OF THE
UNIVERSITY OF CALIFORNIA
18681918
A SURVEY OF
SYMBOLIC LOGIC
BY
C.
I.
LEWIS
BERKELEY
1918
TABLE OF CONTENTS
PREFACE
v
I.
CHAPTER
SECTION
The Scope
and
of
Development
SECTION
SECTION SECTION
II.
Leibniz
III.
18
IV.
De Morgan
Boole
37
51
SECTION
SECTION
V.
VI.
Jevons
Peirce
72
79
107
SECTION VII.
SECTION VIII.
CHAPTER
SECTION SECTION SECTION
SECTION
II.
118
The Postulates
118
and
II.
their Interpretation
Elementary Theorems
General Properties of Functions
122
III.
132
of Equations.
. .
SECTION IV.
V.
of the
Theory
144
SECTION VI.
of the Theory of Inequations. 166 Note on the Inverse Operations, "Subtraction" 173 and "Division"
CHAPTER
III.
ALGEBRA
SECTION
SECTION SECTION
I.
175
175
Diagrams
II.
III.
SECTION IV.
to Classes
184
to Propositions
213 219
to Relations
CHAPTER
SECTION
IV.
222
222
I.
IV
Table of Contents
SECTION
SECTION
II.
III.
The Calculus of Prepositional Functions. Func 232 tions of One Variable Functions of Two or More Variables. 246 Prepositional
Derivation of the Logic of Classes from the Calcu
lus of Propositional
SECTION IV.
Functions
260 269
SECTION
V.
SECTION VI.
of Relations of Principia
Mathematica
279
291
CHAPTER
SECTION
SECTION
V.
I.
Im
292
299
II.
SECTION
III.
The Transformation
of Consistencies
{/)
306
SECTION IV.
The Calculus
of
Ordinary
316
Inference
SECTION
V.
The Meaning
of
"Implies"
324
CHAPTER
SECTION
VI.
340
I.
The
340
View
SECTION
II.
Two
Varieties of Logistic
Method
Peano
Formu343
laire
The Nature
of
of Logistic Proof
SECTION
III.
"Heterodox"
View
of
of the
Nature
Mathe
354
362
The
Logistic
Method
Summarv and
Conclusion
367
APPENDIX.
LEIBNIZ
BIBLIOGRAPHY
INDEX.
PREFACE
The student who has completed some elementary study
of symbolic
logic and wishes to pursue the subject further finds himself in a discouraging He has, perhaps, mastered the contents of Venn s Symbolic situation.
The Algebra of Logic, or the chapters concerning this subject in Whitehead s Universal Algebra. If he read German with sufficient ease, he may have made some excursions into
Logic or Couturat
s
admirable
little
book,
Schroder
These
all
concern the
is
and
his
knowledge
of symbolic logic
probably confined to that system. His further interest leads him almost inevitably to Peano s Formulaire de Mathematiques, Principia Mathematica
of of the
Whitehead and Russell, and the increasingly numerous shorter studies same sort. And with only elementary knowledge of a single kind of
difficult
development of a small branch of the subject, he must attack these most and technical of treatises, in a new notation, developed by methods
which are entirely novel to him, and bristling with logicometaphysical If he is bewildered and searches for some means of further difficulties.
preparation, he finds nothing to bridge the gap.
Schroder
work would
be of most assistance here, but this was written some twentyfive years
ago; the most valuable studies are of later date, and radically
new methods
What
ject
is
one which
system
not only the content of other branches and the alternative methods of procedure, but also the relation of these to
will indicate
The
present book
is
an
common
in addition to this,
some
much
the better.
But
in
this
A
is
gossipy recital
methods,
of
no more use
What
is
and grasp.
vi
Preface
historical
The
of
summary
in
Chapter
development, and no reference, or only passing mention, is given to those studies which seem not to have affected materially the methods of
later researches.
governed by the same purpose. Those topics comprehension of which seems most essential, have been treated at some length, while matters less
omitted altogether.
My
own contribution to symbolic logic, presented in Chapter V, has not earned the right to inclusion here; in this, I plead guilty to partiality. The discussion of controversial topics has been avoided whenever possible
and, for the
rest, limited to
Consequently,
the reader must not suppose that any sufficient consideration of these
questions
accurate.
is
made
will be, I
hope,
"Symbolic
Logic, Logistic,
it
is
facts.
I
Rather,
That would require a volume at least the have tried to set forth the most important and
arbitrarily
critical considerations
is
somewhat
and
and dogmatically,
since there
to provide such a
map
Proof
is
and
of
it is
my
dents
knowledge
mathematics
seldom
command
In any case, this explicitness can do no harm, since no one need read a proof which he already understands.
I
am
indebted to
many
friends
and colleagues
W.
A. Merrill for
emenda
my
H. McDonald and
Dr. B. A. Bernstein for important suggestions and the correction of certain errors in Chapter II, to Mr. J. C. Rowell, University Librarian, for assistance
in securing a
number
of rare volumes,
and to the
officers of the
University
manuscript, eliminated
Mr. Shirley Quimby has read the whole book in many mistakes, and verified most of the proofs.
indebted to
But most
of all, I
am
my
friend
who
first
aroused
my
me encouragement and
due to him.
BERKELEY, July
10,
wise counsel.
Much
best in this
C&gt;
book
is
L LEWIS.
1917.
CHAPTER
I.
LOGISTIC.
we
"symbolic
logic",
logistic",
of
logic",
"calculus
of
logic",
"mathematical
logic",
"algorithmic
And none
because
of these
is
satisfactory.
We
in
it is
the most
commonly used
is
England and
understood.
and
is
because
its
signification
pretty
well
Its
inaccuracy
obvious: logic of whatever sort uses symbols. We are concerned only with that logic which uses symbols in certain specific ways those ways which are exhibited generally in mathematical procedures. In particular,
logic to
relations,
be called
symbols for the logical and must so connect various relations that they admit of "trans
"symbolic"
of
formations"
and
"operations",
of exact statement.
If
we must
is
give
some
definition,
we
shall
Logic
cedure, in ideographic
of these principles
to
g.
to dealing with
is
number and
quantity
and generality
marks
form.
of symbolic logic.
of logic in
any
To
As we
which belongs to it. The modes of Aristotle shall have occasion to point out, the
reasons for the syllogistic form are psychological, not logical: the syllogism,
made up of the smallest number of propositions (three), each with the small est number of terms (two), by which any generality of reasoning can be attained, represents the limitations of human attention, not logical necessity. To regard the syllogism as indispensable, or as reasoning par excellence, is
2
1
And
matter between
it
mark a difference of subject But any such difference is logic. and the really distinguishing mark of symbolic
may seem
There are
between symbolic and other logic. But more important than the making of any such sharp distinction is the comprehension of that ideal of form upon which it is supposed to
and
depend.
so far
we
in
The important
which
characteristics of
this
form are:
ordinary language;
to
may
here be taken
is
mean simply
derived
"exact";
and
Ideograms have two important advantages over phonograms. In the first place, they are more compact, + than 3 than "three", etc.
"plus",
This
is
it
makes
minimum
of effort.
None but
or one without experience of the sciences, can fail to understand the enor
mous advantage
notation
is
of such brevity.
superior to
Many
quite simply expressible in mathematical symbols can only with the greatest
be rendered in ordinary language. Without ideograms, even arithmetic would be difficult, and higher branches impossible.
difficulty
of facts
is
sum
marized in a few principles from which they can be derived, is much more than the mere application of deductive logic to the subject matter in It both requires and facilitates such an analysis of the whole question.
body
of facts as will
most
presentation of the subject matter of logic in this mathematical form constitutes what we mean by symbolic logic. Hence the essential
characteristics of our subject are the following:
(1) Its
The
subject matter
is
that
is,
And,
ideally,
is
excluded.
"
(3)
Amongst the
system)
terms"
of the
Although
"
it
is
nonessential, in
"individuals",
any system
they
prepositional
functions",
or
ambiguously some two or more of these. (4) Any system of symbolic logic will be developed deductively that is, the whole body of its theorems will be derived from a relatively few principles, stated
in symbols,
We
by operations which are, or at least can be, precisely formulated. have been at some pains to make as clear as possible the nature of
logic,
symbolic
because
its
distinction
from
"ordinary"
logic,
on the one
hand, and, on the other, from any mathematical discipline in a sufficiently abstract form, is none too definite. It will be further valuable to comment
briefly
which have
been mentioned.
would not have served our purpose, because "logistic" is used to denote symbolic logic together with the application of commonly its methods to other symbolic procedures. Logistic may be defined as
"Logistic"
the science
which deals with types of order as such. It is not so much a subject as a method. Although most logistic is either founded upon or makes large use of the principles of symbolic logic, still a science of order
symbolic
logic.
We
is
we may postpone any further discussion of it only to make clear the meaning
pages which follow.
It
which
"logistic"
to
comprehends
of such
methods
Its subject
matter
not confined to
logic.
of
logic"
is
name
for our
and
because symbolic logic includes systems which are not true algebras at all. "The algebra of logic" usually means that system the foundations of
which were
laid
by Leibniz, and
after
him independently by
Boole,
and
We
system as the
a
BooleSchroder Algebra
"Calculus"
".
is
"algebra".
By
"calculus"
will
be meant, not the whole subject, but any single system of assumptions
their consequences.
for symbolic logic
and
and
sketched by Leibniz, though the ideal of logistic seems 1 Leibniz left frag to have been present as far back as Plato s Republic.
was
first
mentary developments of symbolic logic, and some attempts at logistic which are prophetic but otherwise without value. After Leibniz, the two
interests
somewhat
diverge.
made by
This type of
contribution
own
was
slight
and not
some extent at
place in
Boole.
England about 1845 and produced the work of De Morgan and Boole seems to have been ignorant of the work of his continental
predecessors, which
proved so
subject,
much
probably fortunate, since his own beginning has more fruitful. Boole is, in fact, the second founder of the
is
and
all later
to his.
The main
line of this
develop
S. Peirce,
of Boole s algebra
and the logical completion of that mode of procedure. In the meantime, interest in logistic persisted on the continent and
fostered
by the growing tendency to abstractness and rigor in mathe matics and by the hope for more general methods. Hamilton s quaternions
was
and the Ausdehnungslehre of Grassmann, which was recognized as a con tinuation of the work begun by Leibniz, contributed to this end, as did also
the precise logical analyses of the nature of
Also, the elimination
from
"modern
geometry"
methods
.of
proof
dependent upon
"intuitions
of
space"
or
"construction"
brought that
1893;
development
of arithmetic
by the
method.
