) by means of the axioms
.

of the science.

has been a matter of perplexity to logicians,

It

conclusion can be drawn from one premise.

how

Ueberweg

The immediateness in the so-called immediate

says
inference is relative.
It implies that this kind of inference

"

"

does not require, as mediate inference does, the addition

first, but at once and of itself

of a second datum to the

yields the derived judgment, which is nevertheless another

judgment, and not merely another verbal expression.
s

(Lindsay s translation of Ueberweg, p. 226.)

It is not very evident, how one equation can at once and
of itself yield another, which has any claim to be called

But if we consider that an identical equation

can be formed, and that the axioms of the science, which
depend upon what is essential in the symbols, allow us to
different.

combine these equations

a great part,

if

not the whole,

of the mystery disappears.

The general axioms of immediate inference are

If an equation and an identical equation be
together ; the resulting equation is true.
108.

I.

added

Thus

if

since

II.

x=y;
z=z,

If

an equation and an identical equation be sub-

The Principle of Identity and

38

tracted the one from the other

the resulting equation

is

true.

Thus

x-y\

if

since

z=z,

xz=yz.
If an equation

III.

tiplied together

Thus

and an

identical equation be

the resulting equation

mul

is true.

x=y

if

since

z=z,
xz=yz.

IV. If an equation and an identical equation be divided

the one by the other ; the resulting equation is true.

Thus

x=y;
z=z

if

since

A common form of application of Axiom II.

109.

in subtracting the given equation

1

For example

1.

the universe being Triangles,

= equiangular
equilateral =
equiangular.
equilateral

consists

from the identical equation

one form of the process of Contraposition.

This

is

110.

An equation, when multiplied by a fractional symbol,

but it is not so general as before.

III.)
of
an
multiplication
equation by such a symbol means
the introduction of an hypothesis.
proposition, which is
true independently of an hypothesis, is true under the hypo
is still

true

(Axiom

The

thesis; but not conversely.

111.

The axiom

that

When

the signs of the terms of an equation are changed

into their respective converse signs, the resulting

equation
reduces Axioms
respectively.

is

true,

II.

and IV.

to cases of

Axioms

I.

and

III.

the

Thus

Axioms of Immediate Inference.

x

if

x=

then

y,

+ is changed into its converse

_i

where

39

and

also

x~~ y

where

changed into its converse

112. Given the principle of Identity
is

x
to

equivalent to

is

-f

prove the principle

is

equivalent to not not x.

x=x,

Since

and

0=1

x= i

+x.

=
x)
= not not x.
i

1,

Axiom of Immediate

II.

Inference.

by Rule of Signs.

(Art. 55.)

COR. i #=1
x)}.
{i
(i
In the case of x we have an even number
Generally.
of I s alternately positive and negative ; and in the case of
i
Xj an odd number of I s alternately positive and nega
}

It is customary with grammarians to point out, that

an adjective, with an even number of nots before it, is
tive.

equivalent to the simple adjective.

x= i

Observation.
is

true whatever

as above, to

When

be

U is

7 be; and

x\

also without

x being

restricted,

single.

singular,

x can only be

x=i
then becomes what

is

called

or o (Art. 92)

and

x)

(i

by some

the Principle of

Con

tradiction.

Given the principle of Identity

113.

x
to prove the principle

equivalent to x,

is

of Excluded Middle

+ not x is equivalent to the whole.

.

x=x,
x x=o

but
.

^+1^=1.

I.

Axiom

1,

of Immediate Inference.

The Principle of Identity,

4O

x + not x is equivalent to the whole, whatever

7 may be singular, plural or infinite.

That

is,

7 be.

The

etc.

Excluded Third, as enunciated by many

is singular.
For example ;

principle of

logicians, presupposes that

Ueberweg states it thus (Translation, p. 260)

Judgments opposed as contradictories (such as A is B,
and A is not B) can neither both be false nor can admit
1

the truth of a third or middle judgment, but the one or the

other must be true.

Let c^be singular, then

U{x+i-x=i}.
Now when 7 is
consequence

x can only be i or o; and

x can only be o or i. Hence either

singular;

in

U\x=i]
or else

U{x*=o}

must be

When

true.

is

plural,

it is

or else

of course not the case that either

U{x=o}

must be

true.

x-{-ix=i

Observation.

does not necessarily presuppose any restriction upon the

signification of x such as to be single.
}

X.

AXIOMS OF MEDIATE INFERENCE.

114. A mediate inference maybe defined as, an inference

from two or more independent equations by means of the
axioms of the science.

115.
I.

true.

The

general axioms of mediate inference are

For example

if

x=y, and a=.b ;

then

If the equations be added, the

resulting equation

x+a=y+b.

is

Axioms of Mediate

two equations be subtracted the one from the

the resulting equation is true.
For example ; if

If

II.

other

41

Inference.

x=y, and a=b;

x a=y b.

then
III. If

two equations be multiplied together

ing equation

the result

is true.

x=y, and #==$/

then

ax=by.

IV. If two equations be divided the one

by the other
the resulting equation is true.

x=y, and

a=b ;

^ = Z.

then

not necessary, that the two equations should

contain a common character ; but it is necessary that both
should refer to the same universe
just as the ordinary
116.

It is

a given investigation must be sup

posed to refer to the same unit. In the latter sense only is
it true, that no
reasoning is possible without a middle term.
algebraic equations in

117.

axioms.

Kant would not consider the above to be true

He says ( Critique of Pure Reason, Meiklejohn s

translation, p. 124)

If equals be

The

added

propositions

to equals, the wholes are equal,

and
If equals be taken

from equals, the remainders are equal,

are analytical, because, I am immediately conscious of the

identity of the production of the one quantity with the pro

duction of the other

whereas axioms must be a priori

But these propositions are not
synthetical propositions.
analytical, in the sense of depending upon convention for
their truth

;
they are true of the symbols of Quantity, as
symbols of Quantity ; and only in that sense may be said
to be analytical.
An axiom appears to be precisely such a

proposition.

Conditions for a Character being single.

42

XL CONDITIONS FOR A CHARACTER BEING

SINGLE.
If x

118.

is single

and positive,

then

x*=x;
and

if x

is

single

and negative,

x2
and

then

=x:

conversely.

For the definition of

see Art. 53.

single,

First

when x

is

positive.

Let y be another single and positive

symbol, and let y be wholly included

in

Then

(fig. 4).

xy=y.
-

Let>

expand

until

FIG. 4

that

when x

Second ;

Let

if

Hence,

By a

is

is

is,

negative.

x=

single

yx

xx =x.
x 2 =x.

then

x being

=x
=

x.

positive.

by first part.
by Rule of Signs.

consideration of the figure

it is

also evident that if

zz^Xj

x must be

single and positive j

x*=
and that if
x must be single and negative.

x 2 =^ix

Hence
is

x,

the necessary and

sufficient condition for

x representing

a single character.
119.

and

if

More

is

generally

single

If x is single and
xm =x ;

and negative,

x=(

i)

m~l

x.

positive,

Conditions for a Character being single*

is

For

(i

120. If

single

positively.
2

-x)

It is

when

x is

is

= - 2x+x*.
is

single

(Art. 68.)

= ix
.-.

i-x

then

positively,

43

by given condition

single positively.

single positively, that

it

(Art. 118.)

means the same

as not x.

121.

Given that

U{x =x}
U{x ( i x) = o}.

to prove that

X 2 = X,
x 2 =x 2

and
therefore

Principle of Identity.
Axiom of Immediate Infer

x^x

x2

x2

II.

ence
that

o=x x2
o=x (i x).

is,

therefore

Distributive

Law,

Art. 68.

Another -Proof.
x-\- 1

therefore

x= i,

Principle of Excluded Middle.

Principle of Identity.
III. Axiom of Immediate In-

x=x,
x (x-\-ix)=x ;

therefore

but
therefore

x z -\-x (i -.#)=.#
x 2 =x,
x ( i x) = o.

ference.

Distributive

Law.

given.
II.

Axiom

of Immediate Infer

ence.

122.

Meaning of these

equations.

means

U s which are x and x are identical with those which are x.

It is true

only

when x

is

positively single with reference

to U.

Since

(Art. 120).

is

single positively,

Hence

ix must be single positively

x means not

x.

Hence

U-{x(i-x) = o}

means
fs which are x and not x are none.

Conditions for a Character being single.

44

Since
is

the condition for

verse

is

U;

x being

single, relatively to the uni

U{x (i#) = o}
mode

the general form of the other

of stating the con

dition.

123.

When

be single

7"

is

(Art. 92).

any fractional character x must

Hence, when that condition is satis

singular,

fied,

and #
are true always.
The latter
expression of the Principle

is

then the exact analytical

It is impossible for an object at a given time to have

a quality and the contrary quality.
This is what is commonly called the Principle of Contra
diction ; but the most general form of the Principle is that

7 may be

enuntiated in Art. 122, where

singular, plural, or

infinite.

124.

As the deduction and

interpretation of

x (ix) = o

of great importance in the Algebra of Quality, I shall

quote the original proof given by Boole.
(Laws of

is

Thought,

p. 49.)

That axiom of metaphysicians which

is

termed the prin

and which

ciple of contradiction,
affirms that it is impossible
for any being to possess a quality, and at the same time not to
it, is a consequence of the fundamental law of thought,
whose expression is x^=x.
Let us write this equation in the form

possess

X^

=Q

whence we have

(ix)

=o,

both these transformations being justified by the axiomatic

Let us, for sim
laws of combination and transposition.
plicity of conception, give to the

symbol x the particular

Conditions for a Character being single.

45

i
x will represent the class of
the formal product of the expressions of
two classes represents that class of individuals which is

interpretation of men, then

Now

not men.

common
class

to them both.
Hence x (ix)
whose members are at once men

will represent the

and

not men.

Now the

formal product of the expressions of two classes re

that
class of individuals which is common to them
presents
both.
Hence x (ix) will represent the class whose

members

men and

are at once

not men, and the latter

equation thus express the principle, that a class whose mem

bers are at the same time men and not men does not exist.
In
other words, that it
same time a

be at the

is

impossible for the

same individual

to

man and not a man.

125. With reference to the above deduction, the Rev.

Robert Harley observes as follows (British Association
From the logical equation x 2 =x the
Report for 1866)
2
x
x
=o
is
derived by subtracting x 2 from both
equation
members, and the result is put under the form x (i-x) = o
:

by the law of

It is to

distribution.

be observed, however,

that at every step of the process the principle of identity

is assumed, and in Boole s interpretation of the final

x=x

result the

x without

same

principle
the brackets

is
is

used, for
identical

is implied that the

with the x within.

it

Further, in the final interpretation not only is the principle

of contradiction employed, but the principle of excluded

middle

is

mean not

also employed.
x,

it

is

tacitly

For in interpreting i x to
assumed that every one of the

things of which the universe represented by unity is made

up is either x or not x. It would thus appear that these
three principles of identity, contradiction, and excluded

middle are incapable of being reduced to more elementary

They are axiomatic, and Boole made use of them

truths.

unconsciously in framing his laws of logical interpretation.

126. Boole s investigation can be defended from the

above criticism; but the defence requires that the definitions

and axioms of the science be laid down in an exact manner
\

4.6

Conditions for a Character being single.

as I have endeavoured to
I

consider

shall

how

do

is

Principles.

x denotes a

investigation, and shall assume that

referred to a definite universe.

127. It

in these

the above remarks affect

Hence
my own
character

true that the Principle of Identity

x=x
is

assumed,

in the sense assigned to

it

in Art.

Mr.

106.

Harley considers it to mean that the one x as a symbol is

identical with the other x as a symbol ; whereas it must
mean that the result of the operation denoted by the one

symbol is identical with the result of the operation denoted

by the other symbol.
In the proof that, under the given condition of

x*=x,

ix means

not x

Middle brought

in

(Art. 120), the Principle of

x=i

x+i
which

is

Excluded

is
j

only another form of

x=x.

(Art. 113.)

be observed that the Principle of Excluded Middle

quoted. in the second proof of Art 121, is that general form
of the Principle which is mentioned at the end of Art. 113.
It is to

The
Principle of Contradiction is not brought in.
and
the
follows
from
the
Distributive
Law,
proof
given
z
condition put into the form x
by means of the
The

x=o

Principle of Identity
Inference.
128.

and the Second Axiom of Immediate

What Mr. Harley seems

Contradiction

is

to

mean by

the Principle of

the truth expressed by

x=i

(i

x).

That principle has been deduced from

x=x
and

is

distinct

(Art. 112)

from

x (i-x)=o,
the one being an identity, and the other a condition.
129. With reference to the same deduction,

Venn

Conditions for a Character being single.

47

observes (Mind^ vol. i. p. 491):

Boole says that the
axiom which is termed the Principle of Contradiction, and
which
has been commonly regarded as the fundamental
axiom of metaphysics, is but a consequence of a law of
thought mathematical in its form," viz. the law whose ex
"

x z =x.

This is doubtless a very elementary

but to regard it as the source of the Law of Contradic
tion surely argues a strange inversion of order.
However
is

pression
truth,

be regarded, nothing can well be considered more

We could not distinguish one thing from another
without it ; we could not, even to go no further than these
symbols, distinguish x from what is not x without making
use of it.
And yet Boole gives a demonstration of their
dependence, a demonstration every step of which demands
that law

ultimate.

the law several times over.

However

130.

Boole

well

this

doctrine, where

no

criticism

distinction

avail

may
is

against

drawn between the

and the operations of thought ; I think that it is

avoided by the investigation in these Principles.
But is
that, by which we distinguish one thing from another, the
subject

Principle of Contradiction

It

does not appear to be the

principle considered in Art. 122.

131. Boole, as above said, considers

x*=x

to be a

law of

He

also directs attention to the circumstance,

thought.
that the equation which expresses that fundamental law of
thought is of the second degree.
(Laws of Thought, p. 50.)

He

ventures to assert, that,

if

that circumstance

had not

whole procedure of the understanding would

have been different from what it is. In reply to a possible
existed, the

objection, that the existence of the equation x*=x neces

sitates the existence of the equation x z =x, which is of the
third degree, he answers
4

The equation x*=x

of logic.

For writing

it

is

not interpretable in the system

in either of the forms

48

Conditions for a Character being single.

we

see that its interpretation, if possible at all,

that of the factor i-}-x, or of the factor

must involve
i

The

x.

not interpretable, because we cannot conceive of

t
the addition of any class x to the universe i ; the latter is

former

is

i is not subject to
not interpretable, because the symbol
the law x (i
o, to which all class symbols are
x)

subject.

Our

132.

thought

selective

position is, that

as the mind can

symbol

x,

x z =x does not

express a law of
a

comprehend with equal ease

which does not

satisfy that condition.

simply the condition for the character being single and

If x is single negatively ; then the equation
positive.
It is

x^=x,
is

actually true.

The

factor i-\-x, considered as a

character, is as interpretable as
does not satisfy the condition

pound

com

x ; and though

x*=x,
it

satisfies

The
and

the condition

source,

whence

it

arises that

sufficient condition for

tf

x being

= =h*

is

single,

is

the necessary
not the mind,

but the actual constitution of things so that, being a law

of things, it ought to be made a rule for the intellect.
;

133. Such, it appears to me, are the relations existing

between the properties of the symbol of Quality, the

Axioms of Operation, the forms of the Principle of Identity,

and the forms of the condition which the symbol of Quality
must satisfy in order to be single. A consideration of these
is of great speculative importance ; for they clearly
that the language of Mathematical Analysis is that

relations

show

language whose existence has been dreamt of

some
philosophers.
by
134. Leslie Ellis remarked that x 2 =x is not true, when
The arguments,
x denotes a relative term, such as father.
which accompany this statement, appear to show the exist
scientific

ence of a supposition, that the organon for reasoning about

Conditions for a Character being single.

49

the characters of a definite universe can be applied to all

departments of reasoning. It is thus of the greatest im

portance clearly to define the nature of that operation,

which the Algebra of Quality denotes by x ; as it is to such,
and such only, that its method applies.
135.

When

a logical equation is given which involves x,

the equation, if we know that

we can eliminate x from

x 2 =x, or x*=x. (Art.

But this must be looked

286.)
upon as another equation as expressing another relation
which the symbols must satisfy. Hence, it is no doubt more
accurate to say that the elimination of n symbols requires

n+i

independent equations to be given, than to say (with

Boole) that no relation whatever can be proved to obtain be
tween the number of terms to be eliminated and the number
of propositions from which the elimination is to be effected.
136. Given that x and y are each positive and
single, to

find the condition that x-\-y

may

be single.

x+y,
that

=x+y,

x+2xy-\-y

is,

.-.

that

See

fig. 5,

(Art. 79.)

Art. 148.

137. Given that x and

find the condition (i) that

and

by given conditions

2xy=o,

xy = o.

is,

(Art. 118.)

(2) that

xy may

and single ; to
may be single and positive ;
and 7iegative.

are each positive

xy

be single

(x-yY = x-y, (Art.

x 2 -}-y 2 2xy=x y.

(i.)

118.)

.*.

that

is,

that

that
(

2 .)

x+y

is,

2xy=xy,

(yxy) = o.

y (ix) = o.
(*-y)*=-(x-y),
(Art.
x-\-y2xy=x-\-y.
is,

118.)

.-.

The

by given conditions

(xxy) = o.

x (i-y) = o.

truth of these results can

be seen from

fig.

5.

Conditions for a Character being single.

5o

If each of the factors of a product

138.

must

be single.

Consider the term xy

Then

(*y)

where x* = dcx,

=*y.
=^dby,

Hence xy must be single.

COR. If x and y are each

139.

the product

2
andj>

=j>.

by given data.
by Rule of Signs.

= ^zxy,
must be
For i

is single,

single

and

positive,

(i

y)

single.

y is

If xy

single (Art. 120).

is

single,

and x

single ;

not necessarily

is

single.

xy;

(i)

From
.-.

but

is

and.* 2 =*.

from

(2)

not necessarily

(2)

x*y*=xy,
xy*=xy,

(i)

i.