See the criticisms of contemporary mathematics and the program for the dialectic and Philebus, Step. 5657.
all
and
more brought
The Formulaire
is
a monumental
But
its
and not
It
all
made
in
clear.
method
mathematics.
The
publication of this
development of the and in the foundations algebra of relatives from De Morgan to Schroder, for number theory of Cantor and Dedekind and Frege, are all brought 2 Further researches will most likely be based upon the together here.
formulations of Principia Mathematica.
We
history
trace in
more
detail the
development
if
of
symbolic
is
logic.
by
are
own
sake.
Rather, we
by which those
results
which most
today have come to be. Many researches of intrinsic value, but lying outside the main line of that development, will of necessity be neglected. Reference to these, so far as we are acquainted with them, will
be found in the bibliography. 4
II.
LEIBNIZ
logistic properly begins
The
In the
say:
6
and
with Leibniz. 5
New
"I
Essays on the Human Understanding, Philalethes is made to begin to form for myself a wholly different idea of logic from
I
regarded
it
as a scholar
it,
diversion, but I
now
2
it is like
a universal mathe
Hilbert, Fieri,
Perhaps we should add "and the modern development of abstract geometry, as by and others", but the volume of Principia which is to treat of geometry has
not yet appeared. 3 The remainder of this chapter is not essential to an understanding of the rest of the book. But after Chapter i, historical notes and references are generally omitted. 4 Pp. 389406. 5 Leibniz regards Raymond Lully, Athanasius Kircher, John Wilkins, and George Dalgarno (see Bibliography) as his predecessors in this field. But their writings contain
little
6
A
As
matics."
would
set itself
and
clear understanding
little
to
man.
possible.
Leibniz
expected that the whole of science would shortly be reformed by the appli cation of this method. This was a task clearly beyond the powers of any
offer
life,
only the
initial
And
and
so,
throughout his
societies
titled patrons, to
Hence
of
his
many
them
of five
results
Leibniz
during his
life,
all
entertained:
language"
(1) a universal
"universal
medium
("universal
or
of
"rational
or
characteristic")
science;
and
"universal calculus")
designed to display
the most universal relations of scientific concepts and to afford some sys
tematic abridgment of the labor of rational investigation in all fields, much as mathematical formulae abridge the labor of dealing with quantity and
number.
that
is
"The
true
thread,
and palpable medium, which will guide the mind as do the lines drawn in geometry and the formulae for 9 operations which are laid down for the learner in arithmetic."
to say, with a certain sensible
This universal medium is to be an ideographic language, each single character of which will represent a simple concept. It will differ from
existing ideographic languages, such as Chinese, through using a
7
combina
editor s introduction to "Scientia Generalis. in Gephardt s Characteristic Philosophischen Schriften von Leibniz (Berlin, 1890), vn, gives an excellent account of Leibniz s correspondence upon this topic, together with other material of historic interest. (Work hereafter cited as G. Phil.}
a"
The
tit.
especially iv
will
which
and vn. But Couturat, La logique de Leibniz prove more profitable to the general reader than any
vii, 21.
some
compound
idea, instead of
So that while Chinese can hardly be learned in a lifetime, the universal 10 The fundamental char characteristic may be mastered in a few weeks.
acters of the universal language will be few in
number, and
will represent
the
"alphabet
of
human
thought":
The
fruit of
many
n
simple."
With
this
this alphabet of
human thought
the whole
symbolism.
we
recognize here
is
the reconstruction of
all
science
a project too
desirability
ambitious,
still
ideal possibility
and the
And
the ideographic
language finds
matica, and
realization in
Peano
method.
Leibniz stresses the importance of such a language for the more rapid
of science
. .
and
of
human thought
in general.
The
will
different nations.
but the
.
. .
Lacking such guides, the mind can make no long journey without losing with such a medium, we could reason in metaphysics its way
.
and
in ethics
very
much
fix
as
we do
characters would
our ideas,
12
geometry and in analytics, because the which are otherwise too vague and fleeting
in
it
in such matters in
be by
the aid of
characters."
"The
The
medium
prevents
cooperation.
human
comparable to a troop which marches without a leader, without order, without any
me
word or other
one another.
of the road,
we run
hither
The
10
11
"
"alphabet
of
human
of
thought" is
more
visionary.
The
possibility
Duke
Hanover, 1679
13
A
is,
cepts
indeed, real
vide the
But we should today recognize a certain arbitrariness in the selection of The these, though an arbitrariness limited by the nature of the subject.
secret of Leibniz s faith that these primitive concepts are fixed in the nature
of things will
believes that
all
be found in his conception of knowledge and of proof. He predicates are contained in the (intension of the) subject
and may be discovered by analysis. Similarly, all truths which are not absolutely primitive and selfevident admit of reduction by analysis into
such absolutely
thing
"real"
first
truths.
And
finally,
as opposed to
"nominal"
possible;
that
is,
the result
in
any concept
is
unambiguously predetermined
the concept
itself.
The
and he
or
"invention",
"art
of
invention".
result of
analysis
synthesis,
character.
always determined, and only one analysis is finally correct, like inverse processes generally, has no such predetermined In spite of the frequent mention of the subject, the only im
of a calculus of reasoning.
portant suggestions for this art have to do with the provision of a suitable
medium and
essay,
To be
and
in the early
De
and combinations or
later regarded this
of
Specimina he says that the utmost which we can hope to accomplish at present, toward the general art of invention, is a perfectly orderly and finished reconstruction of existing science in terms of the
Scientice novce Generalis,
study as of
little
And
in Initia et
After
two hundred
years,
we
are
fields
without any general method by which logistic may be used in as yet unexplored, and we have no confidence in any absolute primi
still
The
for the
is
to be the instrument
and
it is
all
science will be
symbols
The
calculus will consist of the general principles of operating with such ideo14
15
our reasoning
stituting of characters,
images."
is nothing but the relating and sub whether these characters be words or marks or
characteristica universalis
is
logistic
treatment of science
The plan
development of Leibniz s thought, but he speaks of it always as a mathe matical procedure, and always as more general than existing mathematical methods. 17 The earliest form suggested for it is one in which the simple
concepts are to be represented by numbers, and the operations are to be
merely those of arithmetical multiplication, division, and factoring. When, later, he abandons this plan of procedure, he speaks of a general calculus which will be concerned with what we should nowadays describe as "types
with combinations which are absolute or relative, symmetrical or unsymmetrical, and so on. 1 * His latest studies toward such a calculus
of
order"
form the
earliest presentation of
what we now
call
the
"algebra
of
logic".
But
it
is
doubtful
if
that
it
was intended
to
be the science of mathematical and deductive form in general (it is doubtful whether induction was included), and such as to make possible the appli
cation of the analytic
scientific
method
of
mathematics to
all
subjects of
which
essay,
in the fragments which attempt to an algebra of logic. 20 develop Leibniz wrote De Arte Combinatoria when he was, in his own words,
of
mathematics.
consent.
It
was published, he
us,
The
it
by
its title, is
it.
to serve the
As has been
mentioned,
is
to be accomplished
by a complete
is
concerned with the calculation of the possible forms of this and that type
16
17
18 19
See New Essays on the Human Understanding, Bk. iv, Chap, xvn, 913. See G. Phil, vn, 31, 198 jf., and 204. G. Phil, iv, 35104. Also Gerhardt, Leibnizens mathematische Schriften (1859),
Scientia Generalis.
Characteristica,
V,
179.
20
xvxx, G.
Phil.,
vn.
10
of logical construct:
complexes which
of the
moods and
figures
com
the key to the predicates plexity of the subject as a concept being itself which can be analyzed out of it); of the number of propositions from a
given
number
of subjects, given
21
\
number
of predicate relations,
and given
number
of
fact so
much
space
is
of the essay,
and
it is
most
fre
22
quently referred to simply as a contribution to combinatorial analysis. Beyond this the significance of the essay lies in the attempt to devise a
symbolism which
primitive constituents.
The
representation of concepts
itself is of interest.
by numbers
and suggests
it
regarded as elementary and not further analyzable, each of which is given a number. Thereafter, the number is the symbol of that concept. Class 2
consists of concepts analyzable into (definable in terms of) those of Class 1.
By
belongs and
number
of the class
is
7 in Class 2
and Class
is
and
so on.
By
this
any concept
supposed to be indicated by
numerical symbol. 24
21 Leibniz tells us that he takes this problem from the Ars Magna of Raymond Lully. See G. Phil., v, 62. 22 See letter to Tschirnhaus, 1678, Gerhardt, Math., iv, 45163. Cf. Cantor, Geschichte d. Math., m, 39 ff. 23 See the Synopsis, G. Phil, iv, 3031.
24
tit.,
appended Note
vi, p.
554 ff.
The concepts
"Classis I; 1. Punctum, 2. Spatium, 3. intervallum, 4. adsitum seu contiguum, 5. dissitum seu distans, 6. Terminus seu quae distant, 7. Insitum, 8. inclusum (v.g. centrum est insitum circulo, inclusum peripheriae), 9. Pars, 10. Totum, 11. idem, 12. diversum, 13. unum, 14. Numerus, etc. etc. [There are twentyseven numbered concepts in this class.] "Classis II; 1. Quantitas est 14 T&V 9 (15). [Numbers enclosed in parentheses have their usual arithmetical significance, except that (15) signifies an indefinite number .] 2. Indudens est 6.10. III. 1. Intervallum est 2.3.10. 2. Aequale A rijs ll.. 3. Continuum est A ad B, si TOV A j 9 est 4 et 7 B.; etc.
TU&gt;
etc."
11
In point of fact, the analysis (apart from any merely geometrical defects) Leibniz uses not only the inflected Greek falls far short of being complete.
article to indicate various relations of
indicated
by
et, si,
quod,
quam faciunt,
its
In later years Leibniz never mentions this work without apologizing for
it,
main intention
is
sound.
This method
in fact, the
method
of logistic in general.
Modern
logistic differs
from
this
logistic
two respects: (1) modern attempt of Leibniz most would insist that the relations whereby two or more concepts are
notably in
while Leibniz regards his set of primitive concepts as the
analyzed;
(2)
necessary result of any proper analysis, modern logistic would look upon
them
as arbitrarily chosen.