(Art. 72.)

oc

Hence y

COR. The same

is

not necessarily equal to y.

true,

y=

when

a case can be

in such

that

is

be

single within x.
140. If x and y are each single

y*
x,

andy==y

y
xm

y
COR.

i.

=x
y

Let^=^.
Then

Hence

single

(Art. 119);

therefore

in the

^=i.
xn x

Algebra of Quality
V?^*

"

xn

Boole assumes
only.

and positive,

xm _ x

For x m

not in general

=x m

~".

then

It

must

Conditions for a Character being single.

COR.

2.

Since

is

single

and

and o

positive,

single

and

positive,

Observation.

141.

ff

and

single

and

(x+y y z* = (x+y) z,
(Art.
x z* + 2xyz* +y*z* = xz +yz,

find

(i)

is,

that

xz+2xyz-\-yzxz-\-yz.

is,

The

Observation.

118.)

that

xyz=o.

condition

xyz=o

does not necessarily

xy=o.
(x-y)*z* = (x-y) z.
.-. xz
2xyz+yz=. xz

(2.)

.:yz(ix) = o.

(x-yY

(3-)

z*

= -(x-y)

z.

=xz

xz
.

(Art. 118.)

yz.

(Art. 118.)

2xyz -{-yz
+yz.
xz (iy) = o.
.

142. ffx,y, z, u are each positive and

single
conditions for (x-\-y) (z-\-u) being
single.

(z+u).
that

is,

to

find the

(Art. 118.)

xz+xu+yz+yu+2{xzu+yzu+zxy+uxy}
that

since

to

negative.

(i.)

Hence

for (x+y) z being single

; (2) the condition for

z being single and positive ;
the
condition for (xy)z
(3)

being single

involve

See Art. 167.

x, y, z are each positive

the condition

(xy)

=(}".

is,

xzu=o

xzu+yzu+zxy -\-uxy =o.

yzu=o

no term can be negative.

zxy=o

uxy=o}

Conditions for a Character being single.

52

143. Ifx^, x^,

x3

and positive ;
and positive.

are each single

find the

to

condition for x^-\-x^-\-x z being single

*!+#, +# 3

144. If x

is

positive

and

(Art. 118.)

not negative ; then

x*x=w,
where

w denotes a quantity
that

Suppose

which

x and x 2
l

Then
and 2x l x z

(x

and not

is positive

negative.

two positive single parts

consists of

-\-x^-(x

-\-x^

= 2x x
l

and not negative. (Art. 138.)

n
Similarly it may be shown that when x consists of
2
and
x
x
is
the
positive
expression
positive single parts,
is

positive

not negative.

z,

+x where x is single, and x lt x^,

and not negative; then x^ x z x z x are
and every sum of two, and every sum of three, is

7fx=x

145.

-{-x 2 -{-x 3

4L

each positive

each single

also single

x is single, and
number of positive

since

equivalent to the

is

it

sum

of a

must be positive therefore

2
(^i+^ 2 +^;+ * 4 ) = *i+* 2 +* 3 +*4. (Art. 118.)

i.e.

Now

Xi

+^

terms,

2
2

it

+^3 +^4

+ 2^ ^ + ...+^3^4}

=Xi + Wi, where Wi

(Art. 144),

is

positive

2/ 4

But each of these terms

be=o.
= o, and
Hence
So are x x 3 x.
/!

and not negative

^2 2 =

and

is

+2

positive only

therefore

x^=xi

hence each must

that

is,

is

single.

Conditions for a Character being single.

=o

Also x 1 x 2

therefore

therefore

x^x^+x^

The proposition

Xi+x2

is single.

+ X X = 0;

AISO XiX z -\-XiX*

53

(Art. 136.)

is single.

(Art. 143.)

of this Article enables us to draw deduc

from that form of equation which Boole discusses

tions

very frequently

w=A+o+ o C+ o D.
THE

XII.

SIGNS OF INEQUIVALENCE
and

146. Definition of greater than

less

>

AND

<.

than.

x>y

denotes that
y<x

x y

is

and

positive

denotes that

positive

only.

negative and negative

are each single and positive,
is

j>

COR.

If

x and y

then

means

that the

means

that the

is

only.

x>y

includes the

and

Hence

y;

y<x

included in the x.

147. Axioms.
If

I.

For
which

x>y,

and a denote any character; then

x+a(y+a)=xy,
is

given to be positive only.

Similarly

it

follows that

x
II.

and

If

x>y,

a.

a>y

and a denote any symbol which

positive only

is

positive

then
ax>ay.

ax

For
but x

is

positive only,

a
If a

is

ay=a (x

(xy]

and a
is

y) ;
is

positive only, therefore

positive only.

negative and negative only

ax

<

ay.

then

(Art. 77.)

The Signs of Inequivalence

54

axay=a (x

For
but x

is

(xy)

is

and

<

and a negative only; hence

positive only,

y)

>

negative only.

(Art. 77.)

In what follows, unless the contrary is stated,

the fundamental symbols employed are supposed to be
Note.

single

and

positive.

DIVISION.

XIII.

To find

148.

divided by any
x, y,

primary parts into which a

number of independent selective

universe is

the

operations

z, etc.

Suppose we have two symbols x

and y.

x-\-(ix) = i,
y+(iy) = i,

Now

KO-jj)

and

multiply these two equations together,

Since
FIG.

x and y

parts

5.

xy

ixiy

xiy, ixy,

are each positive (Art. 120).

to the whole.

are each positive, the

Also their sum

is

equivalent

is positive, and their sum single,

of
one another (Art. 145) that is,
exclusive
must
be
they

Since each of the terms

each term, each sum of two, each sum of three is single.

These then are the primary parts of i, when we have two

independent operations x and y.

149.

which

There are other

is

Fig. 5.
sets of positive parts, the

equivalent to the

whole

fewer terms than the primary

set.

For example

since

xy+x(iy)=x;
where we have three
150. Suppose that

positive parts.

we have

sum

of

but they each consist of

three symbols, x,

>

z.

Division.

xy+x(iy)-\-(i

Multiply

by
then

It

x)(iy) = i

x)y+(i

z+(iz)=i

xyz+xy(iz)+x(i

y)z+x(i

y)(iz) + (i

x)yz

may be shown,

as before, that xyz, xy (i

z), etc., are
that they are exclusive of one another.
they are the primary parts, into which the universe

each positive

Hence
is

55

and

divided by the three independent operations x, y,

so for any number of independent symbols.

z.

And

The mode,

151.

in

which three independent symbols

by means of a

divide a universe, can be well illustrated

The cube is sup

posed to contain a large
cube.

number

The

of

individuals.

selection

of

the

according to the
character x breaks up
units

the cube into two parts,

which are represented in
from
fig. 6, as separated
one another by a single

The
according to y

plane.

selection

FIG.

6.

again breaks up each of these two portions

into two portions ; the separating partition consists of two
parts,

which are not necessarily

the value of y

is

fixed (that of

in

one plane.

x being supposed

Though

to be fixed

already) the two parts of the y partition are still free

When
to vary, provided they cut off a sum equal to y.
in addition the value of one of the four parts, as xy, is
given, the values of the other three follow necessarily.
The selection according to the third character z breaks up

The partition separat

the whole into eight compartments.
When the
z
consists
of
four
z
from
the
not
the
parts.
ing
value of z is fixed and of xz, yz, and xyz, then the eight
The value of 7
parts are fully determined.
ing condition required to make up the eight.

is

the remain

56

Division.

152. The number ofparts, into which a universe is

primarily
divided by n independent operations, is 2 n
For the parts, as above shown, are obtained
multi
.

by

each of which consists of two terms ; and there is

one multiplier for each symbol; hence when there are n
pliers,

symbols, there are

153.

Any

that

is,

154.

parts.

single positive symbol, a,

same manner as
For
and
therefore

may

be divided in the

i.

a(x+ix)=a,

a(y-\-iy)=a,
a*{xy+x(i-y) + (i-x)y+(i-x)(i-y)}=a*;
a{xy+x(i-y) + (i-x)y+(i-x)(i-y)}=a.

The

tree of

Independent Division may be repre

sented thus.

U
I

and so

12

12

ix
2

12

12

This form is exact and much more general

than the tree of Porphyry.
Porphyry s tree is a division of
the particular universe of Substances, and is a mixture of
the formal and the material.
Only the first two
on.

parallel

rows are given

which correspond to

U
x

ix

12
the remaining part of the tree being left
undeveloped.
155. Formal Division, when thus treated, is not so un
it is commonly
supposed to be. Were it kept
mind by men of science when classifying a given group of
objects, according to a number of independent characters

important as

in

Division.
aid gain in
>ught

be

mt

and exactness.

clearness

to be

made

then such parts,

out of the material division.

first

left

57

remember what

to

and what

can,

the Algebra of Quality.

By

it

we can

can

tell us what, in any given case,

or enable us to state any relation
But there is nothing to prevent the

supposing certain relations to be given

proceeding to show what necessarily
succession of throws with two coins.
the

first

the

-/!

h^ having the

up,

and

second

having the

/2

from the nature of the throw,

!,

their

tail

first

^A,

srs

head up,

hjz, t^h^ A/2 are thus subject

is equivalent to the whole.

sum

completely, three
sd that

independent data

having the head of the

first

the head of the second coin up are

cters, and therefore must be denoted

The reason why they

h
e

in the

Theory of

arithmetical values,

are denoted

Probability,

when

is,

is suffi-

al.

yt

conditions, sufficient
are not given, we

to

may

determine
consider

issumptions to make what are the

In making
all lead us least astray.

Jww*
VWM
^-V
1

Division.

58

these assumptions we are guided by our knowledge of the

nature of
and its characters h^, h 2 etc.

The

may be capable of
is indefinitely large.
In
determination a priori, when
the above example, each of the parts will then be =J.
159. In the Linnsean Classification of Plants we have an
arithmetical values of the parts

example of a formal division by means of two characters

and their subordinate characters. Linnaeus considered the
universe of
Plants having flowers

having stamens of equal length

having

pistils.

Then
Monandria4-Diandria+Triandria+...
... + Dodecandria -f Polyandria =
and

Monogynia + Digynia + Trigynia +

... + Dodecagynia+Polygynia=i (2).
By multiplying these two equations together we get the
. . .

several formal classes.

If

we consider the universe of

Plants having flowers

then

pendent characters

are two inde

two

ber, etc., are dependent characters.

necessarily exclusive of one another.
so, if

and having pistils

and one in number,

having stamens

in

num

The terms
They

of (i) are
would not be

they meant

having stamens

one at

having stamens

Observation.

If

is

then

etc.

at least

Let x denote a single posi

160. Definition of Contrary.

tive selective operation;

two

least

is its

formal contrary.

such that with respect to the given

universe

y=i x
in reality

trary of x.

y the material, but not the formal con

For example ; in the case of the succession of

then

is

throws with a couple of coins, considered in Art. 157,

Division.

59

That

is,
having the tail of the first coin up is identical
not having the head of the first coin up; and it is
therefore the material contrary of having the head of the

with

coin up.
161. Consider the parts of the division

first

#(17)

agrees with xy in the factor

factor as regards y.

Hence x(i

x but has

the contrary

may be

said to be a

y)

contradictory of the first degree of xy ; and xy, a contradic

Again ; (ix)(iy) is
tory of the first degree of x(i y).

contrary to xy, both with respect to x, and with respect to

it is a contradictory of the second degree of xy.
term, which is a product of factors each
Definition.

y : hence

of the form

or

ix,

is

said to be of the

n th

order,

when

it

contains n such factors.

Two terms

Definition.

ih
of the n order are said to involve

tk
with respect to one another a contradiction of the r degree,
factors of the one are contrary to r factors of the

when r
other.

162. Definition.

be

opposite*,

Two

th
terms of the n order are said to

when they involve a

th
contradiction of the n

degree.

For example; xy and (ix)(iy) in the one diagonal

of the square (fig. 5) are opposites so also x(i y) and
The two terms, which lie
(i
x)y in the other diagonal.
in any one of the four diagonals of the cube (fig. 6), are
:

opposite to one another.

Since two opposite terms have

r=n, we can speak of an

th
opposition of the n degree.
COR. An opposition of the

first degree is a contrariety.

objected that. the above definitions are not
accordance with usage, I answer that they are at

163. If
in strict

it is

and equal to the complexity of nature. It is

impossible to use words in strict accordance with a usage

least exact

60

Division.

which

What some have denoted by con

have denoted by contradictory.
The importance of an exact definition of these terms will
be seen from a consideration of the use De
Morgan makes
is

contradictory.

trary, others

of contrary.
that (i
x)(i

provided
But xy

He
y)

says what is equivalent to the assertion

the contrary of x+y.
Now

is

xy=o.

not necessarily =o.

Hence if i x is defined
to be the contrary of x ;
(ix)(iy) is not the contrary
of x+y.
is

In a

164.

division by

n independent

selective symbols, each

primary term has

i

contradictories of the r

2...r

th

degree.

The primary terms are formed by multiplying together

n expressions, each of which is the sum of a symbol taken
directly and the same symbol taken contrarily (Art. 148).
Let a denote directly and b contrarily then the expansion of
by the Binomial Theorem, groups together the contradic
tories of the different degrees.

an
an

~l

a n -i&
si
a n ruh r

denotes the particular term considered ;

a contradictory of the first degree

second
.,

~th
r

..

Now
i

Hence

the

2...r

number of contradictories of the r th degree with

respect to the given term

is

i. 2

Expansion of a Function,

61

etc.

COR. i. A term has but one opposite.

For when

rn

_
i.

COR.

2.

2 ...

there are n symbols, there are 2 n ~ l pairs

When

of opposites.

For each term has one opposite ; and there are 2 H terms
But only one-half of these
(Art. 152).
pairs will be
different ; hence the number of pairs of
opposites is
2"

EXPANSION OF A FUNCTION OF A
NUMBER OF INDEPENDENT SINGLE AND
POSITIVE OPERATIONS IN TERMS OF THE
PRIMARY PARTS INTO WHICH THE UNI
VERSE IS DIVIDED BY THE OPERATIONS.

XIV.

165.

in terms of x

To expand

and

oc

farts into which the universe

ix,

divided by the

is

which are the

selective

opera

tion x.

a and b being independent of

The symbol
value

x.

being single and positive, can assume the

and the value o. Hence the above equation, being
x,

an identity, must be true when either of these values

is

substituted for x.

Let x=i.

Then

absolute constant.

Let# = o.
constant.

ix=o;

and bo also =o,

Then 00 = 0, because a

Hence

for b is

an

Hence 0=1.

b.
o

is

also

an absolute

62

Expansion of a Function,

Observation

oc

occur by

occur by

ix,

then

The above is the expansion for

i.

and

whether

a factor in an expression.

itself or as

itself,

etc.

If

ii

limited to being a possible operation

is

being the factor of

=o

must be

(Art. 79).

Hence x=i.
Observation

which

To expand

general expansion for

x occur by

true whether

is

166.

The

2.

itself

is

or as a factor.

and -- = *+-L(i-*);
...

x L = x*-\-0 x(l-x)+-!-x(l-x)+
.-.

Now x*=x,
(i_^ = I _^.

and x(i

Hence
o

the third

multiplied

-x)

therefore

it

term has the constant

co-efficient is therefore

given condition.

= o.

term consists of an

by o;

o(lxy.

absolute constant

vanishes.

The

fourth

multiplied by a constant o

its

Thus*=*+^(i-*).
X
O
V

This expansion, of course, can be obtained

COR.

i.

is
00

efficients

COR.

2.

always possible.

occur in
is

its

expansion.

not always single.

directly.

For no impossible co-

Expansion of a Function,
For example

etc.

suppose that we consider

x=x(y+z)
y and z denote the
t

where
fig.

circles of

and x the shaded parts of these

7,

circles.

-,
oc

Fig. 7-

for

is

and

single

Now y+z

positive.

recurs; hence

is

not

OC

In the case of

When

3.

is
oc

single within

single;

restricted

is

hence each must be single (Art. 145).

ix

be single within

When

to

being

x.

For each of the terms of the expansion

167.

single.

represented by the figure,

OC

COR.

necessarily

is

is

positive only

must

Therefore
o

(Art. 139).

taken indefinitely, (

Y= x*

\xj

**

=x-\-l

Now

\(ix);

for x* is given

=x.

cannot be more indeterminate than

o
o

(Art. 80.)

-(5)

taken generally

=x +

(i

x).

Expansion of a Function,

64

But

for

x =x;
2

\Z
J

etc.

taken generally

=#

168. To expand x+y.

Let x+y=axy+l>x(iy)+c(ix)y+d(ix)(iy).
Since x and _y are each positive, they can each assume

the values

and

by any con

If they are not restricted

o.

dition, there are four sets of singular values, viz.

yi

x=i
x=i
x=o
*=o
The above

(i)

y=o
y=i
y=o

(2)
(3)

(4)

equation, being an identity, is true when any

Put
sets of values is substituted for x and y.

one of these

in the first set, then

Since

c,

/?,

do not involve any variable

d,

$0

= 0,

Hence
Similarly

= o,
a=

<ro

do = o.

2.

by means of the other

sets of values,

we

get

i.
The above is the general expansion for
the expansion which is true, whether x+y
In the former case the
occur by itself or as a factor.
can
to
be
reduced
equation

Observation
,

that

is,

x+y= 2xy+x(iy)+(ix)y.
Observation

2.

If x-\-y

each of the terms must be

must be

single

be restricted

single.

which necessitates

xy=o.

to

be

(Art. 145.)

single,

then

Hence 2xy

ion

of a Function,

x+y

that

datum.

rt.

is

single;

etc.

65

we can expand

136); hence

y) -\-c( i x)y +d(ix)(i y).

sets of singular values are
possible,

=o
2

y=

to determine the three constants

b,c,d.

WtttfofV

-y-

independent symbols, there are four

can form four equations to determine
as in

the case of x-\-y.

We

get

wmpto
foxM$**$
//Mfaw
^^ Tt
-

by

itself,

the expansion can be re-

(Art. 165.)

he factor o

is

not a part but a co-

an expression by

itself,

the third

Expansion of a

66
COR.

When

2.

in

addition

Fzinction,

is

etc.

possible, the

second

term may be cancelled.

For the expansion must be possible, and it contains only
hence each term must be possible. There
positive terms ;
fore

Now

-y) = o
is

(Art. 79).

an absolute constant, since we are supposing

to occur

by

hence the

itself;

X o formed by

the second

y
term

=o.

is

The
means

may be

truth of this last statement

seen better by

of the following considerations.

When

is

restricted to

be possible,

y cannot

take the

y
we

Hence

taking the value o simultaneously.

have only three sets of singular values, viz.

value o without

x=i
x=o
x=o

y=i
y=i
y=o.

These ought to be sufficient to determine the expansion

but they will be so, only if we write
\ /

"X?

"

y
where the second term of the general expression

Hence
COR.

3.-

If in addition --

To expand

omitted.

=xy+^(ix)(iy).
be restricted to be

the fourth term of the expansion

171

is

is

with respect

single (Art. 145).

to

alone.

Since

=H

single,

(i

y)-

then

Expansion of a Function,
If

is

^-

= o.

y)

are not entitled to

say that the
;

as well as

If in addition

is

y
But

This expansion
multiplied

by

By applying

is

xyx,

which occurs here

y.

= (i
o
j) = o;

then

for x(i

xv]

from that obtained by

expanding
in not having the second term

x and y

It is therefore less definite.

y.

the

single

x+y
method of Art. 168 we obtain

the general expansion for

When we
possible

To expand

upon

differs

with respect to both

This

because the o of the numerator

depends upon the

co-efficient

172.

must be

--=xy-\^o

is

67

= o.

.-.

=o

possible, then the factor multiplying

Hence x

We

etc.

consider

_L_

by

- -

x+y

itself,

and

restrict

it

to

be

then the constituent

multiplying the co-efficient

must be

=o

the co-efficient

(Art. 79).

must be

Also the constituent

multiplying

=o;

for

xy does not contain

as a factor
(Art. 78).

173.

To expand

(Art. 168.)

Expansion of a Function,

68
and

etc.

--

(Art. 172.

multiply together, then

terms vanish, on account of their consti

for all the other

tuents

being =o

(Art. 78),

but-^=i

may be

easily verified by expanding directly.

by means of the results of Arts. 168 and
the purpose of showing that when the function

This result
I

have deduced

172, for

is

expanded
consists of

it

a factor of an expression,

all

the terms, even though

its

true expansion

some of these have

impossible or zero co-efficients.

174.

To expand

X~-

y+z

Since

and

.#=* + o(i-#),

(Art. 165.)

z
(Art. 172.)

y-\-z

When -2-

y+z

constant, the

When
first

occurs by

fifth, sixth,

itself

or multiplied by an absolute

and seventh terms vanish.

in addition -*-

y+z

is

(Art. 79.)

of the
possible, the constituents

and fourth terms must each be equal

fore the terms themselves equal

to o.

to o;

Then

and there

69

Definition.

itions

0,72=0, and.r(i

single, then

(i

s)

= o.

JF)(I

-)

J )(i
#)(i

DEFINITION.
is the establishment
by convenbetween a complex expression and a
?

may be

it

Let

said

x=x

established between the significaracter i

x and the simple character
is

if

i,

adopted

lered
3r

-to

in

a particular course of
but it is some;

throughout

one course of reasoning only, but

established for the sake of

gaining
for the purpose of
expressing
3) the result of the complex expres-

d also

Let x_y=x
symbol the
.

le
jr

y upon

.tional

lead of

result of the

product of

the character x, on which

equation, usually considered

Definition, is related to the
*

^pe
Let

is

lual

symbol denoting the universe Ux.

Ux= U

mrlf

m.m

1*4*

?!&#$%!
A^J/*/^

Inference from one or more Equations

7o

Conventional equations of

manent
178 As an example of

this

kind are usually of a per

nature.

I take the definition

(Manual of Zoology
as follows

it is

U = Ux,

of

given by Nicholson

Vertebrata

p. 383).

When

expressed analytically,

Vertebrata

_ Animals
.

having a body composed of a number of definite segments

arranged along a longitudinal axis

X having a nervous system

in its

off by a partition from

the haemal region of the body

main masses dorsal X completely shut

X having limbs
not more than four in

number X turned away from the neural

aspect of the body

axis known as the vertebral column+without a

a
x having bony
vertebral column X having a notochord
\
(

Here x

is

of the form ax byz

^- u

-)

{d+(i

d}e]

)
"

and

good example of the import of dependent cha

affords a
racters.

XVI.

INFERENCE FROM ONE OR MORE

EQUATIONS OF THE FORM

x=m ;
THAT
179.

of the

equated

IS,

FROM CATEGORICAL JUDGMENTS.

An unknown single character

of the equation.
with
and
first order
unity for its co-efficient is
to a known character, or else to a known arith

Form

metical value.

These are written

x = m,
and

x=m,

The
respectively.
t/ s which are x are identical with
former

and the

is

read

which are

latter

which are x have the arithmetical value of

which are m.
[/ s

m;

of the form

xm.

It is generally sufficient,

made, to make

it

Categorical. )

when an

investigation has to be

kind of equation only ; as

generally true for^the latter kind of

for the former

the result obtained

is

be changed into m.
equation, provided
180. The primary categorical forms comprise four special
cases of the equation

x=m.
When

m=i

the equation

becomes the universal

affirmative.

= o;
universal negative.
= v (a fraction greater than o)
particular affirmative.
=
y
particular negative.
Hence x=m may be called the general form of the cate
i

27

judgment, m is the definite some considered in the

Theory of Probability.
181. The term categorical is frequently ascribed to judg
ments which differ in an important respect from those to
which the term is primarily defined to apply. The forms

gorical

xy=i, xy = o, xy=v,

xy=iv

are not distinguished from the forms

one of the examples of a universal negative

j
judgment given by Ueberweg is (System of Logic,

For instance
categorical
p.

308)

No
But

this is

innocent person

No unhappy
In the
it

new form

was before

person

unhappy.

is

innocent.

of the statement

hence

person remains where

innocent persons

persons and not

the true subject of thought.

for these judgments are

is

is

converted into

The

analytical expressions

Persons {innocent unhappy o},

Persons {unhappy innocent =0}
both of which are of the form

and

Inference from one or more Equations

72

The above
and

is

equation contains a term of the second order,

which contains a

therefore very different from one

term of the

first

The

order only.

equation

where the term on the left-hand side is of the second order,

may be called the general form of the hypothetical equation
of the

first degree.
182. Immediate inference by Contraposition.

If

then

x=m,

ix=i

m.

(Art 108.)

In the same manner,

if

then

The

x=m,
x=i

m.

well-known proposition in the Theory of

and is one of the logical theorems which are
constantly assumed in stating arithmetical problems.
COR. i. Let^=i. Then i x=o.
COR.- 2. Let x=v where v is greater than o.
Then
v; where i v must be less than i and
may be o.
It is said that no inference can be drawn by contraposi
latter is a

Probability,

ix=i

tion from

x=v.
But

this is

not correct.

It follows necessarily that

IX=-LV
iv must
)

what does not follow, is, that

when v is greater than o.
183.

The

be greater than o

Conversion of x=m.
way in which the equation

only

x=m
admits of conversion, consists in exchanging the position of
its

members

thus

m=x.
It is

only to such equations, as involve two unknowns x and

of the form

xm.

Categorical.)

73

y, that a process of conversion in the proper sense of the

For instance

term applies.

may be changed

xy=m
yx=m ;

into

x=y w
y=x+w.

and
.into

184. An attempt, however, is frequently made to convert

the categorical form after the manner of an hypothetical of
It is, for instance, said that
the first degree.
All men are fallible ;

can be converted into

Some fallibles are men.

But several logicians have observed that such a procedure
confounds Accident with Substance. The universe of the
judgment is confounded with the characters of the judg
ment selective symbols are changed into elective symbols.
The result in the above or any other instance does not re
commend the process. The given judgment really means
;

Men

{i

= fallible}

which can be written

Men{fallible=i}.

The

inference by contraposition

Men
185.

{o

= not

is

fallible}.

important to consider the differences in the

It is

signification of

#=i,

as regards variables.
If Per
according to the nature of
sonality (which is the logical Space) is constant, and State

(which

is

the logical Time) varies

then

x=i
expresses
character

that in every state of the given individual the

Similarly,

varies;

is

present.

when

the State

is

fixed,

and the Personality

x=i

expresses that in every individual in the given state, the

character x is present.

Inference from vne or more Equations

74

When

both Personality and State vary

x=i
expresses that in every individual in every state, the cha

racter

is

present.

Examples of these respective cases are

This crow

always black.

is

All crows at a given age are black.

All crows are always black.

The

186.

formal treatment of the equation

x=i
the same, whether the equation express an essential pro
perty, a general property, or an inseparable accident, (for
is

the given universe).

The Algebra of Logic makes no dif
ference in this respect between the necessary truth
Triangles

equilateral

= equiangular}

and the observed fact

(

Animals

<

having condyles on the occipital bone X possessing red blood corpuscles

two in number
non-nucleated
=suckle their young. }

187.
to

x=.m and y=n ;

Given

find what can be concluded about xy,

and

(i

x(iy),

(i

x)y,

(iy).

x)

First, for xy.

Since

but

(ix) (iy)

#)

(i

is

therefore

but x

(i

since

y)

is

x(i

is

it

(Arts.

greater than

120 and 77.)

x+yi

i.

is, xy>x+

(Art. 146.)

y)=xxy,
xy=x-x (iy);

positive only
therefore xy

that

Similarly

y+xy,

positive only;

xy

that

Again

y)=i

(i

(Art. 120.)

is

less

than

that

xy

(Art. 146.)

is, xy<x.

may be shown

<y.

of the form x^m.

x (i

Second, for

Categorical. )

y).

Since

xy=.x+y i-f(i
let y be changed into

then

x(i-y)=x

but

x)y

(i

therefore

And

75

in a similar

is

it

""

x)y;

J>+(i

7)

>

y)>

y;

positive only

x (i

manner

^)( 1

y.

may be shown

that

x(i-y)<x j:

and
Third, for

(ix)y

same

the

is

as

x(iy),

Fourth, for

188.

(ix)

into

into

y, in the limits for xy.

putting in the given data

x=m, andjK=^,
we

obtain
xy>m-}-nJ
<

y).

ix, and y

Hence by

excepting that x and

Hence

have replaced one another.

Change x
Then

<iy.

(ix)y.

i
<m

x)y>m-\-n

Inference from one or more Equations

When
x=m,

Observation.

the data are of the form

y=n,

the above inequalities are still true, provided the

right-hand
members be stroked with a vinculum.
189. From the above we infer; that
any two terms,
which involve two degrees of contradiction, have the sum

of their -minor limits

=o

and the sum of

their

correspond

Also that any two terms, which in

ing major limits =i.
volve but one degree of contradiction, have the sum of their

minor

limits

= it

2
j

sponding major limits

190.

Any one

the

(^}~ I [and

sum of

their corre

=2

of the equations

xy=x+y

+ (i

x)(i -y),

ocy=x-x(-L-y},

xy=yy(ioc),
suffices to give

xy

of the right-hand
191.

Example.

exactly,

member

provided each one of the terms

given as a datum.

is

Suppose that of the persons on board a

ship which was wrecked, the passengers formed two-thirds

and those that were saved in the wreck three-fourths.

How

many passengers must have been saved, how many lost

how many of the crew must have been saved, how many

lost?

Let/ denote

the passengers

Then/+^-=i;
Also J + /=i
;

And/=|;

The passengers saved formed at

the crew

c
.-.

.$

saved

=ip.

/=i

j.

j=f.

least five-twelfths.

I lost.

xtn.

/)(i

s] is

of the form
Since the limit of (i

but with a contrary sign

the same as that for ps

(i

Again.

77

Categorical. )

p)s

>

p + s,

saved formed at least one-twelfth.

Since the limit for/(i
s) is the same as that for (ifl)s
but taken with the contrary sign,
/(i-4>- T\.

The passengers

lost

may have

been none.

192. Suppose the additional information

the crew lost formed one-sixth.

Now /*=/+*- 1 +(i-/)(i-r,r),

Also/(i

=|+|-T+I
= T2= i J (i
j)

J)(i

/),

Also (i-/>=i-/-(i-^)(i~4

given

that

(Art. 190.)

(Art. 190.)

(Art. 190.)

"S"

-A
in

Suppose
number.

in addition that the persons

Then

And

Also,

when the data of

on board were 120

U= 120.
Ups=>]o.

(i)

j)=io.

(2)

Art. 191 only are given

(i)
(2)

78

Inference from one or more Equations

Given

193.

x = m ; y = n ; z=p ;
what can be concluded about xyz.
Since
Xy=x+y i + (i x}(\ y)
multiply by z,

to find

thenxyz=xz+yz-z+(ix)(i-y)z.
But

xz=x+z

+ (i

x)(i

(Art. 108.)

z),

+ (i -*)(i-,)*.

Now (i-x)(i-z) + (i-y)(i- z ) + (i-x)(i-}>)z,

a

sum

itself

(i)

being
of terms each of which cannot be
negative, cannot

be negative

xyz>x+y+z2.

xy=xx(i -y)

Again,
multiply by z

(Art. 146.)

then

xyz=xz

x(iy)z;

=xx(i -z)-x.(i

(2)
}>)z

<x.

For the remainder is minus the sum of two terms each of

which can be positive only. (Art. 146.)
Similarly

it

can be proved that

xyz<y,

Hence, by putting

and

<z.

in the given values,

xyz>m-\-n-\-p

194.

Other two equations

2.

may be

derived from (i) by

but, as the terms of the
first order enter
similarly into (i), the derived equations
will give the same limit as
(i) gives.
cyclical

change of the symbols;

two other equations may be derived from (2)

Each of these yields
cyclical change of the symbols.
two forms but only three limits of the kind considered
can be derived.
Also,

by

of the form

xin.

( Categorical. )

79

195. The limits for any of the other terms of the third
order can be deduced from those for xyz.
For example; to find those for x(\y](\z].

iy=in,
and
x(i

.-.

2=1

y)(i

z)>m

/.
i

n+i p

2-

>mnp.

Also.

196.

x(iy)(iz)<m.

Generally

x x tXt...xn
l

Suppose

it

n
l +x t +x 9 -\-...+xM
n; and multiply by x n+l

>x

true for

Then

(Art. 187.)

Hence

if

the inequality

is

true for

it is

also true for

72+1.
it is true for 2 ; hence it is true generally.
COR. If r of the factors are direct, and n r contrary;
the minor limit becomes

But

Xi+x

+ ...-\-x

For the minor

xr+l

limit

x r+z

...

xn

r+i.

is

r+i.

De Morgan makes some

investigations upon the sub

ject of the preceding ten articles under the name of Numeri
The fundamental data with
cally Definite Syllogisms.
197.

which he deals are inequalities, and he always supposes m

But there is nothing to prevent us from
to be arithmetical.
supposing that the fundamental data are equations, and
A slight reflection upon the
investigating what follows.
nature of statistical data will serve to
of such an investigation.

show

the importance

8o

Inference from one or more Equations

198.

Given ax =m, ay =n, a=p ; to find limits

a xy a x+y-l + (l-x) (1-j/))

=ax +ay a+a ^

axy
a.
x +ay
(

(i-^

x}

to

a xy

(Art. 70.)

>a

=ax -ax(l _y}

a xy

<a

Similarly axy

(Art. 70.)

x.

<a

Hence a xy >m-{-np.
but

<

and

<

m.
n.

COR. Let 0=1.

Then we
199.

get the results of Art. 188.

To find

limits to

Since

where

x) (i

(i

Hence

(ix)

y)

also

it is single.

is

given to be single.

is

single.

y)

(i

when

<

(Art. 170,

x and

xy<x

and

<

x,t.e.<i.

COR.

3.)

<iy;

y;
(i)
(2)

But as

xy=x

(Art. 170), the

expansion can be reduced to

j=*+^-y);
therefore, the important limits are

x
-<x+i-y,
.

and

x
>x.

(s)

of the form
COR.

i.

xm.

Let.#=i and y

2.

Then equation

(5)

=#

becomes

y
i.

(5)

becomes

=i.

m;

xy

200.

INFERENCE FROM ONE OR MORE

EQUATIONS OF THE FORM

XVII.

THAT

Categorical. )

Letjy=i.

Then equation
COR.

FROM HYPOTHETICAL JUDGMENTS.

of the equation. We have .an unknown single

IS,

Form

character of the second order equated to a known symbol,

which may either be a symbol of identity or express arith

metical value.

xy=m
means

that

Cfs which are

means

x and y

are identical with

7 s which

are m.

xy=m
that

which are x and y have an arithmetical value equal

to that of Urn ; or, simply, that xy has an arithmetical value
s

m.
201.

A particular case

is

xy=x;
which

is

an equation universal with respect to x, and means

that

7 s which

are

are y.

xy=y
is

the other universal form.

xy = Q

means

that

which are x and y are none.

F

Inference from one or more Equations

82

Varieties of the equation

202.

There are four

xy = m.

varieties

xy=m.

(i)

x(iy) = m.
(ix)y=m.

(2)
(3)

(i-x)(i-y)=m.
Each

of these has two universal forms,

viz.

xy = x;

(1)

(2)

xy=y.

x(iy)=iy.

(iy)=x;

(i-x)y=y.
(i-x)y=i-x;
)= i-*; (i-#) (i-y)=i-y.

(3)
(4)

(4)

(i-#)

(!-.>

The right-hand member, m, may assume a great

203.

variety offorms.

For example

xy=i
xy=.iy.

jc.

xy=x+yi.
Or we may have

xy=ix.
xy=iy.
xy=x+yi.
The

latter

examples

from the former

differ

in

that the

equivalence asserted does not involve identity.