Leibniz
s later
tion of this first difference, but the second represents a conviction from
At a much
later date
come various
w hich
r
attempt a more systematic use of number and of mathematical operations in logic. 25 Simple and primitive concepts, Leibniz now proposes, should be
symbolized by prime numbers, and the combination of two concepts (the qualification of one term by another) is to be represented by their product.
Thus
if
3 represent
"rational"
and 7
"animal",
"man"
will be 21.
No
a
prime number
animal.
cation:
"man",
will enter
simply a rational
Thus
logical synthesis
is
logical analysis
by
The
analysis of
"rational",
21,
prime factors,
s
3,
and
"animal",
conviction that
all
knowledge
is
analytic
"All
and
all
P"
the proposition
the concept
25
is
be true
if
the
number which
represents
is
divisible
Accordingly the
Dated
April, 1679.
"Elementa
as follows:
Couturat (op. cit., p. 326, footnote) gives the titles of these Characteristicae Universalis (Collected manuscripts of Leibniz in
the Hanover Library, PHIL., v, 8 b); Calculi universalis Elementa (PHIL., v, 8 c); Calculi universalis investigations (PHIL., v, 8 d); Modus examinandi consequentias per numeros (PHIL., v, 8 e); Regulae ex quibus de bonitate consequentiarum formisque et modis syllogis
morum
many
categoricum judicari potest per numeros (PHIL., v, 8f)." These fragments, with others, are contained in Couturat s Opuscules et fragments inedits de Leibniz.
12
(where y
is
a whole number)
some
species whose
"difference",
Similarly
Sx
Hence the particular affirmative, "Some will represent a species of S. S is P," may be symbolized by Sx = Py, or S/P = y x. Thus the uni versal is a special case of the particular, and the particular will always be
true
when the
universal
is
true.
this
scheme.
In the
first
place,
it
the class as genus. presumes that any part of a class is a species within defensible on the ground This is farfetched, but perhaps theoretically that any part which can be specified by the use of language may be treated
as a logical species.
worse defect
lies in
= Py
will
always be true.
will satisfy
For a given S and P, we can always find x and y which = P, the equation Sx = Py. If no other choice avails, let x
or
some multiple
"manangel"
"Angelman"
man"
"ra
tional
invertebrate",
but
it
is
false that
third difficulty arises because of the existential import of the particular a difficulty which later drew Leibniz s attention. If the particular affirma
tive
is
true,
y,
Sx = Py.
The
But
then, be Sx
Py.
And
4=
Py&gt;
this
unworkable because the inequations would have to be verified values of x and y. Also, as we have noted, the equality Sx = Py
4=
true.
Such
difficulties led
still
introducing negative numbers and finally using a pair of numbers, one But this scheme also breaks positive and one negative, for each concept. down, and the attempt to represent concepts by numbers is thereafter
abandoned.
in
Of more importance to symbolic logic are the later fragments included the plans for an encyclopedia which should collect and arrange all known
Leibniz cherished the
G. Phil., vii, xvixx. Of these, xvi, without title, states rules for inference in terms of inclusion and exclusion; Difficultates quaedam logicae treats of subalternation and conversion and of the symbolic expression for various types of propositions; xvm, Specimen Calculi universalis with its addenda and marginal notes, gives the general prin ciples of procedure for the universal calculus; xix, with the title Non inelegans specimen
13
of one concept
when
it is
is
"Equivalence"
two
classes
to include
or
to be included in
27
is
universally
However, Leibniz evidently considers the calculus to have many applications, and he thinks out the relations and illustrates them frequently in terms of extensional diagrams, in which A, B, etc., are
of the subject
A".
line.
is
included in
is
included in
in extension;
and a calculus
may
be inter
and
classes in extension.
tion
may
consequent proposition.
just as the subject
The hypothesis A
all
includes
demonstration
analysis.
means
Thus As
one
title
demonsirandi in abstractis struck out, and xx, without title, are deductive developments of theorems of symbolic logic, entirely comparable with later treatises.
place of symbolic logic in Leibniz s plans for the Encyclopedia is sufficiently by the various outlines which he has left. In one of these (G. Phil., vu, 49), divisions 16 are of an introductory nature, after which come
The
indicated
scientiarum instauratione, ubi de Systematibus et Repertoriis, et de Encyclo paedia demonstrativa codenda. Elementa veritatis aeternae, et de arte demonstrandi in omnibus disciplinis ut in Mathesi.
"7.
De
"8.
De novo quodam Calculo generali, cujus ope tollantur eos qui in ipsum consenserit; est Cabala sapientum. "10. De Arte Inveniendi.
"9.
omnes disputationes
inter
"11.
"12.
"13.
De Synthesi seu Arte combinatoria. De Analysi. De Combinatoria speciali, seu scientia formarum,
Characterismis) sive de simili et dissimili. "14. De Analysi speciali seu scientia quantitatum in genere seu de "15. De Mathesi generali ex duabus praecedentibus composita."
magno
et parvo.
Then
cine, psychology, political science, economics, military science, jurisprudence, theology, in the order named.
27 28
various branches of mathematics, astronomy, physics, biological science, medi and natural
"Generates
Inquisitiones
etc.,
pp. 35699.
14
It is a
failed to
frequent remark upon Leibniz s contributions to logic that he accomplish this or that, or erred in some respect, because he
The
analogy between the various logical relations. It is a part of his sig nificance for us that he sought such high generalizations and believed in their validity. He preferred the point of view of intension, or connotation,
partly from habit and partly from rationalistic inclination.
As a conse
is
likely to overlook
it,
difficulties
and choice
which he might have avoided by an him to make which has since been overlooked and
his
some
which
commentators have
fallen. 29
He
is
existential import of particular proposi In the course of this essay he formulates the symbolism for the four
and the
propositions in
two
All
different ways.
The
first
formulation
is:
30
Univ.
aff.;
is
B:
is
AB =
not B;
A, or
exist.
Part, neg.;
Some
A
is
AB
4=
A, or
exist.
non5
exists.
Univ. neg.;
Part,
aff.;
No A
Some A
B;
is
AB = A
man"
and
AB
=}=
A may
extension indifferently.
is
If all
"mortal
class of
mortal
men
identical in extent
The statements
concerning existence are obviously to be understood in extension only. The interpretation here put upon the propositions is identically that of
contemporary symbolic logic. With these expressions, Leibniz infers the subaltern and the converse of the subaltern, from a given universal, by
impossible,
For example, it led him to distinguish the merely nonexistent from the absurd, or and the necessarily true from the contingent. See G. Phil, vn, 231, footn6te; and "Specimen certitudinis seu de conditionibus," Dutens, Leibnitii Opera, iv,
Part in, pp. 92
footnote.
30
.
29
,C
ff.,
also Couturat,
La Logique
and
p. 353,
15
means
A,
exists.
is
B:
is
AB =
not B:
A.
Some A
AB
4=
A.
Ens.
No A
is
B:
is
Some A
B:
AB before the sign of equality represents the AB s or the AB the region of possible represents existing AB s, or actual members of the class AB. (Read AB Ens, which AB = AB Ens thus represents the fact that the class AB has members; AB AB Ens, that the class AB has no members. A
In the last two of these,
ideas";
"AB
Ens"
"AB
exists".)
=f=
A,
"some
A",
may
be represented by
"some
YA;
exist
YA Ens
ing

members
of that species, or
if
A".
AB = A
(All
is
B),
YAB
A is B); but if AB 4= A, it does not follow that YAB 4= YA, YA. Again, if AB 4= AB Ens (No A is B), YAB B, YAB 4= YAB Ens (It is false that some A is B); but if AB = AB Ens (Some A is B), YAB = YAB Ens does not follow, because Y could assume values incompatible with A and B. For example, some men are wise, but it does
YA
if
(Some
=
for
=
men
32
"foolish"
is
incompatible with
division of
"wise".
The
AB,
a logical
A
s
or of B,
and
AB
Ens, existing
AB s,
is
ingenious.
This
is
our author
intension,
most
and
are
and
of the
calculus"
may
be
concluded in terms of any other letters which satisfy the same conditions; for example, since it is true that [all] ab is a, it will also be true that [all]
be
is
and that
[all]
bed
is be.
2)
"Transposing
for
.
example ab
of a letter in the
same term
is
useless.
4) proposition can be made from any number by joining all the subjects in one subject and all the predicates in one predicate: Thus, a is b and c is d and e is /, become ace is bdf.
.
31 32
is
mine.
33
16
A
5)
"From
any proposition whose predicate is composed of more than one term, more than one proposition can be made; each derived proposition having the subject the same as the given proposition but in place of the
given predicate some part of the given predicate. u a is b and [all] a is c and [all] a is
d."
If [all]
is
bed,
then
[all]
If
we add
ab
is
to the
of
number
of these,
two
principles
(1)
"selfevident
propositions"
included in a;
and
(2)
included in a
we have
logic.
usually qualified
of
by
some doctrine
of the
"universe
of
discourse"
"range
significance",
it is is
The law
the
numbered 2 above
3, the
"principle
what we now
the
"principle
of permutation";
5,
of
tautology";
4, the
"principle
of composition";
"principle
of
division".
And
the two
"selfevident propositions"
are often
There remain
in
XIX
and
XX
of Scientia Generalis:
Char
The
first
Non
inelegans specimen
is
demoninter
rather the
or ab
is
more
esting.
AB
represented
by A+B. And A+B = L signifies that A is contained or included in A scholium attached to the definition of this inclusion relation (est in} B.
distinguishes
from the partwhole relation. Comparison of this and other passages shows that Leibniz uses the inclusion relation to cover
it
(1)
the relation of a
member
(2)
the relation
(3)
a relation in extension;
a relation of intension.
the rela
first
The
of
these
is
our erelation;
is
(2) is
and
(3)
fragments,
terms of
extensional diagrams, in which classes or concepts are represented by segments of a line, and only incidently in terms of the intension of concepts.
The
tinguished.
is
"rational"
and
is
m intension,
then
A+B
will represent
"rational animal".
But
if
and
extension^ then
A B
+
is
the class
made up
of those
or
(or both).
Thus the
inclusion relation,
4.
and
5.
ties of
are stated without qualification because this study is confined to the proper 4. is true also for universal negatives.