The equation
Observation.

xyx -\-y-i
is

an equation of Condition

whereas the equation

xy=x+y-i+(i-x)(iy)
is

an Identity.
204.

For

andj

7fxy=m,

then

xy~m ;

but not conversely.

expresses the identity of the ^7 s which are x

with those which are ;;/. Hence the arithmetical value

xy=m

of xy must be equal to that of

m;

xy = m.

But suppose

it is

given that

that

is,

m.

of the form xy
Since xy

may

83

Hypothetical. )

have the quantitative value

in general

a plurality of ways,

~ni

in

and

xy=.m
is

only one of these ways

it

does not follow that

xy
COR.

Ifxy=x

then

m.

xy=x.

For xy can have the quantitative value x

is

in

only one

xyx

way; and
one way.
205.

There are certain forms of the equation

xy=m,
which can be satisfied only by one or both of the symbols
having
a singular value.
First example

xy=ix.

Then xy+x=i,
xy=o, (Art. 136) for i is single.
i
#=0, by given equation.
.

.-.

and
Second example ;

x=i;

.-.

yo.

xy=x -{-y.
Then*y=o,
.

For both terms are

x=o

xy=x ( i

is left

206.

y)

Then 2xy=x,
xy=o for x is single.
.

positive only.

Third example ;
.-.

(Art 136.)

,T=O.

undetermined.

The equation may

express a condition, which can

be reduced to a simpler condition.

For example; let xy=(ix) (iy).

Then xy= i
.

0=1

xy-\-xy ;
x y;

Inference from one or more Equations

84-

xy = m;

207. Given
to

findx(\
and m.

y),

(ix)y, (ix)

y) in terms of x, y,

(i

x(\y}xxy,
=xm.

First,

(ix)y=ym.
(ix)(iy)=ixy+xy,

Second,
Third,

\xy-\-m.
The

which gives the expression

last equation,

opposite of the given term,

is

for the

commonly called the inference

The other two equations give

obtained by Contraposition.
expressions for those terms, which involve only one degree
of contradiction.

COR.

i.

Let xy=x.

x(iy) = o.

(i)

(ix)y=yx.

(2)

(ix) (i-y)=i-y.
Letxy=x+yi.

(3)

Then

COR.

2.

Then x

COR.

3.

y.

(i )

x.

(2)

208. Let

v,

It

-. x

(3)

(i)

)y=y.

(2)

(i-x)(iy)=ixy.

(3)

xy=v, where v is greater than

Then x(iy)=xv;

o.

though restricted to being greater than

equivalent to x.

than

y}

(i

(i-*)(i-^)=o.
Let xy=o.
Then x(i y)=x.
(

but

(ix)y=i

Hence x(i-y]

is

o,

may be

not necessarily greater

o.

may be shown

in a similar

necessarily greater than

manner

that

(ix)y

(i-x)(i-y) i-x-y+v;
Again.
but v may be equivalent to tf+^i,
.
not necessarily greater than
#)(i .y) is
(i
p

is

o.

o.

not

m.

of the form xy
The

statement contains what

last

85

(Hypothetical. )

meant by the pro

is

position, that the particular affirmative (hypothetical)

ment does not admit of Contraposition.

209. The Conversion of an Equation

consists in.

judg

making

two symbols replace one another as far as possible.

As applied to equations of the form

xy=m
the Conversion

210.

may be

Simple or per Accidens.

-,

The Conversion

is

Simple,

when x and y

enter

similarly into the equation.

For example

if ocy

=o

211. The Conversion \s per Accidens,

enter similarly into the equation.

For example

xy=x.

let

is

when xandy do not

in general greater than o.

yxy

Hence

the equation
does not in general follow

the additional condition required

being

Given
to find minor limits
212.

y=x.
xy=m, andyz=m;
to xz, -x(i

expressed in terms of

m,

z\

n, x, y,

(i

x)z,

+ (i

x)(i

and (ix)(iz)

z.

First for xz.

Since

xzx-\-z

z)

multiply by y,

then

xyz=xy+yzy+(ix)y(iz)

for (i

x)y(i

But
.

Hence, by

z) is positive only.
xz>xyz,

xz>xy+yzy.
putting in the given data,
.

xz>m-\-n

y.

86

Inference from one or more Equations

Similarly; since
x(i-y)z>x(i-y)

and
that

is,

that

is,

+ (i-y)z-(i-y)

y)z

xz>x(i

>xyyzi-\-x-\-yz,
m n \-\-x-\-y-\-z.

>

Second for x(i

z).

Since zy=n,

Thus

(iz)y=yn.

iz

and
replace one another in one of the given
The results, when
conditions, and so also n and yn.
z

be

similarly transformed, will

Hence

x(i
that

true,

z)>m+y
is,

y,

n.

>m

Also x(iz)>m(yti)i+x-\-y-\-(iz\
that

is

z.

m-\-n-\-x

>

In the same manner

(

.may be shown that

m n;
x)z>

it

and

>m

x-\-z.

Also that

(ix)(Lz)>mn+y

and
COR. Let y=

and

xz>m-\-n

i,

z)>m

n, -and

^(i
(i

m-\-n,

x]z>

and

z)>

whenjy=i,
213.

limit

ought in each

>o.

>o.

n+i,
^)(i
are the results of Art. 188.
(i

>o.

and

-case to

>o.

become o

independent of m

the

and

two

limits

for any one of the terms

n.

Since xyz >m-\-n

y,

and x(iy)z>m-ni-\-x-{ y-{

which

is

for

y=o.

The sum of

add,

z.

i.

Then

These
The second

xy

>m-\-n-\-i

then

independent of

i-\-x-{-z-}

xz>

and

n.

Z)

is

m.

of the form xy

87

(Hypothetical. )

together the two limits of any other term, it

be found that the result is independent of m and n.
Hence ; when each of the two limits is
o, each limit

By adding
will

>

must be

When
>

i+x-{-z.

<

each of the two

limits

is

<o,

each must be

i+^-fs.

When

one

is

Given

214.

and the other

>o,

and the

i -\-x-\-z,

>

x}>=m,

latter is

and

(i

<

<o;

the former

is

i -\-x-\-z.

to

y}z=-n;

find minor limits

to xz.

Here the common symbol occurs

contrarily in the

data; whereas in the former case (Art. 212)

it

two

occurred

similarly.

Since

(iy)z=n,

yz=zn.
Hence, if we make n and z n replace one another in
the results of the former case, these results become true
for the present case.

Thus
that

Also

xz>

xz>m-\-(z
is.

>m

7n

n)

y,

y-\-z.

(zn)T.+x-\-y+z.

>

that

is,

>m+ni+x+y.

Similarly for the other terms.

215. The sum of the corresponding limits of any two terms,
which are opposite to each other, is equal to o.

Consider for example those limits of xz and of

x) (iz) which involve y.
(i

They

are obtained from the equation

y{xz-(i-x)(i-z) = x+zi},
by supposing (i x)y(i z) = o and xyz=o respectively.
Hence they are equivalent but of opposite sign and
therefore their sum is =o.
;

216.

limit,

when

m+x,
cannot be negative.

it

assumes either of the forms

m+y, -n+y, -n+z,

88

Inference from one or more Equations

Consider the

first

form.

m=xy,
and
that

is,

xm cannot be

by given equation.
(Art. 187.)

.*)><.#,

negative.
to the other forms.

The same proof applies

Example of application.
Given xy=.m
(i);

From
.

xz>xm

that

is,

xm

217.

z(iy)=iy
y)=x

x(i

(i)

(2).

m;

+ -Ly(iy),

(Art. 212.)

xz>xm;

but
cannot be negative,
the limit for xz cannot be negative.

which

limit,

of either of the forms

is

mxy+i,

yz+i,

cannot be negative.
Consider the first form.

m=xy,
and
.

that

m>x-\-y\

m x y+i

is,

can be shown
cannot be negative.

by given equation.
(Art. 187.)

i,

xy>x+y

cannot be negative.
manner that n

in a similar

It

z+i

Example of application.
Given

From

(i)

from

(2)

xy=m

.-.

limit,

(2).

y)=i

>I

the limit for

.-.

218.

yz=z
xy-\-m;
(i-^Xi-^ssi-^/
- x -y+ m
(i-x)(i-. s)
(i)

#)(i

(i

(Art. 212.)

cannot be negative.
which is of either of the forms
n
m,

(ix)(iz)

cannot be negative.

For

m=xy, and xy

cannot be negative.

Example of application.
Given

xy=m

yz=y
(i);
therefore xz>m-{-yy,

(2).

m.

of the form xy

89

Hypothetical. )

Hence the limit for xz cannot be negative.

When the miner limit is equal to a factor of the term

219.

it is a limit, the inequivalence becomes an

equivalence.
For any term, which can form the left-hand member of
the inequivalences considered, cannot be greater than any

of which

of

factors.

its

Example of application.

Let*y=*

z(i-y)=z

(i);

Then from

(2)

z)y=.y

(i

therefore x(iz)>x -\-y

>x

but

is

(2).
;

y,

a factor of x(iz),

therefore

x(iz)=x.

This corresponds to what is called a universal conclusion.

Nothing can be more elegant than the manner in which
the general analytical conclusion becomes universal,
the data necessitate the conclusion to be universal.

when

220. When the limit is equal to the common symbol, or its

contrary, the conclusion does not necessarily become universal.
For neither y nor i y is a factor of the left-hand

mem

bers of the inequivalences considered, viz. xz,

x(iz),

etc.

Example of application.

Let*y=7

yz=y

(i);

(2).

then xz>y+yy,
>y;

but

not a factor of xz,

.-.
xz=y is not necessarily true.
If we consider the limits of xyz, then
Observation.

is

xyz>y

becomes xyz=y.
221.

To find

(as the case

the condition, to

of m

which the sum or

and n must

may be)
a given limit may be positive only.
First ; when m and n each enter
Example:

m+n

difference

be subject, in order that

positively into the limit.

y.

This expression cannot be negative,

\im-\-n includes y.

if m-\-n>y;

that

is,

Inference from one or more Equations

90

Second ; when

and n each enter negatively

into

the

limit.

Example:

\-\-x-\-y-\-z.

This expression cannot be negative,

is included in
that is, if

m+n

if m-\-n

<x-\-y-\-z

x+y+zi.

Third ; when

and n

one

enter, the

positively,

and the

other negatively.

Example

y -f- z.

This expression cannot be negative,

is,

\im-n

include

yz.

if

n>yz

that

Another form of writing

n

this condition
m<z

is

y,

where the signs of the members have been changed, and

consequence the sign of inequivalence also changed.

in

222. Suppose, for example, that

xy = m ; yz=n.

Then the limit for xyz

is

positive only

\im-\-n

x(iy)z
xy(iz)

>y.

<x+y+zi.
mn>o.

x(i-y)(i-z)

<x-z.

(ix)yz

<o.

(i-x)(i-y)z

>x-z.

.,

(ix}y(Lz)
(i-x)(i-y)(i-z)
Again

Then

suppose that

the limit for xyz

is

positive only

x(i-y)z
xy(i-z)

(ix)yz
-x)(i-y)z
-x)y(i-z)

if

n>yz.

m-\-n>z.

<z.

>x.

m-n<y-z.

of the form xy
223.

The above

m.

91

(Hypothetical. )

investigation

to

applies

equations

of

the form

Let

*=

;=

Then maximum
minimum

value of

limit for

mn

maximum
minimum
The

m+n
_

is

xyz

positive

if

x(i-y)z

xy(iz)

^)j(l-2r)

(l

(i-x)(i-y)(i-z)
Suppose that the given equations are of the form

xy=m; (iy)z=7i;
and

let x,

}>,

z have the

same arithmetical values

Then maximum
minimum
,,

The

limit for

value of m-\-n=-^.

_.

is

positive; if

x(iy)z
xy(iz)
^(1-^)1-5:)
(i-*).?*

(z-.)(z.jr)

-l-s _

(l_^)( I _y

_=f^._

m?i=
=

maximum
minimum
xyz

as before.

>f

fj.
ft-.

>

^.

Inference from one or more Equations

92
224.

xz

m+n

2}

mn

(i-x)z

m-n

(i-x)(i-z)

m+n

yz=n.

,y=&
FIG.

Z
>

8.

mn

xz

C3

xy=-m
j$s

Q|

J [

x(i-z)

m+n o-H

(i-x)z

7n

+n

1\

&-t\

3H

xy=m; (iy)z=n.

X = TU l
>

= TG

FIG.

= rV

9.

These relations are represented graphically in figs. 8

and 9. The values of m+n are represented by distances
from o along the top line of
under the given conditions

fig.

x =TViy=fv

8.

Its possible values,

=f&

minimum value, when

the limit for xyz is positive, is represented by the distance
from o to a stroke pointing in the positive direction ; and

lie

between the

straight strokes.

Its

maximum value, when x(iy)z is positive, is repre

sented by the distance from o to the stroke pointing in the

its

negative direction.

of the form

xy^m.

Similarly, the values of

from o along the second line

93

{Hypothetical. )

n are represented by lengths

when positive by a length to

when

Its
negative by a length to the left.
under
the
lie
between
the
possible values,
given conditions,

the

right,

Its values,

tall strokes.

when

the limit for

xy(i-z)

is

by distances cut off by the stroke

and its values, when the limit for

positive, are represented

pointing to the right

(i

y)

pointing to the

And

2) is positive,

(1

on

so

22. When

by distances cut

off

by the stroke

left.

for the others.

mn) has

m-\-n (or

a value inside

],

then

the two limits are simultaneously positive; when inside

the two limits are simultaneously negative ; when
]
[,

outside

] or ]

the one limit

[,

is

positive

and the other

negative.

226.
to find

Given

what

and x, y,

yzn;

xy=m,

relations necessarily exist between m-\-n or

Write each of the equations in

its

four equivalent forms,

as follows

xy = m.

(!

(i)

(ix)y=ym.

(2)

x(i-y)=x-m.

(3)

-x

(i

-y) = i -x-y + m.
yz=n.

(4)

(5)

(i-y)z=z-n.
y(i-z)=y-n.
Xi-^) (i-*)= i -y-*+.
From

z.

(i)

and

(6)

xy

(i

y)z>m+z

but 7(1
.

y)

= o,

o>m-\-zni

n<

z.

(6)
(7)
(8)

(Art. 188)

94

Inference from one or more Equations

From

(i)

and

(8)

we obtain by

o>m-\-i

similar reasoning
i

z-\-n

/. m+n<y-\-z.

By

method we obtain

this

From

and

(i)

(2)

(6),

(8),

m+n<y+z.

(6),
(8),

(3)

(5).

(4)

(5),

(7),

(7),

227.

Given

find what
and x, y, z.
to

n<iz.

>y+z-i.

n>

>-(i-#).

m
m

n<x.

(iy)z = ?i;

xy=m,

relations necessarily exist between m-\-n or

Write each of the equations in

its

xy=m.

(i)

)y=y-m.
x(i-y}=x-m.
-x)

(i

(2)

(3)

-/) = i -x-y+m.

(4)

yz=zn.

(5)

(i-j;)0=.

(6)

^(i-*)=y-*+i

(7)

(i-y)(i-z) = i-y-n.

Then by
we obtain

applying the

From

(i)

and

(3)

(6),

m-n<y.

m -n>-(i-y).

(8),

m+n>o.

(5X

w4->Jc+2-i.

(5)>

(i),

Article

m-{-n<i.

(6),

(7),

(4)

(8)

method of the preceding

(8),

(2)

four forms as follows

(T.-. x

(i

mn>x+yzi.

n<x+y-z.

m+n<x+z.

m.

of the form xy

95

The merit of this investigation is

The truth of any of the individual

Observation.
haustiveness.

can be established
first

( Hypothetical.

directly.

ex-

its

results

Consider, for example, the

of the above relations.

ocy

<

and

that

To find

228.

when

~z

y)

<

y,

is, m-\-n< i.

expressions for the limits independently of one

of the two data m,

First ;

n.

the datum, which

to

is

be eliminated, enters

positively.

Example. If xy= m, (i) and

can be shown (Art. 214) that

From

(2)

n>

(iy)z=n;

y+ z

an d

(2)

>

it

(Art. 217.)

>0j

(i-*)Xi-*)>-Mj

and

then

(3)
z.

m-\-y

Second ; when the datum, which

is

(4)

be eliminated,

to

enters negatively.

Example.

From

Under

the

same conditions

(2)
...

<i

(i-#Xx
i.

y and

<z,

jO(x-*)>-*j

and
Observation

as above,

>m

(5)

+ ixyz. (6)

In each case one of the pair of

expres

sions obtained cannot be positive.

The

expression (3) cannot be positive

for

is

positive

cannot

only.

The expression (5) cannot be positive

include more that x.
Observation

The

2.

be verified as follows

The

for

truth of the other two results

inequality (4) asserts that

that

is,

(ix)y(iz)>y(ix)z,
that

which

is

evidently true.

is, z>y(i

(Art. 193.)

x)z;

may

Inference from one or more Equations

g6
The

inequivalence (6) asserts that

that

is,

>(i-x)(i-y)-z,

that

which

is,

evidently true.

is

z>(i-*x)(iy)z

229. To find the condition which

order that

it

may

(Art. 193.)

or n must

be impossible for the limit

satisfy, in

of a given term

to

be negative.

is required, ;/ must be elim

When the condition for
inated from the general form of the limit in the manner
shown in the preceding Article. One of the results obtained
is

necessarily negative ; hence not more than one condition

If
enters positively into the other result ;

can be found.

then the required condition is, that m include the remainder

If it enters nega
of the expression with its sign changed.
tively ; then it must be included in the remainder.

For example;
is

m>x-}-y-^-zi,

the condition which

Article 226, in order that

of (i

x)(i

x)y(i

z)

Two MIDDLE

Given

under the data of

for the limit

z,

\im<y

the limit for (i

230.

may

satisfy

be impossible

be negative.

y)(i -~z) to

Similarly,

must

it

cannot be negative.