17
A +B =
in
L,
may
is
".
This
is be interpreted either in intension or in extension as little confusing to us, because we should nowadays invert
"A
intension to extension;
instead
B"
A +B
from
If
"both
and
(in
intension) to
mal",
"either
or
B"
(in extension).
A
If
is
"rational",
"ani
and
"man",
then
B = L
is
true in intension,
"man".
"rational animal"
"man"
or
"rational"
is
contained in
A, B, and
L are
classes
common
segments
to
and
and B.
The
relation
If
A +B
is
exclusive.
does not require that A and B should be mutually a line, A and B may be overlapping segments; and, in
intension,
and
B may
"triangle"
and
"equilateral",
detractio.
A =
+
N signifies
is
contains
and that
[ ]
if
be taken from
the
if
remainder
N.
The
B =
L,
it
==
B, because
.
overlapping
(in
terms, communicantia)
If
Hence
A+B
= L and
M.
(If
take a
clear.)
line,
L, in which
and
In
extension,
it
is
A may
and
calls
be interpreted
"L
which
is
not
".
In intension,
"rational"
more
difficult.
"man"
=
is
"brute",
"man"
"rational"
not
"nonrational man"
"nonrational".
35
In intension, the
But there
"
Say that
pp. 37778)
man"
35
r
"carpenter"
w^~
and
"man"
"whiteskinned"
^^
&
Couturat in commenting on
this (op.
cit.,
says:
"Ailleurs
"Mais il oublie que le ne"ant Nihilum. Sed A nonA est Absurdwn. (nonEns) n est pas autre chose que ce qu il appelle 1 absurde ou 1 impossible, c estadire
est
le contradictoire."
Nonexistence may It may be that Couturat, not Leibniz, is confused on this point. be contingent, as opposed to the necessary nonexistence of the absurd. And the result of abstracting A from the concept A seems to leave merely nonEns, not absurdity.

18
A
"Caucasian".
Then
"
Caucasian
"
"
"
carpenter
"man"
"white
skinned"
penter"
"woodworking".
"whiteskinned",
Hence
("Caucasian"
"carpenter")
"car
because the
common
constituent
is,
"man"
has
"carpenter".
That
which
the abstraction of
"Caucasian"
from
"Caucasian
carpenter"
leaves, not
is
but
"Caucasian"
wholly absent in
because
is
"man"
We
"whiteskinned
man"
is
abstracted, nor
"man":
"whiteskinned animal"
because
"animal"
contained
in
"whiteskinned"
as a pure abstraction.
Such
abstraction
to carry out
and
an instrument
of logical analysis.
manuscript, and
his
theorems
in
Leibniz s illustration is scribbled in the margin of the seems clear that at this point he was not thinking out terms of intensions.
it
Fragment by
[
].
This
loss,
[ ]
and because [+ and XX is more carefully developed more of the simple theorems are proved, and more illustrations are given. Otherwise the definitions, relations, and
interpretation
:
symbolized both because of the difficulty of are not true inverses. Also
methods
of
these fragments might well have proved sufficient foundation for a satis
III.
calculus of
Castillon,
and
others, all
made
Of these, the
of Ploucquet,
Lambert and
it
intends to be a calculus
success,
But
is
result
of the others
relations in intension
most
Of Segner
36
we can give no account, since no Venn makes it clear that Ploucin this country,
secure
There seem to be no copies of Ploucquet s books them from the continent have so far failed.
and attempts to
19
quet
calculus
was a calculus
of intension
and that
it
most
made from
different beginnings;
those which precede and refer to their theorems; and yet the development
is
will
be found
set forth
more
at length in another.
And
discussion of
lems of the theory of knowledge and of scientific introduced. But the important results can be presented as a continuous
Lambert
of his symbols:
The symbol
of equality (Gleichgultigkeit)
=
+
addition (Zmeizung)
abstraction (Absonderung)
opposition (des Gegentheils)
universality
particularity
copula
given concepts (Begriffe)
a, b, c, d, etc.
undetermined concepts
n,
m,
I,
etc.
unknowns
the genus
the difference
x, y, z.
7
5
developed entirely from the point of view of intension: the letters represent concepts, not classes, + indicates the union of two
calculus
is
[
The
some part
term
and
represents the
common
"multiplied"
into them.
difference of a.
Much
use
is
qualify any Thus ay represents the genus of a, ad the made of the wellknown law of formal logic
8
7 and
that the concept (of a given species) equals the genus plus the difference.
(1)
37
ay + ad = a(y +
6)
Johann Heinrich Lambert (172877), German physicist, mathematician, and astrono He is remembered chiefly for his development of the equation x n +px = q in an infinite series, and his proof, in 1761, of the irrationality of TT. 38 In Logische und philosophische Abhandlungen; ed. Joh. Bernoulli (Berlin, 1782),
mer.
vol.
i.
20
A
is
ay + a8
As immediate
consequences of (1),
(2)
we have
also
(3)
]
ay = a
it
ad
ad
]
ay
Lambert takes
tions.
If
and
We
+
b.
two concepts,
(a
b)
then
have already noted the difficulties of Leibniz on this point. a and b, have any part of their connotation in common, b will not be a but only that part of a which does not belong
"European"
"+
also to
If
and
"carpenter"
have the
common
is
part
"man",
then
("
European
"carpenter")
"man".
minus
"carpenter"
]
not
"European"
but
"European"
minus
And
[+
and
will
inverses.
But this difficulty may be supposed to disappear where the terms of the sum are the genus and difference of some concept, since genus anddifference may be supposed to be mutually exclusive. We shall return
to this topic later.
of
may
any given a
also a concept
and can be
a
Proof:
= a(y+
2
&lt;5)
= ay
+ ayd + ady + ad 2
ad
ay = ayy + ayd
ad.
and
= ady +
add
But a = ay +
That
is
Hence Q.E.D.
to say:
giving
its
one wish to define or explain a, one need not stop at genus and difference, but may define the genus in terms of its
if
genus and
difference,
similarly.
Thus
is
equiva
lent to the genus of the genus of a plus the difference of the genus of a plus
a.
This
may
be called a
"higher"
definition or
"explanation"
of a.
Obviously, this process of higher and higher "explanation" may be carried to any length; the result is what Lambert calls his "Newtonian
formula".
We
shall best
understand this
if
step. Suppose the explanation carried one degree further and the resulting terms arranged as follows:
6 )
+ 767 + dyd
+ dyy + ddy
The
two y
s,
and one
might be summarized
21
the three arrangements of two 6 s and one 7 by 8752. With this convention, the formula for an explanation carried to any degree, n, is:
by 3y
d;
This
"Newtonian formula" is
Two
(6)
a
Proof:
ad + ayd + ay 2 d + ay*d +
etc.
= ay +
ad
But
and
2 ay = ay + ayd 2 2 ay = ay* + ay d
ay*
(7)
= ay 4 + ay*d,
2
etc. etc.
. .
a
Proof:
= ay + ady + ad y + ad*y + a = ay + ad
ad
etc.
But
and
ad 2
ad*
= ady + ad = ad y + ad*
2 2
=
is
a by ay 2
n
etc., so a species of
which a
is
genus
may
be represented by ay~
2
,
l ,
and a species
ay
genus of the genus by ay~ etc. In general, as a. genus above a, so a species below a may be represented by represents
of
is
which a
ay~
a
or
yn
Similarly afny concept of which a
is
ence
etc.,
may
be represented by
ad~ n
or

Also, just as a
tion",
a(y +
n d)
,
where a
is
d)
its "explana
so
a,
where
^n ls ^ ne concept and a
the
"explanation"
of
it.
"division,"
need to be observed. 40
40
22
A
is
(ay + ad)y
2 7 + ady
= a(y+d)y = ay(y+d) = ay
while
of a.
is
Also
 y must be
7
which a
is
77
.
 y
is
species x of
the genus,
e.,
(9)
7 =
But ay/y
is any species of which the genus of a is the genus, x such that a and x belong to the same genus. species
i.
e.,
any
We
turn
now
is
which have a
and
common
in
part.
Similarity
so far as,
identity of properties.
Two
if,
they comprehend
identical properties.
In respect to the
ab represents the
common
properties of a and
a.
b.
a
a + b
from
to recognize in
b
his
system the law a + a a; else he need only have written a + If x and a are of the same genus, then
ab.
xy = ay
If
and
ax
= ay = xy
is
now we symbolize by a
b that part of a
which
different
from
42
6,
then
(10)
a\b + b\a + ab + ab
=
x
a+ b
Also
x
ax
a
x a
ay,
or
= ay + x\a
ad
ax
ad
ax + ad
a
ad
ax
41 42
= ay = xy
sometimes a
:
Ibid., p. 10.
b for this,
b.
23
And
since
x
ax + a x ax
= ay
ax + x\a
a\x
a ax
x\a
x\a
a\x
=
is
ax
The
y
fact that y
a property comprehended in x
= xy
or
by y + x y
x.
The manner
first is
in
interesting.
If
is
a property
x+y
Hence
in this case,
2xy + x y
\
=
=
x+y x +y
And
since y
xy,
2y + x\y
Hence
y+x\y =
This
is
He
2y
=
y&gt;
follows,
probably be
concerned.
There
is
for elimination or
solution.
Formulae
They
and
of dubious application.
It is difficult to
judge of possible applications because in the whole course is not a single illustration of a
illustrations of
any kind.
The shortcomings
There
is
too
much
and
reliance
upon
Some
uninterpretable, as for
and
1.
but 2y either
has no meaning or requires a conventional treatment which is not given. And in any case, to subtract y from both sides of 2y = x + y and get y = x
represents no valid logical operation.
of the relations
b,
43
employed
is
lacking,
and
may have
common
24
x
A
man =
from
x.
s
Suppose, for
procedure,
rational + animal.
Then, by Lambert
animal.
we
= man
There
is
still
common
"
thing"
ing
(Lambert
Hence the
and a
difficulty will
not really
a8
ay =
ad, a
= ay
use of
of a
in a calculus
In fact this consideration vitiates altogether the based on intension. For the meaning
[ ]
ventional inverse of
becomes wholly doubtful unless be treated as a wholly con + ], and in that case it becomes wholly useless. The method by which Lambert treats the traditional syllogism is only
b
[
its
He
reconstructs the
whole of Aristotelian logic by the quantification of the predicate. 45 The proposition "All A is has two cases:
B"
(1)
A =
it
is
(2)
&gt;
is
comprehended concept A. The particular affirmative similarly has two cases (1) A B, the case in which the converse is a universal, the subject comprehended within the predicate B.
:
&lt;
(2)
The
is
particular.
comprehended within a subsumed species of the predicate and the predicate within a subsumed Lambert says species of the subject.
subject
is
this
may
mA
not
&gt;
and
&lt;
nB
must
&lt;
A
is
&gt;
mA, and
mA
A.
mA
Ibid., p. 12.
45
7m,
25
Hence
mA
in
&gt;
&gt;
We
see that
comprehended
Lambert here
translates
"
Leibniz, that
any subdivision
as
some
species
Lambert says the subject and predicate each have peculiar properties by virtue of whose comprehension neither is contained in the other. But if the peculiar properties of the subject be taken away, then what remains is contained in the predicate; and if the peculiar properties of the predicate be taken away, then what
remains
is
is
repre
m
The
is
and
A&gt;
B n
two
cases:
i.
e.,
when some
A.
This
is
expressed by
.4
&lt;B
(2)
When
it
In this case a
contained in the predicate, and a sub sumed species of the predicate in the subject.
subsumed
mA
Either of the signs,
terms.
&lt;
&gt;
B
&gt;,
and
&lt;
nB
by transposing the
"
and
&gt;
may
be reversed
/,
And
if
&lt;
Q,
P =
IQ.
Also,
multi
plication"
and
"division"
&gt;
is
equivalent to
A = mB nA = B
kB]
]
mA =
IA A_
= nB
or
pA = qB
_B
k
or
m
A =
"
B_
It is
26
lents of the
"
A
inequalities"
the
full
expression of these
equations:
(1)
A = mB:
All
is
and some
is
not A.
(2)
(3)
nA = B: Some A is not B and all B is A. mA = nB: Some, but not all, A is B, and some, but
not
all,
is
A.
(4)
=
No A
is
B.
that
The
first
is
A/m = B/n
is
transformable into
nA = mB and
fact,
can mean nothing different from (3). Lambert has, in four different propositions, if he sticks to the laws of his calculus: only
(1)
(2) (3) (4)
A = B: All A is all B. A = mB: All A is some B. nA == B: Some A is all B. mA = nB: Some A is some
B.
These are the four forms which become, in Hamilton s and De Morgan s A little scrutiny will show that treatises, the four forms of the affirmative.
Lambert
all, it is
treatment of negatives
necessary that
"
is
a failure.
For
it
to be consistent at
should not be transformed. But Lambert constantly makes such transformations, though he carefully re frains from doing so in the case of expressions like A/m = B/n which are supposed to represent universal negatives. His method further requires that m and n should behave like positive coefficients which are always
fractions"
greater than
4= n.
This
is
unfortunate.
It
makes
it
"entangling alliances".
as affirmatives
book and treated negative propositions with negative predicates, he might have anticipated the
s
calculus of
De Morgan.
In symbolizing syllogisms, Lambert always uses A for the major term, for the middle term, and C for the minor. The perfectly general form of
is:
proposition
mA
P
Hence the
nB
(
1
46
:
AT
46
mA
p
nB
q
Major
Ibid., pp. 102103. Ibid., p. 107.
47
27
Minor
nC
vB
Conclusion
fj.n
C =
mv
pp
irq
The indeterminates
letters.
in the
The conclusion
is
gives
B =
np
A.
The minor
gives
B =
irv
C.
Hence
np
AC.
irv
syllogism,
it
Lambert
scheme of
28
nB = A
IV.
Fideleo
C
7T
X.
Pilosos
nB = A V.C = B
nC
= A
XL
Romano
nB = mA C = B nC = mA
nB = mA C = B
7ijjiC
XII.
Somnio
= mA
The
since
difficulty
about
"division"
it is
if
does not particularly affect this scheme, one of the premises involve fractions",
"
It will
mood
Hilario
is
identi
with Somnio.
The reason
for this
nB =
mA
(see
above).
interpretation, as their
the negative. indicate; Into the discussion of the other three figures, the reader will probably not care to go, since the manner of treatment is substantially the same as
in the above.
names
method
There are various other attempts to devise a convenient symbolism and for formal logic; 48 but these are of the same general type, and
they meet with about the same degree and kind of success. Two brief passages in which there is an anticipation of the logic of
relatives possess
attributes",
some
interest. 49
Relations,
Lambert
says, are
"
external
to the object
an
"
sich.
Metaphysical"
e.,
by
Greek
letters.
For example
if
=
/
fire,
h
: :
=
h
heat,
and a = cause,
The symbol
48
Also fragments
i,
"Uber
die Vernunftlehre", in
190 jf.
49
ec/is
29
"relative product".
called a
Fire
h
is
Fire
is
*
&lt;*
Heat
is
to
fire
as effect to cause.
The dot
here represents Wirkuny (it might be, Wirklichkeit, in consonance with the metaphysical interpretation, suggestive of Aristotle, which he It has the properties of 1, as is illustrated elsewhere 50 gives to Ursache).
by the fact that 7 may be replaced by this symbol. Lambert also uses powers of a relation.
If
=
&lt;p
b,
and
=
&lt;p
c,
=
&lt;p
(p
(p
And
if
=
&lt;p~
c,
o
&lt;p
a c
j and
V7
a \
indeed
it
is
difficult to see
how
felt
Yet
it is
that the powers of a relation ought to be logically important, and that he here hit upon exactly the concept by which the riddles of "mathematical
induction"
Holland
bert.
51
attempt at a
it
logical calculus
is
contained in a letter to
Lam
He
himself calls
an
"unripe thought",
and
in a letter
some three
if
logic
is
But
this
study
is
of particular interest
because
it
symbolic logic from the point of view of extension from the time of Leibniz
to the treatise of Solly in 1839.
method
lines,
Ibid., p. 21.
52
Joha n. Lamberts deutscher Gelehrten Brief wechsel, Brief m, pp. 16 ff. See Ibid., Brief xxvn, pp. 259 ff.
30
"
A
"men"
relation of
to
mortals" is
inter.
He
is
apparently not
aware that
this
as
we
(1)
shall see.
If
S represent the
subject,
TT
signify
is
unde
a part a P.
is
termined variable
numbers, S/p =
P/TT will
come
to:
part of S
of P, or certain of the
P s,
all
or (at least) an
is
This expression
is
possible judgments, as
A member
If
is
and
in
IT
both
cases,
is
either
finite or infinite.
(3)
We
can be understood.
p =
less
attains
it
its logical
can
still
all S, and in this way S/p cannot become less than 1, Since, then, p disappear and consequently cannot become negative.
1
in S/p, then
S/p as
many
as
maximum.
TT.
The same
(4)
is
true of
Therefore p and
TT
TT
less
than
If
or
(5)
infinite, the concept becomes negative. If /expresses a finite number 1, then the possible forms of judg
&gt;
becomes
ment
are as follows:
All
(1)
=
^
=
^
j
is all
P.
(2)
All
is
some P.
Now
an
expresses negatively
what
To
say that
straight,
means
exactly:
No
part
of a curved line
straight.
(3)
~ =
1
All
CO
is
not P.
(4)
j
.
Some S
p
is all
P.
o
(5)
 =
Some S
is
some P.
(6)
S  = P
Some S
is not
P.
(7)
53
oo
=^ 1
All notS
is all
P.
See
31
S
(8)
P j
P
All notS
is
some P.
S
(9)
All notS
is all
notP.
(1),
(2),
and
and
(9)
(8),
says are universal affirmative propositions; universal negatives; (4) and (5), particular affirmatives;
Holland
a particular negative.
As Venn has
Boole.
If
numerator
indicated by
where
Boole
is
&lt;
&lt;
1,
and
/co
.=
QS.
But the
and
form
The
fractional
little
may
be cleared
of fractions,
and
s
this
would give
which are
logically uninterpretable.
But Holland
calculus.
notation can be
That he did
made the basis of a completely successful not make it such, is apparently due to the fact that
he did not give the matter sufficient attention to elaborate the extensional
point of view.
He
Example
1.
All All
men
H are mortal M
E are men H
Europeans
P
Ergo,
[All
Example
2.
P =
P
All plants are
no animals
oo
Ergo,
 =
A

32
A
Example
3.
All
men
are rational
//
P
T&gt;
P =
CO
men
vH P=00
= R/p
into
pH =
"Some
R, which
is
not legitimate, as
the rational
It
men
are
all
is
also misinterpreted.
should be,
some
men".
correct reading
Lambert
method, pointing out, correctly, that Holland s calculus would not distinguish the merely nonexistent from the impossible or
contradictory (no calculus in extension can), and objecting to the use of c in this connection. It is characteristic of their correspondence that each
of the intensional
pointed out the logical defects in the logical procedure of the other, and neither profited by the criticism.
Castillon s essay toward a calculus of logic
is
contained in a paper
Academy in
1S03. 54
The
letters S,
A,
etc.,
represent
is
an indeterminate, S +
the withdrawal or
of
S and M, S
from
S.
 M,
subsumed,
of
in
 M.
Consonantly S + M,
to S, represents a
"further specification"
which contains
(in intension)
the concept S.
is
The
contained in the
Thus
"All
is
A"
is
represented by
S =
A
is
M
+
is
The
universal negative
"No
S

is
A"
symbolized by
= ( A)
M
withdrawn
classes,
is
The concept S
lusory".
is
something,
M, from which A
no A.
two
"real"
and
"il
the
"Memoire
un nouvel algorithme
logique",
in
de Berlin, 1803, Classe de philosophie speculative, pp. 114. sur la Logique", loc. cit., 1802.
33
whose converse
also
is
particular.
The
real particu
A = S since this
is
is
M
The
illusory particular affirmative
the converse of S
= A
M.
represented by
S =
A=?
M
A
which
Castillon s explanation of this is that the illusory particular judgment gives us to understand that some S alone is A, or that S is got from A by ab
straction (S
= A  M), when
==
abstraction (S
M + A).
in reality
this
it
it is
is
Thus
put +
M]
judgment puts by S = A ^ M.
M where
drawn from S by
it
should
The
fact
of course, that
"Some
if
indicates nothing about that they are not incompatible. except one or both be further specified in proper
is
"
might well be symbolized by S + N is really suspect that Castillon s choice of S = A =F governed by the consideration that S = A + be supposed to give may
fashion, the results will coincide.