TERMS.

xy=m,

zu=p ;

yz=n,

minor limits to the terms of thefourth order having

x and u as factors, which can be expressed in terms of m, n,

to find those

p, x, y, z or

u.

xyz>m-\-ny,

(Art. 212.)

and zu=p,
.

that

xyzu>m-{-ny-{-pz;
>m-\-n-\-pyz.

is,

+x-\-y+z,
and zu=fl,
m n i -{-x+y -\-z-\-p

Also, x(iy)z>

that

(i
is,

y)

zu>

>

n-\-pi-\-x-\-y.

(Art. 212.)
(i)
(Art. 212.)

z;
(2)

of the form

m.

xy

97

{Hypothetical. )

Again xy (iz)>mn,
and (iz) u=up,
that

Also,

>mnpL-\-z+u.

is,

x(i

and
.

Now

y)

(i

(3)

y) (iz)>m-\-n-{-xz, (Art. 212.)

(i

z)

(i

u=up,

m-\-n

u>

z)

p~

\-\-x-\-u. (4)

and only four terms of the second

order which can be formed by means of the middle symbols
y and 2, viz.
there are four

yz,

Hence

(i-y)z,y(i-z), (iy)(iz).

the above are

all

the terms of the fourth order

which contain x and u as factors.

COR. Each of the four limits obtained is a limit
For xu is greater than each of the terms.
231.

The

have x and
factors,

xu.

to

limits for the terms of the fourth order,

iu,

or

can be found

x and

u,

in a similar

or

x and

which
u as

manner, or deduced from

the above results by making the appropriate transforma

tions.

For example

to

deduce the

(l-*)(l-J>)(l

xyzu

and

(i
(i

limit for

-*)(!-*).

m -\-n-\-p-y-z ;
(iy) = ixy-{-m.

>

x)

-y) (i-z) = i -y-z+n.

>m

232.

The

-\-n-\-p-\-\

x y

u.

the simplest example possible of a

conclusion derived from three equations of the kind con
Sorites

is

The premises
x

sidered.

IT S which are

are

are

y;

7 s which arejy are z ;

C/ s which are z are u ;

xy=x.
yz=y.
zu=z.

(i)
(2)

(3)

Inference from one or more Equations

98

Now

xyzu>m+n+p
z

>x

that

is,

(Art. 224.)

z,

yz,

+y+

by premises*

>x,

=x.

The

conclusion

is

7 s which are x

What

is

which

is

We

are

u and y and

called the conclusion

commonly
L7

(Art. 219.)

read

z.

is

which are x are u ;

evidently only a part of the

full

conclusion.

have shown in the preceding Article that

(!_#) (i-y) (1-2) (i-u)>m+n+p+i-x-y-z-u.

x +y + z -f- 1 * y z u,
Hence by the premises
>

>iu.

which are not u are not x and not^ and not

This conclusion

is

complementary

z.

to the other.

The investigation of the two preceding

Articles shows how much truth there is in the statement
that no conclusion can be drawn when more than one of
Observation.

the premises is not universal.

233. Given the conclusions of Art. 230 ; to deduce minor
limits to the terms of the third order having x and u as
factors.
It

has been shown that

y z.
x(i y)zu>mn+p-L+x+y.
xy(iz)u>mnfi+z+u.
x(i-y(i-z)u>-m+n-#-i+x+u.

(i)

xyzu>m+n-\-p

By adding

(i) to (3)

(2]

(3)

(4)
(5)

xyu>2m-i-y+u.

(2)

(4); x(i-y)u>-2m-2

(i)

(2); XZU>2p-l+X-Z.

(3)

(ti

+ 2x+y+u.
(

6)

(7)

,x(i-)u>-*P-*+x+*+
(8)

Now
comprise

y,
all

i-y,*>

*- 2

the different expressions which can form the

third factor of a term of the third order having

x and u

of the form xy

m.

Hence

constant factors.

99

(Hypothetical. )

the above comprise

all

the re

quired limits.

COR. i. Each of these four limits is a limit to xu.

COR. 2. Each of these limits is less than the one which
can be obtained directly.
For example consider (5).
;

xy=m, and u=u,

.

Now

but
.

m y

the limit (5)

Observation.
limits.

(2m
is

xyu>m+ui.

(Art. 188.)

(m+ui)=my

y+u)

is

necessarily negative,
less than the one obtained

directly.

The

integer 2 appears in each of these

presence indicates that the conclusion is a

Its

fortiori.

To deduce a minor limit

x and u as factors.
Add (5) and (6) ;
234.

to the

term of the second

order having

then

xu>

+ 2*+ 20.

(9)

By adding (7)
(8) we get the same expression.
COR. i. The total number of limits for xu which can be
and

deduced by the method under consideration

For it is

is 3*.

COR.

The sum

2.

of the primary limits for

pendent of m, n, and/.
For the sum of (i), (2),
235.

(3), ( 4) is

xu

is

inde

To prove that

where n denotes any integral symbol.

Suppose it is true for any one n.

Then
and
.-.

that

is

xy>nx-}-ny~2n-{-~L,

xy>x+yi.

x*y* >nx+Hy2n
is,

(Art. 187.)
i

(Art. 187.)

^>(+i)*+(*+iXy 2(*+i)+i.
true for any n, it is true for -fi.
But
hence for 2, and hence generally.

Hence,

if it is

true for

+ i+x+y-L

it

TOO Inference front one or more Equations

236.

Given

to find the relations

m-\-np,

xy = m ; yz = n ;
which necessarily

mnp, and

mn+p,

Write the equations as follows

zu=p ;
exist between

m+n+#,

x, y, z, u.

xy=m.

(2)

(ix)y=y-m.

(3)

x)(i

(i

(i)

x(iy)=xm.
y)

y-\-m.

(4)

yz=n.

(5)

y(iz)=yn.

(6)

(i-y)z=z-n.

(7)
(8)
(9)

u)=zp.

z(i

(10)

(n)

(i-z}u=u-p.
From(i),

w)==i

z) (i

(i

u+p.

(12)

($),(n)x}>z(i-z)u>m+n-p+u-2. (Art. 193.)

but z
that

The

is,

m-\-np<2

may be

truth of this result

It asserts that

that

(iz) = o,

o>xy-\-yz

is,

u.

verified as follows

zu+u

2>xy+{yz+u(i

2,

z)}

and xy<i ; hence their sum

nowyz+u
By the above method it may be shown

is

<2.

(iz)<i,

from(i)

(5)

(n)
(12)
(9)

(6)
(

(7)

that m+n-p<2-u

m +n+p<i+z+u
m-n+p<2-y

IQ )

m - n -p

(9)

m-n+p<2-z

(10)

<2

-y- Z

m -n-p<2-2Z

(12)
(8)

(9)

m+n+p<i+y+z

(10)

w+-/<i+7

(n)

m+n-p<i.+y+z-u

(12)

m.

of the form xy
By
will

( Hypothetical. )

taking (2) instead of (i) inequivalences involving

And so on.

oi
>

be obtained.

237.

To find

the condition or conditions

differejice (as the case

must

may

satisfy in order that it

limit to be negative.
and
Suppose that

which the sum or

of any two of the three m,

be)

be impossible

may

n are the known

data.

n, p,

for a given
If

occurs

positively in the given limit, then either of its minor limits

is to be substituted ; if it occurs negatively, then either of
its major limits is to be substituted.

Example of first

case.

When xy=m,yz=?t, zu=p,

yz.

xyzu >m-\-n -\-p

Now/
.

xyzu>m-\-n

and also

Hence

>z-\-u

>m+n

z.

u;

m-\-n>y-\-\

latter

m-\-n>y-\-z.

of these conditions

Example of second case.

Under the same conditions
xy(iz)u>m

is

and

also

>m

impossible.

as above,

np

and

Hence

>o,

y-\-u

the required conditions are

and

The

and

\-\-z-\-u.

<u,

i-\-z.

the required conditions are

n>i

u,

and mn>iz.
Both of these are possible.
To find the condition which any one of the three
m, 11, pj must satisfy, in order that it may be impossible for
a given limit to be negative.
Consider the condition for m.
Suppose first that n and / each enter positively into the
238.

given limit.

IO2

Inference from one or more Equations

Example. When xy=m,yz=n, zu=p,
xyzu >m-\-n -\-p yz.

Now

72>jy+

p>z-\-u

hence

n-\-p>y+z

and
and

i
i
i

>o,

>o;

-\-z-\-u

i.

(i)

i.

(2)

i.

(3)
(4)

>o.

Therefore the required conditions are

(2)
(3)

(4)

The

conditions

(2), (3), (4)

are evidently impossible.

when n and/ enter with opposite signs.

Example. Under the same conditions as above

Second ;

x(i

i -\-oc-\-y.

y)zu>mn-\-p

Now

n<y

and<,

p>z-{-ui

and

2+

>o;

i.

(i)

>-!.

(3)

>-*

(4)

Therefore the required conditions are

2.

m<x-\-z-\-u

(i)

<^i.

The
The

(2)
2.

(3)

i-

(4)

condition (2)

is

evidently impossible.

condition (3)

is

impossible

Similarly

m>x

+y

because

iz.

Third ; when n and /each enter negatively.

Example.

Under

the

same conditions

as above,

m.

of the form xy
xy(i

n<y

p<z
.

z)u>m

Now

(Hypothetical. )

and
and

103

i+z+u.

p
<z,

<u

(i)

n+p<y+z

(2)

(3)
(4)

<z+u.

Hence

the required conditions are

u,

(i)

>y+Lz.

(2)

m>y

(3)
(4)

The

conditions

are impossible.
239. If m, n, p, be all eliminated from a limit in the above
manner, only one of the resulting expressions is such as can be
(i), (2), (4)

positive.

Consider the third example of the preceding Article

n p i -\-z-\-u.
xy(i
z)u>m

Now

m>x-\-y

and
and
and

n<y

p<z

.*.

>o,

<z,

<z/,

np>xiz
>x

>x

(i)

(2)

+y-i-2Z

>x+y

(3)

(5)

>-y-u

(6)

>-2Z

(7)

>-z-u.

Hence

(4)

>-y~*

xy(iz)u>x-\-u

(8)
2

(i)
(2)

Z^-2

(3)
(4)

>-.?+- 1
i
i

(5)
(6)
(7)

(8)

IO4

Inference

from one or more Equations

Only (3) can be positive. It is the

means of the proposition of Art. 196.
Given a system

240.

ofn+i

limit obtained

by

equations of the form

where there are n middle terms ; then there are z n limits

n~l
of the first kind, n2
of the second kind, and generally

ri
The number
symbols

is

n
.

of primary terms formed by n independent

Hence there are
limits of
(Art. 152.)
2"

the first kind.

kind

limit of the second

is

formed by adding any pair

of terms which involve only one degree of contradiction.

But a term of the nth order has n contradictories of the

So for each of the 2* terms; but

degree.
(Art. 164.)
of
one-half
the
sums
so
obtained are different.
Hence
only
first

there are n2

n~l

four terms, which differ

For example

Now

second kind.
is formed
by adding any
with respect to two symbols only.

limits of the

limit of the third

kind

each term has

degree (Art. 164).

Hence

contradictories of the second

there are

n(n

of these only one-fourth are different.

^Zl)
I

I
>

sums; but

"

Therefore there are

2 -2 limits of the third kind.

th
Generally for the limits of the r kind.

They
factors.

are obtained

by adding terms, which

But there are to each term
contradictories of the r

differ in

_^

degree>

(Art. 164.)

m.

of the form xy

105

(Hypothetical. )

But of the sums obtained only

are different

hence

there are
its

COR. The

241.

total

sum of the

of the

limits is 3*.

ONE MIDDLE TERM OF THE SECOND ORDER.

Given

xyz=m,

to find limits to

xyzu,

yz=m
.-.

(Art. 188.)

y.

(i)

>m-\-n

z.

(2)

=xm

that

is

>

Nowyz>y-{-z
.-.

yz,

>m+n

and u( i yz) -=-un,

yz)
x(iyz)u>xm-\-un(iyz), (Art.

x( i
.

z.

yz U-=n,

xyzu>m+n

and
Again.

yzu=n-,

in terms of m, n, x,y,

etc.,

188.)

i+x -\-u-\~yz.

and

>o,

x(iyz)u>mn2+x+y+z+u.
and also
m n i+x+u.
>

COR. xu is greater than any one of the above limits.

For it is greater than xyzu, or x(iyz)u.
242. Given
xyz=m andy(\ z)u?i ;
to find minor limits to xyzu, etc., in terms
of m, n, x,y, z,

(3)

(4)

u.

y(i-z)u=n,
.

yzu=yu

n\

and xyz=m,
xyzu>mn+yuyz.

(Art. 188.)

yz>y+u

y.

(i)

>y+u

z.

(2)

>o-y.

(3 )

>o-*.

(4)

On

io6

certain forms of the

Therefore xyzu >mni-\-u.

>m

The

i+y

(i)

z+u.

(2)

>m-n-y.

(3)

>mnz.

(4)

any of the other terms of the fourth order

can be reduced in a similar manner.
limits for

ON CERTAIN FORMS OF THE

XVIII.

DISJUNCTIVE EQUATION.
243. One species of disjunctive equation consists of a
single positive character equated to the sum of a number of
The general form of the equation is
positive characters.

given in Art. 145, and several properties are there deduced.

Its chief function is to express the characters which a given
character comprises.

The simplest case is

of two terms only, as

when

the disjunct

member

consists

the reading of which is

s which are x are identical with those that are

together with those that are z.

244. Another species of disjunctive equation consists of
a single positive character x equated to the sum of a num

ber of terms multiplied by the character

forms are

and

We

x.

The

principal

x=x{y(i-z)+z(i-y)}.

(i)

x=x{yz+y

(2)

(i

-z)+z (i -y)}.

have already discussed

x=xyz;
and the remaining one of the four

x=x{yz+y(i-z)+(i-y)z+(i-y)(i-z)}
is

an

identity.

(Art. 153.)

107

Disjunctive Equation.

Equation (i) is read

s which are x are either y and not z or z and not.}

is

Equation (2)
U s which are x are either
5

read

and not y.
COR. Since

yz-\-y

y and

(iz)=y

z or

jy

and not z or

equation (2) can be reduced to

that

is,

Z7

which are x are either y or not

Similarly
Observation.
it is

put equal to

245.

The

x=x{z-\-(i

j>

and

z.

z)y}.

When
is a conditioning character.
the equation ceases to be conditional.

Here x
i,

equation

x=y+z
take the form

may

x=y+w,
where

is

but single and positive.

indefinite,

nature of the equation,

From

the

follows that

it

wy=o.
The

equation

x-=y+w

Only
and

it

is

is

read

which are x can be y

evidently the equational

mode

of stating the in

equality,
x>y.

246.

Conversion of the disjunctive equation

x=y+z.
the Conversion of a disjunctive equation is meant its
transformation into an equation having one of the terms of

By

the disjunct

member

as the

new

left-hand

member.

Ifx=y+z,
that

is,

which are y are identical with those which are x

excepting those which are z.
s

On

08

certain forms of the

If x=y-\-w,

theny=xw;
that

is,

Cf

which are y must be x.

If the original equation

is

written

then the form deduced by conversion

is

y<x.

247.

Contraposition of the disjunctive equation

x=y+z.
The

contraposition of a disjunctive equation consists in

deducing the expression for the contrary of the left-hand

member.
then
that

Ifx=y+z,
x=i y

z;

is,

U s which
are

y and

are not x comprise all

excepting those which
excepting those which are z.
If

then
that

is,

Observation.

x=y-\-w
x
;

expressed by
then

>y

x=i y w

expressed by
248. If

which are not x are not y.

If

is

is

x=y-\-w,

x=iy w

xy-\-z;

then

x<

y.

xy=y, and xz=z: but

not con

versely.

multiply by y,

then

xy=y 2 +yz

=y.
xz=z.
xy=y and xz=z-

Similarly

Given

(Art. 145.)

then
.

xy+xz=y+z,
x(y+z)=y+z

(Art. 291.)
t

Disjunctive Equation.
-

But

is

109

=y+z

not necessarily

(Art. 173.)

hence

it does not follow that

x=.y-\-z.
249. If x=y+w, where
has the meaning assigned to
in Art. 245
then the derived equation

it

xyy
is

equivalent to the original.

Now

(i

jy)

it is

such that

and

it is

is

single, since

is

single (Art. 145),

and

Hence

it is

not restricted in any other manner.

equivalent to w.
Observation.
x=y-\-w, x>y, and xy=y, express the
same truth. To express the truth considered by means of

an equation, either three symbols must be used, or else a

term of the second order introduced.
250. Inference

common

from two

disjunctive equations

having a

character.

First ;

when

the character occurs oppositely in the two

equations.

x=y+z.

=iy+z

X+X =T.+Z+Z

Add;

that

Second;

(i)

when

is,

x=i

(2)

-\-z-{-z

the character occurs similarly in the two

equations.

x=y+z.
Subtract

x
that

x =y+J.

(i)

is,

=z

x=x +zsf.

(2)

On

io

certain forms of the

When

Observation.

and

are each indefinite, the

former conclusion amounts to

Only

The

which are x are not x

conclusion under the same conditions amounts

latter

to

Some 7
For

251. If

by equation

x>z,

x=x(y+z),

which are x are x

(i).

then

x=x{y(i-z)+z(i-y)}- and

conversely.

x{iyz} = o,

.-.
...

that

is,

that

is,

that

is,

As

2y2z-\-2yz] =o,
x{i-y(i-z}-z(i-y)}=o,
x=x{y(i z)+z(i y)}.

x{i+y-{-z

this

must be

proof can be reversed, the converse proposition

true.

The

Observation.

an expression which

The

*2{i-y-*}2=o,

process here exemplified of squaring

equal to o is of great importance.

is

expression

*{i-X I -*)-*(i-j)}
is

of a form which cannot be negative

while the expression

x{iy~ z}
is

of a form which
252.

.-.