It
= A
M.
^F
We
S =
the universal to give its subaltern, and that A will also give S = A =F M, that is to say, the real particular
M,
= S which
M
is
"All
is
S"
S =
A"
=F
M.
Thus
"All
"Some
is
A"
may
is
be
is
and from
is
S",
which
a de
sideratum.
The
is,
correspondingly,
=F
M
symbolism.
The S =
is
A
is
M gives A
= S
M, and
S+
=
vice versa.
The
universal negative
directly convertible.
S =
ticular
A+M
J/
gives
A =

J/,
and
vice versa.
The
S
illusory
par
also convertible.
S =
A T
gives
=F
M.
Hence
A =
=F
comes back to
S = A
M, which
M.
subaltern
A
S
universal gives
its
S =
=
A

gives
S =
A
=
=F J/,

and
=T=
A+M
gives S
A
M,
M.
And a real A = S 4
S =
34
A
its
subaltern,
S =
A ^ M,
figures of the syllogism
A =
M.
moods and
may
be symbolized
which involve particular propositions being valid For example: particular and for the illusory particular.
Mis A
M
.
SisM
Sis
,4
.
All
No
All
M
S
is is
M
:.
M
A
No Sis A
All
(N + P)
is
M
.
.
Some S is Some S is A
is
= A+N S = M =F P S = (A + N)
or
=F
S =
or
M
P
+ N)
S = (A
 P
This
The
and
not
in
difficulty
about
"subtraction"
in the
Lambert
"
"
or
M which
If,
Leibniz,
"M
but
not
P
is
".
If it
mean
this,
and
are not
true inverses.
M
P indicates the
abstraction from
the concept
difficult or
of all that
P
is
cannot be represented by []. How does it happen, then, that Castillon s notation works out so well when he uses [] both for abstraction and as
the sign of negation?
It
his calculus
ought to involve
him
in
both kinds of
is
difficulties.
method
by good luck, hit upon a which nothing is ever added to or subtracted from a determined concept, S or A, except an indeterminate, or N or P, and this indeter
in
The answer
M
]
it is
and []
And when
the sign
A,
serve as the sign of negation, because no difficulty arises from supposing the whole of what is negated to be absent, or abstracted. Castillon s calculus is theoretically as unsound as Lambert s, or more
it
may
so
if unsoundness admits of It is quite possible that it was worked degree. out empirically and procedures which give invalid results avoided.
35
Whoever studies Leibniz, Lambert and Castillon cannot fail to be con vinced that a consistent calculus of concepts in intension is either immensely difficult or, as Couturat has said, Its main difficulty is not impossible.
the one which troubled Leibniz
s system the failure This can be avoided by treating negative propositions as affirmatives with negative predicates, as Leibniz did. The more serious difficulty is that a
[ ]
Lambert
and which constitutes the main defect in of + and [] to behave like true inverses.
calculus of
"
"concepts
is
and
admirably
in a letter to
Lambert. 55
He
Lambert
method,
All triangles are figures.
All quadrangles are figures.
T =
tF
Q = qF
tQ
Whence,
T =56
or
qT =
the conclusion nA = be drawn, the calculus cannot determine whether the ideas nA and
general,
if
from
==
mC
and
B = nC
mB mB
to the
matter."
calculus,
A"
A, then
this
mode
is
not fallacious.
And
this observation
brings
down
The
more
The
calculus of Leibniz
is
successful than
s is
continental successors
unless Ploucquet
period between
It is
no accident that the English were so quickly was aroused; they habitually think of
of
"intension"
it
is
relations of concepts
which the
Ibid., pp.
Brief xxvu.
36
A
The beginning
of
England is marked by the proposing some modification of the traditional logic by quantifying the predicate. As Sir William Hamilton 57 notes, the period from Locke to 1833 is singularly barren of any real con
thought upon
publication of
numerous
treatises, all
tributions to logic.
About that time, Hamilton himself proposed the As we now know, this idea was as old at
Both Hamilton and
his student
Thomson men
but this new burst of logical study in England impresses one as greatly concerned about its own innovations and sublimely indifferent
to its predecessors.
establish his
Hamilton quarrelled at length with De Morgan to 58 This is the more surprising, since priority in the matter.
New
2. 3.
4.
5.
6.
7.
8.
is

X in toto = Y ex parte X in toto Y ex parte; X in toto = Y in toto; X in toto Y in toto; X ex parte = Y ex parte; X ex parte Y ex parte; X ex parte = Y in toto; X ex parte 7 in toto.
;
!
    
(
"diversity").
in
But Hamilton was certainly the center and inspirer of a new movement logic, the tendency of which was toward more precise analysis of logical
significances.
Thought are
there
Bayne s Essay on the New Analytic and Thomson s Laws of the most considerable permanent record of the results, but
fervid discussion of logical topics in various peri
was a continual
was
odicals; logistic
in the air.
This movement produced nothing directly which belongs to the history of symbolic logic. Hamilton s rather cumbersome notation is not made the
basis of operations, but
is essentially only an abbreviation of language. scheme of representing syllogisms was superior as a calculus. But
Solly
57
See Discussions on Philosophy, pp. 119 ff. This controversy, begun in 1846, was continued for many years (see various articles in the London Athenceum, from 1860 to It was concluded in the pages of the Con 1867).
temporary Review, 1873.
37
in fact
it
movement accomplished two things for symbolic logic: though not always in name the point of view of
it
emphasized
extension,
and
aroused interest in the problem of a newer and more precise logic. These may seem small, but whoever studies the history of logic in this period
will easily
convince himself that without these things, symbolic logic might Without Hamilton, we might not have had
record of symbolic logic on the continent is a record of failure, England, a record of success. The continental students habitually
The
DE MOP CAN
of symbolic logic only
De Morgan 59
permanent value
of certain
tions.
is
known
to
most students
through
of
"universe
the discovery
new types
of propositions,
and a beginning
new
ready
and the
clarity
and
yeoman
of
service in breaking
down
in
His important writings on logic logic. are comprised in the Formal Logic, the Syllabus of a Proposed System of
"mathematical"
methods
Logic,
and a
sophical Society?
59
Augustus De Morgan (180678), A.B. (Cambridge, 1827), Professor of Mathematics London 182831, reappointed 1835; writer of numerous mathematical treatises which are characterized by exceptional accuracy, originality and clearness. Per haps the most valuable of these is "Foundations of Algebra" (Camb. Phil. Trans., vu, vm); the best known, the Budget of Paradoxes. For a list of his papers, see the Royal For many years an active member of the Cambridge Philosophical Society Catalogue. Society and the Royal Astronomical Society. Father of William F. De Morgan, the novelist and poet. For a brief biography, see Monthly Notices of the Royal Astronomical Society,
in the University of
xii, 112.
60
The
idea
is
"Let
as
man and
notman.
plain that between them they represent everything, imaginable But the contraries of common language embrace, not the whole
idea. Thus, of men, Briton and alien are contraries: The same may be said of be one of the two, no man can be both. integer and fraction among numbers, peer and commoner among subjects of a realm, male and female among animals, and so on. In order to express this, let us say that the whole idea under consideration is the universe (meaning merely the whole of which we are considering parts) and let names which have nothing in common, _but which between them contain the whole of the idea under ^nsideratiqn, be called contraries in, or with respect to,
every
man must
that
universe."
61
after
(Formal Logic, p. 37; see also Camb. Phil. Trans., vm, 380.) Formal Logic: or, The Calculus of Inference, Necessary and Probable, 1847. to be cited as F. L.
Here
38
Although the work of De IN [organ is strictly contemporary with that of Boole, his methods and symbolism ally him rather more with his prede
cessors than with Boole
follow.
Like Hamilton, he
is
bent
upon improving the traditional Aristotelian logic. His first step in this direction is to enlarge the number of typical propositions by considering
all
and F, and
their
negatives.
De Morgan
affirmative or negative of the proposition are indicated, these being sufficient to determine completely the type of
That a term
^before or after
A"
is
distributed
is
indicated
by writing
half a parenthesis
A"),
it,
letter, thus:
or (X.
An
undistributed term
A"(,
or
is
distributed
is
undistributed,
is
F".
X()Y
distributed, that
is,
"Some
The negative
of a term,
is
X,
is
indi
cated by
of
F by
y, etc.
indicated
by
a dot
will
is
not
Y"
an affirmative proposition. All the different forms of proposition which De Morgan uses can be from two types, the universal, "All is generated ./ and the
A"((F.
be
64
Two
particular,
tive,
.r,
"Some
F and
y.
.,"
X and
its
nega
Syllabus of a Proposed System of Logic, 1860. Hereafter to be cited as Syll. Five papers (the first not numbered; various titles) in Camb. Phil. Trans., vm, ix, x. The articles contain the most valuable material, but they are illarranged and inter
spersed with inapposite discussion. Accordingly, the best get these articles and the Formal Logic, note in a general
way way
to study
De Morgan
is
to
then use the Syllabus as a point of departure for each item in which one is interested. 62 He does not speak of "distribution" but of terms which are "universally spoken
cr
"particularly
of"
spoken
of
",
or of the
"quantity"
of a term.
63 This is the notation of Syll. and of the articles, after the first, in Camb. Phil. Trans. For a table comparing the different symbolisms which he used, see Camb. Phil. Trans.,
ix, 91.
64
It is
criteria
whether
De Morgan
them
X and
(Syll., p. 13):
"Let
He
Thus Every A is Y is Some things are neither Some things are neither
things are neither
or x; negative, which is true of and x, false of and X. affirmative: is is true; is x is false. But Every Every
notA"
A s nor F s A s nor A s
is
is is
also affirmative,
true,
A s nor z s
false."
39
A))
x))y
MIX
All
is
Y.
is
All not
A"
not 7.
X))y
is
not7.
7.
x))Y
:
AllnotZis
and
for particulars
we have
(5)
X()
Some
X X
is
7.
is
x()y
Some notX
not 7.
X()y
x()Y
Some
Some
is
not 7.
is
notJT
7.