To

may be negative.
convert the equation

xy=x(\-z\
x
.>

COR.

=-(

i.

-*>,

Multiply by x,
then xy=x(i

z)

See Art. 289.

Disjunctive Equation.
which involves as a part of
i.e.

COR.

.*.

2.

truth

xy-xy(iz\

which are x and y are not

i

Multiply by

x(i

its

1 1

y)(i

z)

= o,

z.

y,

for neither

term can be negative

..x(i-y)=x(i-y)z;
7 s which are x and not y
i.e.
Similarly it may be shown that

are

z.

xz=x(iy);
and x(i
253.

To convert

z)=x(i

z)y.

the equation

x=x{y+(i-y)z}.

Now

x=xy+x(i

y)

z,

-z)+(i -x) z+

(i-x) (i-z\

This expression differs from the corresponding one ob

tained from

in

xz

having

COR.

in addition.

Multiply by x,

i.

then

hence xy
includes

COR.
then

is

xy=x (iz) -- xz ;
-\

not, as previously, identical with

x (12), but

it.

2.

Multiply by

o=* (i-^)(i -z)

y,

x
+-^(i^j) (i_^) +?- (i-y)z;

1 1

On

certain forms of the

none of the terms of the right-hand member

therefore, since

can be negative,

x(i-y)(i-z)=o,
is, x (i
y)=x (i

that

z=x (i

Similarly

y)

z.

+(*#) +*?-

x=x(y+z)
xy=xy(iz), and x (iy)=x (iy) z ; and

254. If
then

y)

con

versely.

x=x (y+z),
* aB *{y(x-*)+jr(i- y)},

...

(Art. 251.)

Multiply by y ; then

xy=xy(i-z).
Multiply by

To

(i)

y ; then
x(i-y)=x(i-y)z.

(2)

prove the converse.

From
From

that
.-.

xyz=o.

(i)

x(iy)

(2)

by(i)

Observation.

It

is

(i

z)

= o,

y z-\-yz}=o,
x=x(y+z).

x{i

is,

commonly

said that the disjunctive

proposition

7
is

are either y or z

equivalent to the four hypothetical propositions

/ s which are
z.
jy are not
(i)

Now

which are not y are z.

which are z are not y.
which are not z are y.

from

there follows,

(2)

(3)
(4)

i=y-{-z

by making

x=

in the

above equations

y=y(i-z).

(i)

i.-y=(i-y)z.
z=z(i-y).

(3)

-* =

(!-*)>>.

(2)

(4)

Disjunctive Equation.
But

(i)

and

(2)

\ i

without (3) and (4) are together equi

and so are (3) and (4) with
;

valent to the given equation

out (i) and (2).
255.

If

x=x{y+(i-y)z};

then

x (i-y)=x(i-y)

and conversely.

=x\yz+y (i -z)+(i -y) z],

=^{i-(i-^)(i-^)},
*(i-y)(i-*)=o;
that is, x (i
y)=x (i y) z,

(Art. 148.)

Since the proof can be reversed, the converse

proposition

must be

true.

XIX.

THE ARISTOTELIAN FORMS OF

INFERENCE.

256. I shall assume that the Aristotelian

moods do not
mere application of a general proposition to
an individual case, but of the combination of two
general
I shall further assume that the data are
propositions.
equations not of the first but of the second order in other
words, that the premises each involve one hypothesis.
They are mere cases of the general form of inference
consist of the

discussed in Article 212,

If

viz.

xy=m, and yz = #,

then xyz

and a fortiori

>

-f- ;/

y;

xz>m

-\-n-y.
The analytical reductions, which are sometimes required
to fit the premises of a mood for
being put into the above
formula,

correspond to the reductions indicated by the

significant letters in the

The

name

of the mood.

number of Aristotelian
a prime number suffices to throw doubt
upon
the completeness of the scheme.
Observation.

moods

is

fact that the

1 1

A ristotelian forms of Inference.

The

257.

The First Figu re.

17 s which are y are z ;
s which are x are y.

Barbara.

(i)

y=yz.

x=xy.

(i)

(2)

(2)

xz>x+yy;

.-.

=x
that

U
U

Celarent.

is,

X = xy.

(i)

z)>x+y

=x ;

(i)

(2)
(2)

y;

which are x are not

s.

which are 7 are z ;

Ts which are ^ and j are some.

Darii.

^==7*.

that

z.

which are y are not s

which are x are y.

y=y(l-z).
.-. #
(i
that

(Art. 219),

CTs which are # are

is,

^7

is,

^=e;.

(i)

(i)

(2)

(2)

which are x and z are some.

U
Z7

s which are jy are not z;

which are x and JF are some.

j=><(i-4
.-.

^=^

(i)

(i)

(2)

(2)

x(iz)>v+yy,
>v
}

that

^7

is,

258.

Tlie

which are x and not z are some.

Second Figure.

The

Cesare.

premises are

z=z(iy).

and x=xy.

(i)

(2)

premise contains iy, and the second y,

one of the two must be transformed. In this case it is the
transformation of the first premise which leads to the
Since the

first

ordinary conclusion.

From

(i)

zy

= o,

TJie Aristotelian forms

Hence

The

of Inference.

115

x (\z}>x+yy,

transformation

here

made corresponds
name of the mood.

to

that

indicated by the letter s in the

Observation.
The conclusion in the above case is
deduced more shortly by multiplying together the two
data.

Camestres.

From

zzy.

a.ndx=x(i

(i)

y=y(i-oc)

(2)

y).

(2)

x(lz)=X.

Here both equation (2) and the immediate conclusion

are transformed by a process which corresponds to
simple
conversion.
Festino.

z=z(iy).

(i)

From(i)

and xy=v.

(2)

(!-*>=> ,

/.

x (Lz)>v+yy,
>v.

Baroco.
It is

z=zy.

and x(i

(i)

y)

= v.

(2)

necessary to transform (i) so as to contain

Now

(i

)(iy)=iyz+yz
= i-y

Hence

^(i

The method

of reduction here

z}>v+i

(i

iy.
always,

by(i)

y)

>v.

to the contraposition of the

employed corresponds
major premise. The ana

shows the reason why the mood

lytical investigation
gave so much trouble to the Scholastic logicians, and
also shows how readily problems, which are difficult
to the unaided mind, may be solved with the
help of a
true organon.
The ductio per impossible is as follows.
If ^(i
z) is not greater than o, it must be equal to o;
for

it

cannot be negative.

The

1 1

Let

it

A ristotelian forms of Inference.

be equal to

o.

xz=x.

Then

Multiply (i) by x, then xz=xzy.

Hence
that

x=xy
x(iy) = o.

is

x(iy)=v
not =o

But
hence

x(i

x(iz)

therefore

is

a factor of the left-hand

In the rest of the moods y

of the conclusion.

the corresponding

factor.

259. The Third Figure.

y=yz.

Darapti.

andj/=y#.

(i)

Hence

(2)

xz>y+yy,
>y.

The

7 s

but

it

commonly said to be
which are x and z are some

conclusion

is

evidently amounts to more, viz.,

Ts which are x and z include those which are y.

y=y(iz).

Felapton.

Hence

a.ndy=yx.

(i)

x(i

(2)

j,

z)>y+y
>y.

Disamis.

yz=v.

andy=yx.

(i)

Hence

(2)

xz>v+yy,
>v.

Datisi.

yyz.

and xy=v.

(i)

Hence
Bocardo.

Hence
Ferison.

Hence
260.

(2)

greater than o.

is

iy

Here

Observation.

member

by

2;)

(2)

xz>v.

y(iz) = v.

(i)

andjv=^.

(2)

x(iz)>v.

y=y(i-z}.

(i)

^(i

Hence

(2)

z)>v.

77^ /^r/^ 7%^^.

z=zy. (i)

Bamalip.

and xy = v.

and;- =_y^.

xz>y+zy,

(2)

is

The Aristotelian forms of Inference.

The
But

conclusion

it

is commonly said
7 s which are x and z
really amounts to

7
or, in

1 1

to be

are some.

which are x include those that are

2,

other words,

Only

The

character

the character

which are x can be

z.

a conditio sine qua non with respect to

is

z.

Calemes. z=zy.

audy=y(i
z>y+z
y

(i)

Hence

x)

(i

x).

(2)

=z.

The
but

it

conclusion in this form means

Only 7 s which are not x can be z

can be converted into

x=x(iz),
that

is,

^7

Dimatis. zy=v.

which are x are not z.

and jy =j-;r.

(i)

Hence
Fesapo.

(2)

xz>v.

z=z (iy).
From

andy=yz.

(i)

(2)

y=y(i-z),

(i)

x (i-z)>y+y-y,
>y.

This conclusion
the character

(i

not particular merely, but asserts that

z) is a conditio sine qua non with respect

is

to the character y.
Fresison.

z=z(iy).
From

y=y(

(i)
.\

and xy=v.

(i)
I

(2)

z ),

x(iz)>v+yy,
>v.

Here the conclusion

investigation

of

is

For an
merely particular.
Syllogistic forms see the

De Morgan s

Examples, page 135.

1 1

On

Probability.

ON PROBABILITY.

XX.

261. It is important to draw a sharp distinction between

the combining of two general propositions, and the appli
The
cation of a general proposition to an individual case.
is investigated by the Theory of
the latter by the Theory of Probability.
Any

former of these processes

Necessity

one who clearly perceives

this distinction, is in

no danger

the absurdity of maintaining that true

syllogistic reasoning involves the fallacy of the Circle.
262. The Probability that a given member of a universe

of

into

falling

has a character x

is

measured by

the arithmetical value of

referred to that universe.

The

probability here spoken

of,

or at least

its

measure,

The value of this objective proba

in which case we consider
known
be
not
fully
bility may
The latter depends upon the
its most probable value.
is

entirely objective.

state of our

knowledge concerning the objective proba

consequence varies with that state.
263. Ifx is equivalent to a function of a number of charac
ters, the probability of an individual
being x is equal to
bility,

and

in

that function taken arithmetically.

For every

logical equation

is

true arithmetically.

As an

example,
let

x=2-,
z

then#=>>+-(i-*)

Hence

(Art. 199.)

the results of the preceding Articles may be made

by taking the right-hand

to yield theorems in Probability

arithmetically, and supposing

an individual member of the universe.

members of the equations

to denote

On

1 1

Probability.

which
264. The probability of a
by the arithmetical value of y in Uxy

is

being

y is measured

Let

Ux=U".

The

probability of a
arithmetical value of y in

But
a

Ux

is

U which

is

value ofy in

identical

x being y

Uxy

is

it is

measured by the

is

(Art. 262.)

therefore the probability of

measured by the arithmetical

COR. The probability of a

For

being

U y.
with U ;

U which

equal to that of y in xy
.

,\

equal to

is

being y

is

equal

to

(Art. 64.)

oc

equal to -^.

265. If the probability of a

U which

is

being y is equal
then either is

U which not x being y,

equal to the probability of a U being y.
to the

probability of a

is

xy

(\x]y
-

~=>

ix

xy _^yxy

Any one

1x

given.

of the three equations

i-x

expresses the condition for x and y being really indepen

dent of one another.
(Art. 62.)
which is x being y,
266. Required the probability of a

having given that

(i)

or (2)

U{x=y+z};

U{x=yz\.

On

120
First Case.

Probability.

Probability required

(Art. 264.

="?

00

=/+*
Second

case.

267. Given
required
or x.

t/ie

>

Probability required

x=x +x +x
1

probability of a

=?

+Xt;

which

= *fo+*

Probability required

COR.)

is

being either

Art<

2 64),

268.

U which

required (i) the probability that a

the probability that it is not y.

Since

is

is

y,

and

(2)

zu=z(iu)u(iz),
x=y+z(iu)u(iz).

Now x cannot be negative, and z(i u) and u(iz) are

exclusive of one another ; therefore y must include
u(i
z).
Hence
(i.)

x=y{iu(iz)}+z(iu).

Probability that a

U which

y+zu
^

y+zu
y

u-\-uz

is

\?>

=~

On
Suppose that uz

not known.

is

y~U

prob.

>.)

Probability.

>

y+zu

_,

Probability that a

and

U which

Then we

affirm that

<--.

y+zu

is

is

not;/

=*(*_

zzu

_
Hence

if

s&

is

not known,
prob.

269.

The

is

y+z-u

U
to

is x, and the pro

find the probability

y.

data are

Now_j>=

affirm that

-.

<-

Given the probability that a

a
which is x is y ;

bability that

that a

we can

x=p, and xy =pq.

It is

required to findj?.

~(I

Hence
The problem can

also

be investigated as follows

*)y-

Hence
Hence

J<i

j/

<

x+xy
p -\-pj

and
and

(2)

Fundamental

122

relations, etc.

FUNDAMENTAL RELATIONS BETWEEN

THE SINGLE FUNCTIONS OF A NUMBER

XXI.

OF INDEPENDENT CHARACTERS.
270.

By

the

fundamental relations between the single

x^ x*...x n is meant the different

forms of the relation connecting the part x l x 2 ...x n with
the other parts into which the universe is divided by the

functions of the characters

characters.

The

two independent characters

They suffice to combine any

relations in the case of

190.

are given in Article

number of premises, which are linked together by common

As, how
characters after the manner shown in Article 240.
problem are not necessarily of that
fundamental
important to consider all the

ever, the data of a

nature,

it

is

which exist in a given case.

271. To find the fundamental relations, when there are

relations

three characters x, y,

Expand

z.

the term which

is

the opposite of xyz,

(ix)(i-~y)(i--z)=i--xy*+y*+*x+ xy*y*
Expand a term which involves two degrees

(i)

of contradic

tion,

(ix)(iy)z=zxzyz+xyz.
Expand

(2)

a term which involves one degree of contradic

tion,
(

x)yz =yz

There are other two equations

two similar to

is

(3)

hence,

xyz.
similar to (2),

(3 )

and other

if

3
also counted, there are 2 altogether.

which exist may be derived by

For example
relations.
fundamental
transforming these
becomes
then
into
z
iz,
(2)
change

The

other relations

Fundamental
272.

When

there are

relations,

etc.

n characters, there are

2"

123
fundamental

relations.

For each primary part into which the universe is divided

by the n characters yields a fundamental relation ; and there
n
are z such parts.
(Art. 152.)
273. Consider the relation

(ix)(iy)(iz)+x+y+zyzzxxy+xyz=i.
If the arithmetical values of

any seven of the eight terms

known, then that of the remaining term can be found.
If the arithmetical values of
any six of them are known,
then that of the remainder can be found ; and so on.
Suppose that the remainder can be reduced to two parts
/! and /2 having the same sign.
are

Then /H-f a
.-.

Hence A and

tl

=/

= nt

must each be less than

Suppose that the two parts t^ and f2l to which the
mainder is reduced, occur with opposite signs.
/2

".

Then

/1

-/

and

tz

re

>n.

274. The following example illustrates the manner

which these relations can be applied to solve questions

in
in

Probability.

x=d ; y=b ; xz=dp ; yz=bq ;

Given

and
required

By

(2)

(i

x) (i

(i

*) (i-jy)

y)

z=o

z.

(Art. 271)
.

.-.

z=xz+yz

z=zxzyz+xyz.

xyz

by

z<xz+yzxyz-\-\. (i)

<xz+yzyxz-\-i. (2)
<xz+yz.
.

z<i

a+ap.

(3)

(i)
(2)
(3)

last

datum;

(Art. 188.)

Fundamental

124

(\x)yz=.yz-\-zxzyz, by
.

z=xz-\-(i
.

It

(ix)yz=yzxyz.

relation (3) (Art. 271)

By

etc.

relations,

may be shown

the last datum.

x)yz.

z>dp.

in a similar

manner

that

z>bq.

Boole deduces these results by a

much

longer process.

GENERAL METHOD OF DEDUCING A

CONCLUSION OF A REQUIRED FORM FROM

XXII.

GIVEN DATA.
275. When we know a necessary relation existing be
tween the forms of the given data and of the required con
clusion, and between these forms alone, we can find the

conclusion by simply substituting the data in the necessary

If the known relation involves other forms in

relation.

addition, it
conclusion.

may be made
If,

however,

to yield limits to the required

the forms of the given data do

all

not exist in the relation, the limits deduced may not coin
cide with the actual limits necessitated by the data.
is

276. I have discovered a method, by means of which it

easy to find the necessary relation which exists between

the forms of any given data and the form of the required
The forms are supposed to be functions of

conclusion.

single (Art. 118) symbols.

Equate the form of the conclusion to the

sum of the forms

of the data each multiplied by a constant, and with a con

stant added ; then substitute each of the sets of singular
values

The equations
which the characters can assume.
the relations which necessarily exist be

obtained express

tween the constants ; they may

suffice to

determine each of the

General method,

etc.

125

constants exactly, or they may require one or more

of the
constants to have indefinite values.
It is supposed that the functions do not contain
any

inverse factors.

277. Given

x = m,x(i

y)

= n,(ix)y=J ; required xy.

Since x and y are given to be single and

(we may sup
pose) positive, each can assume the values i and o. Hence

from
from
from
from

x=i,y=i,
x=i,y = o,

x=o,y=i
^=0,7 = 0,

The determinant
equations

i=a+b.
= a+b+c.

(i)

=a

(2)
-\-d.

0=0.

of the

right-hand

(3)

(4)

members

of these

is
i

1000
which

is

equal

to

Hence

i.

a definite

relation exists

between the form of the conclusion and the forms of the

Definite values can be found for the constants,
given data.
viz.

Hence

xyxx
m

(i

y),

n.

The problem here discussed can of course be solved by

much simpler methods. Its solution by the general method
is

introduced to illustrate that method.

2 78.

Given y

= m, xyn;

required x.

Let xa-\-by -\-cxy.

Then i=a+b+c.
1=0.
o

<7.

(i)
(2)

(4)

General method of deducing a Conclusion

26

The determinant

of the right-hand
i

members

is

ooo
ooo
Since the elements of the fourth column vanish, the
is o.
Hence there is no definite

value of the determinant

between the forms considered.

relation

The

fourth row

left-hand

The

member

and

o.

and of

o,

made
and

identical

restricting

by putting

L=a+b+c.

(i)

=a.

(3)

The determinant

of these equations

is

ii

00
o

the value of which

is

And

i.

v= -o

c=i.

o
x= --o
o
o

Hence

=(!-)+.
O
Since

lies

between

and

o
x>n

and

o,

<i

(2).

in

within the limits

Then we have
o

but the

not identical with that of

is

equations can be

instead of
i

identical with the second;

is

of (4)

m-\-n.

of a required form from given data.

The fundamental

Observation.

2 7

relations

(ix) (i-y) = -x-y+xy,

and (i y) x=xxy,
i

readily

show

that
x<i

y+xy

The

and

>xy.

an instance of the general

following problem
in
the
problem investigated
preceding Article
The probability that it thunders upon a given day is /,
the probability that it both thunders and hails is q, but of
the connection of the two phenomena of thunder and hail,
279.

is

nothing further

Required the
on the proposed day. Boole,

hails

it

Laws of Thought,
Let

supposed to be known.

is

probability that

p.

276.

U = a succession of

states of the

atmosphere

at a

given place, an individual state being of the length of a

day.

x= containing a

hailstorm;

Then

y=p

containing a thunderstorm.
t

xy=q.

Hence, by the preceding Article,

Hence x cannot be
less

than

greater than

q+

/,

and cannot be

q.

280. Let us take the three data

y = m, xy = n,

y=/,

which are equivalent to two independent data, and

inquire
how this is detected by the method, when we seek for the
relation connecting

with the given functions.

Let x=za-\-by-\-cxy+d(iy).
(i)

+d.

(2)

(4)

(3)

128 General method ofdeducing a Concision

The determinant

Now

this

of these equations

determinant has the sum of

fourth columns identical with

value

is

minant

is

its

first

its

column

second and
hence its
;

Thus the

analytical reason why the deter

vanishes indicates the real reason of the insuffi
o.

ciency of the data.

281

Given xy+x(i y) + (ix)y =/,

andx(Ly)+(ix)y+(ix) (iy)=q;
to find x (iy)+y (ix).

Let x(iy)+y(ix)=a+Z>{x+(ix)y}

Then o=a-{-b.
(3)
(4)

The determinant

is

which the second and third rows are identical. But the
equations (2) and (3) have the same left-hand member.
Hence one of them may be struck out.

in

Then

Hence

The above
blem

x
is

(i

y)+y

(i

x)=

the general solution of the following pro

of a required form from given data.

29

The probability that one or both of two events happen is

p, that one or both of them fail is q.
What is the proba
bility that

only one of these happens

It is

282.

Given

Let

x, y, xz, yz,

_!+/+
and (ix)(i

z=ax+by+cxz+dyz+e (i

y) z ; required
x) (i

y)

z.

z.

Then by means

of the eight sets of singular values,

which the three characters can assume, we obtain the
equa
tions

i=a

(3)
(4)

i=

o=

b.

The determinant

(5)
(6)

=
= 0.

(7)
(8)

expression

is

Since the determinant contains five columns and seven

rows, two of the equations are identical or must be made
identical.
Now the sum of the third and fifth rows is
identical with the

equations are
,

first,

and

but the values of the

corresponding
Denote this value by

respectively.

subject to the restriction that

nor less than

rows

is

i.

it

cannot be greater than

Also the sum of the fourth and sixth

identical with the second,

i

and the values of the

cor-

are 2

Since the limits of

are

and o; hence

than

nor

less

than

</=-,

and
where

i,

the limits of

^-cannot

be greater

o.

*=5^H-(i -*)(!--.?)*

Thus

a more general form of the pro

blem considered in Article 274.
in which the equa
283. It is evident from the manner
Observation.

This

is

when an additional datum is given,

in the equations obtained for the
made
alteration
the only
constant into some
fewer data is the introduction of the new
tions are derived, that

of the equations.

considered involve three

Suppose that the forms
Then
z as a factor.
has
each
that
and
characters x, y, z,
be dispensed
characters
the
of
may
the term independent
reduce to the form
with and four of the eight equations
284.

= 0.

of each of the terms, the four sets of

For, z being a factor
the
in which z=o cause each side of
values
singular
identity to

become

o.

On

ON BOOLE S GENERAL METHOD.

XXIII.
285.

Boole s General Method.

The method

of the preceding chapter, while ad

mirably adapted for problems in which the data are what
may be called explicit equations, is not adapted for pro
blems in which the data are implicit
For the
equations.
latter

we

require the special method invented by Boole.

I
propose to consider the elements of that method in the
light of the principles of this work.

To

286.
is

Since
value

above

o.

(ix) = o.

equation f(x)

= o,

x(ix) = o,
By

identity,

Hence

x from an

eliminate

given that

x can assume the value

putting these singular values of

it

and the

into the

we get/(i)=a, and/(o) = &

/(i)*+/(o)(i-#)=o.
(i)-/(o)}*=-/(o).

Multiply together (i) and

then
{/(i)-/(o)} *

but

when

(i)

(2),

(i-*)= -/(i)/ (o);

*(i-#) =
,

/./(i)/(o)=o.
Observation i.
The condition /(i)/(o) = o will be
fied independently of
x(i-x) = o, if/(i)-/( ) = o

satis
;

that

the given function does not contain x.

It may also
be satisfied independently of either of these
conditions; for
/( J ) /() generally involves fractional symbols.
Observation 2.
The equation /(i)/(o) o is true, when
is, if

#=i, and when x = o-t therefore when

x=,
o

where

o
denotes any value lying between i and o.
The contradic
tory values of x are united into one in the same manner as
those of a in Art. 278.
Hence

On

132

Boole s General Method.

expresses a relation among certain characters of the universe

U, which holds true whatever the value of the character x.

If the given equation isf(x)

Observation 3.
m; it can
be shown that the result of the elimination of x from the

equation

is

287. To eliminate x and y from an equation f(x, y) = o

x and y being such that x (i x) = o, andy(iy) = o.
As x and y are independent, x can be eliminated first,
and then y. The elimination of x by the preceding Article
,

gives

a double application of the

By

we

to y,

same process with respect

get

/(i, i)/(i, o)/(o, i)/(o, o)

288.

The method

is

example.
Let the equation be

To

= o.

by the following simple

x=y-\-z.

eliminate x, write the equation in the form

Now f(i)=i
that

illustrated

f(x)=zx y z=o.
(y+z), and/(o) =

jy=o,

is

provided y and z as well as x are known to be single and

positive.

To

eliminate y, write the equation in the form

Then
that

(i+zx)(z- x)=o,
z(ix) = o.

is,

289.

:77z

ra:0//

power of any

expression which

is

a func

tion of single symbols cannot be negative.

For the expression can be expanded in terms of the
primary parts into which the universe is divided by the

The second power of the expansion is equal to

symbols.
the sum of the second powers of the terms of the expansion,
The second power of each term
as the products all vanish.

On

Boole

General Method.

133

composed of the second power of the co-efficient multi

by the second power of the part. The former cannot
be negative (Art. 77) and the latter cannot be negative
is

plied

(Art. 138)

the

sum

hence

of a

cannot be negative. But

terms, each of which cannot be

their product

number of

negative, cannot itself be negative.

290. To prove that x -\-y-\-z
2xy

2xz -\-2yz cannot be

negative.

For

{xyz}

equal to
It is evident that the expression will be as negative as
Its value is then z.
possible, when #=1,^=1.
it is

291. If each equation of the system

= 0, # 2 = O, ... XH =0,
member of a form which cannot

^!

has

its

left-hand

be negative ;

then the equation

#i+#a + "-+#*=0
is

equivalent to the system of equations.

When the equations of the system are true, the

pound equation

is

true, for

it

is

com

formed by adding the

equations of the system.

(Art. 115.)
When the compound equation is true, each equation of
the system must be true ; because the left-hand member of
the compound equation is a sum of functions each of which
cannot be negative.
292. If any one equation of the above system, x r =o,
has a left-hand member x r of a form which can be negative ;
then

x r2 =o and not x r =o must

be

put

into the

compound

equation.
If x r were put into the compound equation, then the
truth of each of the original equations could not be inferred
necessarily from the truth of the compound equation.

But xr2 cannot be negative (Art 289).

Hence when x r2 = o

truth and the truth

compound equation,
of the others can be inferred from the truth of the com
is

put into the

its

pound equation.
whether x r

is

or

from x r2 =o we can deduce xr =o,

not of a form which can be negative.

Now
is

On

134

Boole s General Method.

Hence the equations of the system can be inferred back

from the truth of the compound equation, provided x r be
squared.
293. Example.

The equation

x=y

gives

y = o.

Now x y

is

of a

form which can be negative.

Hence it is (x y) 2 o
which must be put into an equation, which is to be equiva
lent to x-=y together with other equations.
It is curious to

observe that (x

that

is,

x(iy) = o,

that

is,

x=xy,

Thus
sidered

y)

andjy(i

=o

x)

gives

=o

andjy=_y,x.

the single equation x-=y must for the purpose con

be resolved into the two equations x=xy and

EXAMPLES.
the follow
page 56 of his Syllabus, De Morgan gives
the name of
under
of
forms
inference,
ing twenty-four
Metaphysical Syllogisms
1. Dependent of dependent is dependent.

AT

2.
3.

Essential of independent
Independent of essential

is

independent.

is

independent.

essential.

4.

Essential of essential

5.

Dependent of

inessential

is

inessential.

6.

Inessential of dependent
Repugnant of alternative

is

inessential.

7.

8.
9.

is

is

Repugnant of independent
Independent of alternative

10. Alternative of

repugnant

is

dependent.
is

inalternative.

is

irrepugnant.

essential.

11. Alternative of inessential

is irrepugnant,
Inessential of repugnant is inalternative.
13. Dependent of repugnant is repugnant.

12.

14.

15.

Essential of irrepugnant is irrepugnant.

Irrepugnant of repugnant is independent.

16. Essential of alternative

17.

Dependent of

is

alternative.

inalternative

18. Inalternative of alternative

19.

20.
21.

22.

inalternative.

Repugnant of essential is repugnant.

Repugnant of irrepugnant is inessential.
Irrepugnant of dependent is irrepugnant.
Alternative of dependent is alternative.

23. Alternative of inalternative

24.

is

is inessential.

Inalternative of essential

is

is

independent.

inalternative.

In each case three characters of a given universe are

Let the first-mentioned character be denoted
considered.
The first term in
the second
y, the third by z.

by

x,

by

the proposition expresses the relation existing between

136

Examples.

and y

the second that existing between

third the relation

and

which

y and z ; and the

consequence between x

exists in

z.

By x

being dependent on y
independent of

is

,,

x (i
x (i

meant
,,

essential to

y)

= o.

y)

v.

(ix)y=o.

inessential to

(i

#).J;=P.

repugnant to
irrepugnant to

xy=o.
xy=v.

,,

(ix)(iy) = o.
(ix)(i y) = v.

alternative to

inalternative to

Some of these syllogisms coincide with Aristotelian forms,

the latter have the meaning
assigned to them in Articles
256-260.
The truth of the conclusions can be proved by
means of the theorem enuntiated in Article 256 excepting
those of the sixteenth and twenty-second forms, which are
not true as enuntiated.
if

Consider the sixteenth.

xy=y

Substitute xy fory in

then
that

is

data are

y+(iy)z=i.

(2).

(2),

xy+(ixy)z=i,
(ixy) (iz) = o.

is,

Hence
It

The

and

(i)

xy, but not y,

evident that

is

is

alternative to z with respect to i.

alternative to z with respect to y.

The conclusion of the twenty-second form must be corrected

by introducing the same condition.
25. For every man in the house there is a
person who
is aged
some of the men are not aged. It follows that
some persons in the house are not men. De Morgan,
;

Syllabus, p. 29.

Let

people in the house

#/

Then
From

m<d

= man,

and

(i).

im>

(i)

but from (2)

/.

and

is

im

#=aged.

m(ia) = v.

id

greater than
is

o.

greater than

o.

(2).

Examples.

37

Most men in a certain company have coats most

same company have waistcoats therefore some
De Morgan.
in the company have coats and waistcoats.
27. Of the inhabitants of a certain country those having
black hair form one-tenth, and those who are short-sighted
26.

men

in the

What

proportion must be (i) black-haired

and not short-sighted,
and
not
black-haired
short-sighted, (4) neither black(3)
haired nor short-sighted ?
28. Eighteen out of twenty-one F s are X; fifteen out
one-twelfth.

and

short-sighted, (2) black-haired

of twenty-one

Fs

are

Z;

therefore twelve

Zs

are X.

De

Morgan.

Here the

Fs

form the universe, and the arithmetical

value of the universe

is

given.

which is F; some Z s are

For every Z there is an
De Morgan, Syllabus, p. 29.
not F s.
The two conclusions which De Morgan deduces can be
proved in the following manner
and
The data are xy>z (i).
z(iy) = v. (2).
29.

From

z]yy-{-v_ z,

(i

(2)
.-.

x (iz)

>z

-\-y-\- v

zy

>v.

Again, from (i)

30.

Of a

certain

x) (i

(i

y)

>

xy+z,

community, which comprises 12,000

persons, three-fourths are Presbyterian, two-thirds Liberal,

and one-sixth on the electoral roll. What is the least and

what the greatest possible strength of each of the twelve

classes formed by taking two of the characters; as, for
example, Presbyterian and Liberal, Presbyterian and not
Also what is the greatest and what the least
Liberal, etc. ?
possible strength of each of the eight classes formed by
taking three characters

as, for

and not Liberal and not on the

31.

example, not Presbyterian

electoral roll ?

In a series of twelve throws with two dice, ace and

138

Examples.

two occurred together in one-half of the throws, ace and

not-three occurred together in one-third, and ace occurred
in two-thirds.
often, at least, must two and three

How

have occurred together ?

32. In a series of throws with two dice, ace and two
occurred together in one-sixth ; ace and four in one-twelfth.
How often, at least, must two and not four have occurred
together ?

Of

33.

the matriculated students of a certain University,

three-fourths are under the age of
;

two-thirds are medical

21

and

five-sixths are natives of the British Islands.

Also

the medical under the age of 21 form one-half; and the

colonial or foreign under the same age one-twelfth.
The

number of students is 2400. What is the least, and what

the greatest possible strength of each of the
remaining ten
classes of the second order ; also of each of the eight classes
of the third order

From

34.

follows Q,
cannot both be true

and S
be true.

U= the

Let

is

true,

Then

and from
;

therefore

follows

Pand

is

but

p-= when

collection of times considered.

when

S;

R cannot both

true, etc.

the data are

p=pq

(i).

r=rs

gs=o.

(2).

(3).

Multiply together (i) and (2),

then pr=pqrS)

= o.

Every X

by

(3).

one only of the two,

or Q; every Fis
both
and Q, except when
is Af, and then it is neither ;
therefore no XisY.
De Morgan, Formal Logic, p. 124.
Let U denote the subject of which x, jy,/, q are charac
35.

is

ters.

Then

is

not a character of

/,

but of Up.

Hence

the data are

Multiply together the two equations

then each of the

139

Examples.
four terms of the right-hand
^K

36.

=
?)

every

member

because

will vanish,

and (i-/)/ 1 _ w =o.

either P, Q, or J? ; but every

is

every R

J/",

<2

-/) =

Hence xy = o.

Every
is

is

J/; therefore every

is

is

J/.

M,
De

Morgan.
37.
is

38.

What

Dilemma.

neither

B nor

Constructive Dilemma.

which are 2 are y

Therefore

The

is

either

is

Therefore what

C.

and 6^

is

or

is

C;

what

which are x are jy

are either

is

Z>

not A.

or not

x and

s
0.

are^.

x=xy

data are

Substitute for

z=zy

(i)

and *+(i x)z=i

^ and z in (3) by means of

Then

(2)

(3)

and

(i)

(2).

xy+zyxz=i,
i-\-xz

x+z

=
Hence
39.

1.

^7

arej.

The Method

of Difference, which bulks so largely in

some Systems of Logic, can be disposed of in a few lines.

In the states of a certain substance when the property A
and when the
is present, the property B is also present
;

property

is

absent, the

property

is

also

absent.

What follows ?
The data are

a=ab
From

(2)

(i-0) = (i-*)(i-).

(i)

(2)

ia=-iab-\-ab,
.-.

.-.

b=ab,

b=a

from

(i).

is present are
the states in which the property
is present.
identical with the states in which the property

Hence

40

Examples.

40. The members of a board were all of them either

bondholders or shareholders, but not both ; and the bond
What con
holders, as it happened, were all on the board.
clusion can be drawn ? Venn, Mind, vol. i. p. 487, discussed

by Jevons, Principles of Science, p. 90.

Let C/=members of the company.
=bondholder s= shareholder.
^director

Then
and
Put

d=d{b(i-s)+s(i-b)}.

(i) into the

(2)

form

d{i-b(\-s)-s (i-)} =

and

(i)

bbd.

(2) into the form

(i-0 = o;

then their left-hand members consist of functions which

Hence

be negative.
together equivalent to
cannot

these

two

equations

d{i-b(i-s)-s(i-b)}+b(i-d) = o.

are

(Art. 291.)

By means

of this equation any one of the three d, b, s,

can be expressed in terms of the remaining two. For

example
,

bs+b

(i

= b(is)-{

- *) +

^-(i

-) *+o (i -b) (i -

&)s from the nature of

(i
o
with the condition bs=o,

=b-{

by taking

in

s).

d,

bso.

For the sake of illustrating its application I have here

employed Boole s Method of dealing with implicit equa
tions.
The conclusions can be deduced by a simpler
method. Thus

and

d=d(b+s)

(i).

db=b.

(2).

(Art. 251.)

Examples.

d=b+ds ;

Hence
which

is

41

equivalent to

Multiply by
then

d=b-\

s.

s,

ds=bs+ds,
bs=o.
.