The
its
rule for transforming a proposition into other equivalent forms may be stated as follows: Change the distribution of either term that is, turn
parenthetic curve the other way, change that term into its negative, and change the quality of the proposition. That this rule is valid will appear if we remember that "two negatives make an affirmative", and note
that we introduce one negative by changing the term, another by changing the quality of the proposition. That the distribution of the altered term
should be changed follows from the fact that whatever proposition distrib
utes a term leaves the negative of that term undistributed,
and whatever
term.
Using this rule of transformation, we get the following table of equivalents for our eight propositions:
(a)
(1) (2)
(3)
(b)
(c)
(d)
XY
x))y
= X)(y
=

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

X))y
X).(Y = x((Y
a)Q/
=
(4) .r))7
(5)
X((y
X()Y
T()
7 = X((y
*x(.(Y
.=
=x)(y
=
=x).)Y
(G) x()i/
(7) (8)
XQy
a()7
=x(.(y
X}(y =X))Y
is
It will
terms positive,
A and
Y.
Selecting these,
we have the
of propositions:
40
A
(la)
(2c)
A))F
All
A"
Y.
is all
A ((F
X)(Y
X()Y
r
Some
F;
or, All
is
X.
(36)
NoZis
Some
Some
Some
All
Y.
is
(4d)
Everything
either
X or
Y.
(See below.)
(5o)
(6c)
A ()F
X)(Y
X((Y X))Y
A
A"
is
7.
X nor
F.
(See below.)
(76)
not Y.
or,
(Sd)
is
not some Y;
is
Some 7
any one
is
not X.
indicated,
of these propositions
may
that
is
is,
w ith Y
r
subject
(4d)
and
X predicate X
is
preserved.
and
We
"Some
not
but we hardly get from that the difference between X()Y and some is not (any) A ((F, "Some Also, A"()F is equivalent to uniis not some would make it par and the reading, "Some versals,
F".
F",
ticular.
"Xo
A"()F
is
equivalent to
cc))
F,
"All
notZ
is
F",
and
to x)(y,
notX
is,
is
notF".
of these
A"
and F
"Everything
either
or
F
or
(or
"
both)".
(6c),
X)(Y, we should be
";
"All
is all
F",
X and
F are equivalent
its
equivalents
is
notF".
are
particular
propositions.
of this in
The equivalent
neither
terms of
xQy, X and
is
"Some
notX
is
plainly,
"Some
things are
X nor
F".
(1) will
We
X))Y
X((Y
A&gt;(F
contradicts
"
(76)
X((Y
(Sd)
"
AF
X
",
(36)
(5a)
(6c)
XQY
X)(Y
X()Y
65
which it might seem that De Morgan himself has fallen. See e. g., is all F SylL, p. 25, and Camb. Phil. Trans., ix, 98, where he translates X)(Y by "All or "Any one X is any one Y But this belongs to another interpretation, the "cumular", which requires X and Y to be singular, and notX and not Y will then have common members. However, as we shall note later, there is a real difficulty.
error into
".
An
66
De Morgan
he
calls contradictory
propositions
"contraries"
p. 11), just as
calls
41
Thus the
of terms
rule
is
two
cases.
The
rule
may
any symbolized proposition be read or written backwards, provided the distribution of the terms
from universals, particulars from particulars.
"Each
For the
rest,
we have
67
the rule,
same
quality".
For
and
as follows:
"There is inference: 1. When both the premises are universal; 2. When, one premise only being particular, the middle term has different quantities
in the
two premises.
conclusion
is
"The
[parenthetic
curves]."
found by erasing the middle term and its quantities This rule of inference is stated for the special
arrangement of the syllogism in which the minor premise is put first, and the minor term first in the premise, the major term being the last in the
second premise.
Since
this
X))Y,
"All
X
is
is
F",
Z",
give X)(Z,
X

is
Z".
A&gt;(F,
"No
is
not
Z",
give
A")
(Z, or
A ) (Z, which
"Some
nor
Z."
The
any
of
De Morgan
De Morgan
pound syllogisms
"1.
X)0)Forboth A))Fand
AF
All
Xs
sides are
2.
s.
For
both
A ))Fand A ((F
X((Y and X((Y
All
3.
X(0 (F
)
or both
all
are
s.
and
some things
4.
besides.
A O (For
X)(Y
and
F and
neither.
67
Sytt., p. 16.
42
A
5.
A&gt;(Fand
X()Y
6.
X(O)Y
or both
X()Y and
XQY
or
Y and
some things
both."
Each
sign
of these propositions
for the
meaning
of the
The
slightly different:
or distribution, of
negative.
We
change at the same time the quality of the proposition. This difference is due to the manner in which the propositions are compounded.
The
"
syllogistic",
compound
or 6,
any two be joined, each of which is [of the form of] with the middle term of different quantities, these premises
clusion of the
1, 3, 4,
yield a con
same kind, obtained by erasing the symbols of the middle term and one of the symbols [O]. Thus X)O(Y(O)Z gives X)O)Z: or
if
nothing be both
X and
7 and some
things both,
if
everything be
tivo lots
either
other
"
follows that
all
and
of
it
follows that
may
be written for
one place, without any alteration of the conclusion, except )O( )O) reducing the two lots to one. But if this be done in both places, the con clusion is reduced to or and both lots disappear. Let the reader

examine
which one
of the premises
is
cut
down
to a
simple universal.
"The Letters written following exercises will exemplify what precedes. under one another are names of the same object. Here is a universe of 12
P s;
5 are
and the
remainder Q
s;
7 are
Z s and
R s.
XXX PP PP PPPPP
YYY YY QQ QQQQQ Z Z Z Z Z Z Z RRRRR
We
can thus verify the eight complex syllogisms
X)0)Y)0)Z P(0)Y)Q(R
70
P(0)Y)0)Z
P(0(Q(O)Z
X)0)Y)0(R
X)Q(Q(0(R
P(O(Q(O(R X)O(Q(O)Z
Sytt., p. 23.
45f
be seen that the two lots in the middle form the quantity of the particular proposition of the conclusion."
will
In so
too
of his
work
as
we have thus
far reviewed,
simplifies traditional
it is still
He
inference from universals to particulars nor observe the problems there involved. 71 He does not seek a method by which any number of terms
may
be dealt with but accepts the limitation to the traditional two. And his symbolism has several defects. The dot introduced between the
parenthetic curves
is
read ()
as,
"It
is
make
is )
it
possible to
(,
so that this
is
formation applicable only when no more than two terms are involved
prepositional
relation.
term
in a proposition
may
be distributed or
"mentioned universally",
and
De Morgan,
fails
This
is
is
like
"All
A"
all
7",
and
really
is
"Some
things are
Mathematical symbols are introduced but without any mathematical operations. The sign of equality is used both corresponding for the symmetrical relation of equivalent propositions and for the un".
nor
72
73 symmetrical relation of premises to their conclusion. His investigation of the logic of relations, however,
is
more
successful,
and he
is
This topic suggested to him by consideration of the formal and material elements
laid the
field.
in logic.
71
He
says:
7l
simple term has members. ideas, is tacitly claimed for the terms of every syllogism". 72 A universal affirmative distributes its subject in the sense that it indicates the class to which every member of the subject belongs, i. e., the class denoted by the predicate.
Similarly, the universal negative,
all inference (in extension) of a necessarily based: the assumption that a class denoted by a He says (F. L., pp. 110), "Existence as objects, or existence as
notA
is F, indicates that every is not Y, every Y is No The particular nega particular proposition distributes a term in that sense. tive tells us only that the predicate is excluded from some unspecified portion of the class and Y in this sense only. Comparison with denoted by the subject. A)(F distributes
.
No
its
it
can
tell us,
of
A",
only that
it is
specified portion of notF; and of Y, only that it is excluded from of notA. is wholly included in Y, or cannot infer that
We
it.
In one passage (Camb. Phil. Trans., x, 183) he suggests that the relation of two C. premises to their conclusion should be symbolized by A B
&lt;
74
Camb.
44
there
A
"Is
of the
machinery?
Is not consequence an action not logic the science of the action of the machinery?
act of the
Consequence
to ask,
on,
always an
mind
the
What
kind of act?
What
is
act,
of
certain functions;
it is
competent to those
functions
validate
because
. . .
it
its use.
it
The word
which
identifies,
does not do
is
its
work
because
identifies,
a transitive and
;
convertible
motion:
is
that which is
B means A
is
,
is
and
A
is
is
B
fit
means
Hence every
transitive
as
and by the same proof in each case. Some forms are valid when the relation is only transitive and not Thus if Y represent convertible; as in give and 7 connected
.
by
is
valid, as in
EveryZ
... In the
7,
No
X
:
Y,
therefore
No
is
X~Z.
exclusion of matter,
Hypothesis
(Positively true)
is is
animal
Y
a
/?
has existence.
X 7 Every X 7 a of X 7
(Probability
(3)
X has existence.
is
a transitive relation.
&lt;
is is
a fraction
or or
1.
a of .Y
a fraction
&lt;
1.
The
point about
"...
except the transit! veness of the copula. 75 I hold the supreme form of the syllogism of one middle term to
is
be as follows: There
there
is
is
in the relation
the probability
the probability a /3
"...
to 7;
is
is
in the relation
whence there
76
The copula
of the
post hoc
is
and effect, of motive and action, of all which form and propter hoc (perhaps) of the matter, will one day
of cause
77
logic."
77
45
De Morgan
general,
is
and to consideration
thus led to a study of the categories of exact thinking in of the types and properties of relations.
His division of categories into logicomathematical, logicophysical, logico78 metaphysical, and logicocontraphysical,
is
inauspicious,
and nothing
"The
much comes
of
it.
But
it,
in
connection with
this,
sum]
the
compound
[logical
product] of the contraries of the aggregants: the contrary of a is the aggregate of the contraries of the components." 79
compound
For the
L, of
logic of relations,
X, Y, and
will
M, N,
.
relations.
LY will
is
signify that
A"
some one
of the objects
in the relation L, or is
one of the
. .
of 7. 80
X LY will signify that X not any one of the L s of Y. X (LM) Y or X LM Y will express the fact that AT is one of the L s of one of the M s of y, or that X has the relation L to some Z which has the relation M to Y. X LM Y will mean that X is not an L of any M of Y.
.
.
It should be
L and
.
M
. .
is
.
what
.
we should
If
"
call
today
.