In a certain class of sub

41. It has been observed that
are present, the pro
A
and
the
where
properties
stances,
C are present, A is
and
is present ; and where
C
perty
Does it follow that B is present where C and A

present.
are present?

Let
a

The data
Hence

7= the

class of substances.

= having the property

are

ab=abc

A,

etc.

bcbca.

and

ab(L-c)+bc(*-a) = o.

(Art. 291.)

-=i.

Now

(Art. 78.)

And

is

impossible.

(Art. 78.)

Therefore ab

(IC) = Q

and fa(ia) = o; which are the given equations taken

(Art. 79.)
each by itself. The sixth term is cancelled.

Add

co-efficient
together the terms whose

ca

then

Hence

it

= abc+(i

does not follow that

are present.

is
,

b).

is

present where

C and

42

Examples.

42. Suppose that an analysis of the

properties of a par
ticular class of substances has led to the
following

general

conclusions, viz.

and
ist, That wherever the properties
either the property C, or the
property D,
but they are not jointly present.
2d,

B are combined,
is

present also

That wherever the properties B and C are combined,

A and are either both present with them,

the properties
or both absent.

and
are both
3d, That wherever the properties
and
are both absent also ; and
absent, the properties

vice versa,

and

where the properties

D are both absent,

C and

B are both absent also.

Let it then be required from the above to determine

what may be concluded in any particular instance from the
presence of the property
absence of the properties

with respect to the presence or

B and

C, paying no regard to the

Boole, Laws of Thought, p. 118.
Here the universe under consideration is a particular
class of Substances.
Let a= having the property
;

property

D.

without the property

Then

ab=ab

A;

and so on.

(c+d).

(i)

bc=bc(ad+a d ).
a b =c d
equation, when put into

(2)

Since the

first

(3)

the form

contains a function which can be negative, it must be

squared before being put into the equivalent equation.
This squaring amounts to writing the equation
(Art. 292.)
in the form

ab=ab{cd +dc
Similarly (3)

equation

must be squared.

}.

Hence

the equivalent

is

ab{i-cd -dc }+bc{i-ad-a d }+a

(i-c d

+c d (i-a
As d

enters into this equation,

and

is

)
)

= o.

not to enter into

Examples.
the required conclusion,

it

143

must be eliminated by

Art. 286.

Now
f(i)=abc+bca

and
is

d may

true whatever value

positive

and

+a

= ab

have, consistent with being

single.

a=

Hence

-bc
o

c
+-b
o

Hence

the substances, which have the property A, are

identical with those which have the property C and are
without the property B, together with an indefinite part of

those which are without the property C.

The conclusion can be deduced very simply by the

fol

lowing method

ab=ab(c+d\

(i)

bc=bc{ad+(L-a)(i-d)}.

(2)

(x-. fl)(i_J)=(i-,)(i-<f).

(3)

Multiply together (i) and (2),

then
abc=abc(acd-\-ad\
2 abed ;

but abc

is

single,
.

abc=Q.

Again, multiply (3) by c,

then c(i
a)(i
.

(4)

=o
=
c(iab) Q,
&)

by

(4)

This expression for a coincides with that obtained above.

43. Find also what may be concluded from the presence

and B.
of the property C with reference to the properties
44. In a certain class of substances, those which have

144

Examples.

and the property X, together with those

B and are without the property X,
are identical with those which have the
Also
property C.
those which have the property
and the property X,
together with those which have the property E and are
the property

which have the property

without the property X, are identical with those which have

the property F.
What relation necessarily exists between
the properties A, B, C, D, E, F?
The data are

ax+b(i

(i)

*=L

From(i)

From

x)=c.

f_

x=---

(2)

Hence, as the equations are true simultaneously,

c-b_f-e
~

.-.

45.

af+bd+ce*=ac+bf+cd.

In a certain succession of throws with

those

six dice,

having a two and an ace, together with those having a three

and a six, were identical with the whole and those having
a four and an ace, together with those having a five and a
What relations existed
six, were identical with the whole.
;

between having a two, having a

having a

Having

The

three, having a four,

five ?

a two

here means

having one two at

least.

equations are
(i)
(2)

Multiply (i) by d and (2) by b, and subtract

then (adbc}x=db ;

_ d-b
ad

be

and

Examples.
Expand

~^bc

the universe

is

in terms of the

divided by

a, b,

P rimary P arts

c,

each of the co-efficients of

sible,

Then

d.

since

must be

which

int

o.

is

pos

We

thus

=o

(2)

obtain

=o

(i)

-0)(z-*)fl/so

(3)

(i-*)(i-</)

The

(i-a)J(i-r)(i-</)

l-*l-Jl-<=o

equations (i) and (2) can be combined into the

equivalent equation

Similarly (3)

Equation

(5)

and

b(i-c)(i-d) = o.

(5)

(4) into

(6)

-a)(i -b)d=o.
can be put into the form
(i

*=*{H-(i-*)rf},
that

is,

four,

the throws having a three had a four

not a

or, if

five.

Similarly

From

d=d{a+(i a)b}.
y other two

the expansion for

relations

can be

obtained, viz.

a=a{c+(i-c)d}
and c=c{a+(i

a)b],

46. Boole solves the

following highly complex problem
Let the observations of a class of natural
productions be
supposed to have led to the following general results
:

That in whichsoever of these productions the

pro
perties A and C are missing, the property
is
found,
together with one of the properties
and D, but not with
ist,

both.

That wherever the properties

and
are found
and
will either both
missing, the properties
be found, or both be missing.
wherever the property^ is found in
3d,
conjunction
^That
with either
or E, or both of them, there either the
pro
or the property
will be found, but not both of
perty
them.
And conversely, wherever the property
or
is
2d,

while

is

46

Examples.

found

singly, there the property

tion with either

B or E,

A will

be found in conjunc

or both of them.

then be required to ascertain, first, what in any

particular instance may be concluded from the ascertained

Let

it

presence of the property A, with reference to the properties

C, and D; also whether any relations exist indepen
C, and D.
Secondly, what
dently among the properties

B)

B>

may be concluded

and the properties A,

The

manner respecting the property B,

and D. Laws of Thought, p. 146.

in like

C,

equations are

(i_*)(i_,) = (i-0)(i-^(iW)-H/(i-)}
ad(i-e)=ad(i-e){bc+(\-V) (i-*)}

ab+ae(i-b)=d(L-c)+c(L-(I)

(i)
(2)
(3)

The

function of (3) must be squared before being put

Then e can be
into the single equation.
(Art. 292.)
eliminated by means of Article 286. By expanding the func

tion for

in

terms of

b, c,

d>

we

get

a=,(l-^ + *(x-4 + (x-J)(l-4(l-rf).

By taking the function for b in terms of a,
panding it, we obtain

acd=o

and

(3),

a) c(i

(i

(i-a)(i-c)d=o.
By means

d)

c,

o,

d and

(j)

ex

(4)
(5)

of (3), (4), and (5) equation (2) can be reduced

to

47.

certain philosopher has observed that his library

the following conditions

(i.) Languages, which contain Dictionaries and Poetry,
contain Mathematics.

satisfies

(2.)

Languages, which contain Novels and Mathematics,

contain Natural Philosophy.

Examples.

147

Languages, which contain Natural Philosophy, con

Mathematics or Dictionaries, but not both.

(3.)

tain either

(4.) Languages, which contain no Novels, contain either

Poetry or no Natural Philosophy.
which contain Novels, Dictionaries, and
(5.) Languages,
no Mathematics, contain either Poetry and no Natural

Philosophy, or Natural Philosophy and no Poetry.

which contain Poetry and
(6.) Languages,

Natural

Philosophy, contain Dictionaries and Mathematics, or no

Dictionaries.

Are these data all independent of one another ? Is there

any special relation between two only of the subjects?

What

are the relations

Natural Philosophy
48.

Let

The
2.

the subjects independently of

To find (i # y) in terms of x -\-y, xy, and

(i-x-y)*=a+b (x+y)+c(x-y)+dxy.
z

Then

is

among

i=a+2b

xy.

+d.

value of the determinant of the right-hand members

the equations are solved by the method of

When

determinants,

22 04

we

a=

get

--

c=

a=

Hence by Articles 78 and 79

= I,
d2.
0=1,
=
x
i
and
(i
yy
(x+y) + 2xy.
<T=0,

I introduce this

of the
49.

example to illustrate the truth and power

method by indeterminate co-efficients.
There are three bags, one containing 2 white and 3

red balls, the second containing 5 white and 2 red balls, the
third containing 4 white and 7 red balls ; and all the bags
are equally likely to be drawn from.
white ball has

been drawn
the

first

bag?

what

is

the chance that

Gross, Algebra,

p.

225.

it

was drawn from

148

Examples.

Let
a

U = succession of drawings of a ball,

= from the first bag b from the second
c = from the third bag.
;

a w =from the
#y

bag and white,

and red.

first

The

bag

data are

=J

a=\

c=\
r= J 1

=! 1

*.=H
*=*

*r=S?
*r=i

-ft

Now

+^+^=i)
.-.

w=*a m

Probability required

+Z>

+ cw

(Art. 71.

=~
w

l+f+ft
Suppose that in the drawing of balls from an urn
attention had only been paid to those cases in which the
balls drawn were either of a particular colour, white, or of
a particular composition, marble, or were marked by both
these characters, no record having been kept of those cases
in which a ball that was neither white nor of marble had
been drawn. Let it then have been found, that whenever
the supposed condition was satisfied, there 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 that in the next
drawing, without reference at all to the condition above men
tioned, a white ball will be drawn ; also the probability that a
marble ball will be drawn. Boole, Laws of Thought, p. 262.
= long succession of drawings of a ball from the urn
50.

U
x

producing a white ball,

.=
producing a marble ball.
y

Examples.
Let

149

x-\-(ix)y=a.

Then

a x =dp,

required to find

It is

Now

jf

and ay =.aq.

x and y.

={x+(ix)y}x

=x.
ay

Similarly

Hence

=y>

/=_

#+(i

and

JL_

^r -*

From

(i)

From

(i)

hence from

and

+ (!

(2)

py=qx.

(2)

(Art 64

(l)
#)_)>

(3)

x=j{x+(i-x}y},

(3)

Boole proceeds further on the assumption that ocyxy.

That is only the most probable assumption among an
infinite number, and the value for x deduced
by its means
is only the most
probable value of the probability of a
given drawing producing a white

Let the assumption

ball.

be introduced,
then

And

similarly

The

51.

probability that a witness

/i, the probability that another witness

is

fa,

is

speaks the truth is

the truth

B speaks

and the probability

3.

What

statement

Let

is

is

that they disagree in a statement

the probability that if they agree, their

true ?

Boole,

Laws

U = Testimonies by A

of Thought,

and by

= in which A told the truth,

y = in which B told the truth.

p. 279.

B about

a fact

50

Examples.

Then

the data are

and

x(i-y)+y(i-x)=j

t.

(3)

required to find the value of the dependent* xy in

It is

==

Required value

Now

from

xy
(Art. 64.)

xy+(i-x)(i-y)
2xy=x+jf p st
+A At^+(i

(3)

=>i

Also from

Hence

(3)

from

(i )

#)(i>>)=i

required value

""^

=|^-JT^
I-/3

COR. The value of the dependent

and

2)

/,.

(ix)(iy)

in

To deduce

these conclusions Boole employs a long and

He does not see that xy can be deduced

elaborate process.

by very simple algebraic operations from the given data.

At the end of his investigation he remarks that the number of
the data exceeds that of the simple events which they involve.
considerations of Article 151 show that the statement

The
is

not correct

that the data are just sufficient to determine

the single functions of x and y necessarily.

52.
says that a certain event took place ;
says that
required the probability that the event did take place,
all

/! and

speaking

being

the

As

truth.

and

It s

respective probabilities of
the Educational

Mathematics from

TimeS) vol. xxvii.

must suppose that a succession of statements of A

s statements about an event taking place have been
It is required to deduce the probability that
observed.
one of these statements refers to an event which actually

We

about j5

took place.

Examples.

U= statements of A about B

Let

=
=
which
y
The

made by

truly reported the event.

data are

*=A
Now

statements about an

event taking place.

which truly reported a statement

and y

is

#,=AA-

(0

(2)

required.

xyxy

Hence
and
There has been much discussion about the true answer
to the above question.
No fewer than four different solu
tions are given in the Reprint, viz.

AA H AX

1
Todhunter s Algebra
Artemas Martin
A(A/a+(i
American Mathematicians and Woolhouse

Cayley

AA+(i

The meaning

of

^8

and K

by the following paragraph

in

solution

A)A)}-

A A-

-A) + *(i -)(i

Cayley

"

"

is

~A>

explained

B told A

this

that the event happened, or he did not tell

s statement that
told him
the only evidence is

A-

that the event

i

"AX*

AX

B ut
>

happened; and the chances are

in the latter case, either

B told A

an d

that the

event did not happen, or he did not tell him at all ; the
chances (on the supposition of the incorrectness of
s

and the chances of the three

an ^ ( I ~^)( I A)- On the

statement) are j3 and i

cases are thus
/2( J

A)

suppositions of the

and second cases

A>

first

/?;

respectively, the

52

Examples.

chances for the event having happened are

fa and i
on the supposition of the third case
here there
(viz.,

fa

no

is

information as to the event) the chance is

K, the antecedent
probability ; and the whole chance in favour of the event is

AA+/?(i-A)(i -fa) +
Todhunter assumes that
house assumes that

is

in terms of the

expresses

The

it

is

(i-/3)(i

equal to

equal

to

o.

unknown

-A).
and Wool-

fa

Cayley

quantities

solution

and

/3

K.

next example contains the solution of the

general

problem, when there are n persons.

53.

says
of

A A A
2,

<u

AJ

that
says that A* says that
s
M
says
a certain event took place.
The probabilities
s
n respectively speaking the truth are

A,,

that

Required the probability that the event

Pn-

A>

took place.
Let
succession of statements of

U=

about

saying

etc.

which reported truly a statement of

XH\ =
AH
xn = which reported truly the event.
Let x
denote that x 2 is formally dependent on x
that is, x lx
-

x<i

Then

Xl

x2

x,

xn .

*!#,#,.;. ^-i

and

NOW X

-^

X*

X*

(i

Pi fa

COR.
truly.

i.

A-iA +

Suppose that each of the persons always reports

Examples.
etc.,

/2>

each equal to

x o=
XH = +
o

and therefore
COR.

153

Suppose that each person excepting

2.

An

always

reports truly.

Then
COR.

Xn =p,,

An

Suppose that

3.

and

*=-^(i-AA

COR.

An

always reports

pn =

Then

always reports

A-i).

falsely.

Then

XH
5.

Suppose that any other of the persons excepting

4.

COR.

falsely.

=.
o

Suppose that each reports

falsely

as

often

as

truly.

Then each p

Hence

When

equal to

is

xn
n

the range between the limits

complete uncertainty.
54. A goes to hall / times in a consecutive
days and sees
there q times.
What is the most probable number of
times that
was in hall in the a days ? Whitworth, Choice
is i

that

is

is,

infinitely great,

there

is

and Chance,
Let

U=
x

y
The
and

on which
on which

went to
went to

hall,
hall.

data are

it is

Now

=
=

p. 239.
the consecutive days

required to find Uy.

y=xy-\-(ix\
o

154

We
which

Examples.
cannot proceed further without assuming a condition
is not given.

Assume that B is as likely to go

then
go, as when A does go

when

to hall

does not

P
55.

The

probabilities of

The

d 2 respectively.

two causes A^ and

probability that

if

are a 1

the cause

and

pre
(whether as a

will accompany it
sent itself, an event
consequence of the cause AI or not) is p it and the proba

will
cause A z present itself, that event
whether as a consequence of it or not, is / 2
cannot appear in the absence of
Moreover, the event
both the causes A l and A 2
Required the probability of

bility that if the

accompany

it,

the event E.

The

Laws

Boole,

causes

a definite universe.

of Thought^

p. 321.

and the event E are characters of

z respec
Let them be denoted by x

and

}>,

tively.

The

data are

x=d,
xz=d fi
l

y=d

(i)

and

(2)

yz=d #

(3)

(4)

(i-x}(iy)z=Q.

(5)
Article 282 shows that the probability required cannot be
determined exactly from the data. The limits, between

which it must lie, are given in Article 274.

Boole gives the relations which exist among the data.
These relations can be determined very readily by means
of the

method

of Article 226.
i

x=i

dl

that

dl

is,

Thus
and

xz=dipi,

Similarly

Again

#(i

xz=d
z=d

z)=x
and

155

Examples.

that

Similarly
And there

is,

d^

<

#2

+ # 2/2-

which the process can be

applied to the given data ; hence these are all the relations
which exist among them. The solution of this problem
is

no other way

in

was discussed by Cayley, Boole, and Wilbraham in the

Philosophical Magazine^ 4th series, vol. vi. p. 259 ; vol. vii.
Wilbraham
p. 29 and p. 465; vol. viii. p. 87 and p. 431.
gives as the solution

where

is

necessarily less than either a l ^ l or a 2 # 2

He

maintains that we can get no further in the solution without

further assumptions or data, and remarks that the disadvan
tage of Boole s method in such cases is, that it does not
show us whether the problem is really determinate.

The method by

indeterminate co-efficients gives a solu

which agrees with Wilbraham

s (Art. 282), and in every

case supplies the desiderated information as to whether the
proposed problem is determinate.

tion

What is given by Boole s solution is not the mathematical

but the most probable value of
,
probability of the event
the probability which can be deduced from the given data.

EDINBURGH T. AND A. CONSTABLE,

PRINTERS TO THE QUEEN, AND TO THE UNIVERSITY.
:

UNIVERSITY OF TORONTO
O

LIBRARY

<*-<

rH
to
CD

,0
<D

bi

H
cd

DO NOT
REMOVE

Q)

THE

Q)

CARD
FROM

ra

0)

n.,

C!

cb

fl

THIS

POCKET
5

o o

a,
LO

CO

rH

aramn

1
IH3H