.
their
relative product
.
that
. .
is,
A"
together give A^
LM Z, but A
"brother
of"
LY and A
is
the relation
"
and
MY do not give
"aunt
is
the relation
Y".
will
mean
is
a brother of an aunt of
of
(Do not
MZ LM Y. A LM Y of say hastily, X
LY and Y
A"
",
"
is
uncle of
7".
"Brother
an
aunt"
is
not equivalent to
written by
"uncle"
since
sisters.)
L, or
M,
itself, will
represent
M,
that
is,
LY
X which
a brother of T. 81
is,
to the form in
it is
which the
necessary
copula
may
L of Accordingly every M, that is, something which has the relation L to every member of L*M is to indicate an L of none the class (say, a lover of every man). but The mark of quantity, * or *, always s (a lover of none but men).
is
LM*
to signify an
M
78 79
See
ibid., p. 190.
Ibid., p. 208.
"On
See also
Sytt., p. 41.
3,
summary
the ideas
of the paper,
80
81
and on Logic
Camb.
We
1 tried at first to
make De Morgan
follow the order of the paper from this point on. s symbolism more readily intelligible by intro
ducing the current equivalents of his characters. But his systematic ambiguities, such as the use of the same letter for the relation and for that which has the relation, made this impossible. For typographical reasons, I use the asterisk where he has a small accent.
46
it,
but
L*M
L
of
is
read as
if
*]
modified
suggests
To
De Morgan
"An
L*M
an Z,
be read,
"
"An
every!, of If;
an
but we
none but
s".
M and
1
of
none but 3f
s:
LMX*,
of
an
of
an
M of
.
If
Two more L is
".
The converse
of";
L is
l symbolized by L~
"lover
L
is "beloved
if
is
"aunt",
L
is
"niece
or
nephew
ized
The contrary
(or as
we should
L is symbol
:
by m. by 1; the contrary of In terms of these relations, the following theorems can be stated
Contraries of converses are themselves contraries.
The contrary
If
of the converse
is
(4)
the relation
L
l
,
be contained
is
converse of L,
L~
the
the
contrary of
M, m,
is
I.
For example,
if "parent
is
contained in
"ancestor
of",
(a)
"child
of"
is contained in "descendent
"not
of",
and
(b)
"not
ancestor
of"
is
contained in
parent
of".
compound relation is accomplished by converting both components and inverting their order; thus, (LM) 1 = M~ l L~ l
(5)
.
The conversion
of a
If
X be teacher of the
child of Y,
is
When
a sign of quantity
its
If
X be
is
(6)
When,
compound
relation, there
its
a sign of quantity,
if
each
contrary, and the sign of quantity be shifted the other and its position on the letter changed,
thus
LM* =
l*m and
L*M =
A
lm*.
lover of every
man
is
of none but
men
is
47
When
compound
relation involves
no sign of quantity,
its
contrary
to the other.
"Not
The contrary of LM is IM* or L*m. is "nonlover of every man" (lover of a man) but nonmen"; and there are two equivalents, by
((&gt;).
or
"lover
of
none
But
if
is
taken by dropping that sign and taking the contrary of the other component.
The contrary
"Not
of
LM*
"
is
IM;
of
L*M
is
Lm.
of a
man";
(lover of every
man)" is "nonlover
".
and
"not
(lover
of
none but men) is "lover of a nonman So far as they do not involve quantifications, these theorems are familiar though
seems not generally known that they are due to The following table contains all of them:
it
to us today,
De Morgan.
Combination
Converse
Contrary
l
LM
M L~
l
/J/* or
1
L*M
J/*
/ 1
or
l
m~ Ll
LM*mhm
L*Morlm*
The
M*~ L~
l
or
or
M^L
m~Hm* l~
}
*
l
IM
M~
l
l~ l
l
Lm
is
m~ L~
"contained
in"
said to be
or to
"imply"
is
contained in
Every grandfather
not contained in
is
The
uses
"All
relation "grandfather
is
contained in
"ancestor
father of
is
contained in
is,
is
M",
where
L and
L))M.
If
is
that
is,
L))M
gives m))l.
Applying
(8
)
this
theorem to compound
relations,
we have:
LM))N
If
(8")
))/.
l
l
.
Proof: If
But an
of
48
every
M of an
M~
of
Z must
of Z.
l
Hence
nM~ ))l.
l
.
Again;
if
LM))N,
l
then ri))L*m.
relation converseofZ to
Whence L~ n))L L*m. But whatever has the an L of none but m s must be itself an m. Hence
K"
L~ n))m. De Morgan
(9)
If
from
1
its
LM
= N,
then
Proof: If
for
any X,
LM MM~ X
1 1
and L~
LX
l
theorem.
We
do not have
L =
==
NM~
and
l
M
=
==
L~ N, because
1
:
it is
not generally
of the of the
parent of
then
X"
will contain
l
X.
1
The
relation
MM~
or
cancel out.
MM~
1 1
and
if
XMM~
YMM~ X. If X is child of a parent of Y, then Y is child of a parent of X. But MM and M~ M are nctf exclusively reflexive. XMM^X does
If we know that a child of the parent of X is a celebrated we may not hastily assume that X is the linguist in question. linguist With reference to transitive relations, we may quote 82
"A
relation
is
transitive
when a
relative of a relative
is
a relative of
so on.
ZZZ))ZZ))Z; and
transitive relation has a transitive converse, but not necessarily a transitive contrary: for L~ l L~ l is the converse of LL, so that ZZ))Z gives
L L
1
its
From these, by contraposition, and also by theorem contrapositions, we obtain the following results:
*))L
l
.
and
L
I
is
contained in
LL
*,
ZZ"
1
,
Z!Z ...... Z 1
descendent.
Camb.
For
XLL^Y,
this discussion of transitive relations, De Morgan as also reflexive, though not necessarily
exclusively reflexive.
49
ancestor
is
always an ancestor of
all
descendents, a nonancestor
none but nondescendents, a nondescendent of all nonancestors, and a descendent of none but ancestors. A descendent is always an ancestor of
of
all nondescendents, a nondescend ent of none but nonancestors, and a descendent of all ancestors. A nonis always a nonancestor of all ancestors, and an ancestor of none but nonancestors. A nondescendent is a descendent of none but non
descendents,
and a nondescendent
all
of
all
descendents.
Among
all
nonnonall
descendents.
Among
all
nondescendents
are
contained
nondescendents of
ancestors."
M,
may
be tabulated as follows:
and
l
replacing
.
it
contraries.
1M~ Z
gives
for
Ibid., p. 350.
50
A
When
the copula of
all
X..LY
I
X.LY
Y..L1Z
X..LY
Y.L~ Z X L~ Z
1
Y..LZ
X ..LZ
X .LZ
Y ..LX
IV
Y .LX
Y ..LX
1
Z.L Y
1
X.LZ
Here, again, in the logic of relations,
Z..L~ Y X.L~ Z
1
Z..LY X..L~
1
Z"
De Morgan would
done better
if
he had
left
The
introduction of quantifications and the systematic ambiguity of L, M, etc., which are used to indicate both the relation and that which has the
hurry him into complications before the simple analysis of rela tions, and types of relations, is ready for them. This logic of relations was
relation,
destined to find
its
importance
in the logistic of
of,
mathematics, not
in
any
Aristotelian logic.
And
these compli
De Morgan
s,
due largely to
had to be discarded
Boole
s
algebra and
relation theory.
The
reintroduced by the work of Frege and Peano, and more especially of Whitehead and Russell, in the logistic development of mathematics. But
it is
there separated
was De Morgan who laid the foundation; and if some part of his work had to be discarded, still his contribution was indispensable and of
that
it
permanent value.
84
85
:
85
51
here the general idea of relation emerges, and for the first time in the history of knowledge, the notions of relation and relation of relation
are symbolized.
And
here again
is
is
the
difference of form at one step of the ascent difference of matter at the next. But the relation of algebra to the
manner
in
which what
higher
developments of logic
It will hereafter
it
is
tion of
Omnis homo
est
animal, Sortes
est
homo,
etc.,
atmosphere of syllogism, the unceasing composition of was admitted that such an atmosphere existed." M
V.
BOOLE
The beginning from which symbolic logic has had a continuous develop 87 is that made by His significant and vital contribution George Boole. was the introduction, in a fashion more general and systematic than before,
ment
of
Indeed Boole allows operations which have no direct logical interpretation, and is obviously more at home in mathe matics than in logic. It is probably the great advantage of Boole s work
that he either neglected or was ignorant of those refinements of logical theory which hampered his predecessors. The precise mathematical
mathematical operations.
development of logic needed to make its own conventions and interpreta tions; and this could not be done without sweeping aside the accumulated
traditions of the nonsymbolic Aristotelian logic. As we shall see, all the nice problems of intension and extension, of the existential import of universals
and
particulars, of
It
is
empty
classes,
and
and demand
consideration.
enough
well that, with Boole, they are given a vacation long to get the subject started in terms of a simple and general procedure.
s first
book, The Mathematical Analysis of Logic, being an Essay toward a Calculm of Deductive Reasoning, was published in 1847, on the
86
Boole
omit, with
of
De Morgan
contributions to prob
and judgment. (See SylL, pp. 6772; 38487, and 393405.) His work on this topic is less closely connected with symbolic logic than was Boole s. The allied subject of the "numerically definite syllogism" (see Syll., pp. 2730; F. L., Chap, vm; and Camb.
Phil. Trans., x, *355*358)
87
ability theory as applied to questions of authority F. L., Chap, ix, x; and Camb. Phil. Trans., vm,
is
also omitted.
George Boole (18151864) appointed Professor of Mathematics in Queen s College, Cork, 1849; LL.D. (Dublin, 1852), F.R.S. (1857), D.C.L. (Oxford, 1859). For a biographi cal sketch, by Harley, see Brit. Quart. See also Proc. Roy. Rev., XLIV (1866), 14181.
Soc.,
xv
(1867), vixi.
52
o
~*same day as
Calculus of
article
The next
"The
Logic,"
This
summarizes very
posed by Boole.
in
briefly clearly the important innovations pro But the authoritative statement of his system is found
An Investigation of the Laws of Thought, on which are founded the matical Theories of Logic and Probability, published in 1854. 89
Boole
s
Mathe
is
based squarely
The
upon which
(1) symbols"; (2) the Jaws of thought expressed as rules for operations upon these symbols; (3) the observation that these rules of operation are the same which would hold for an algebra of the numbers and I. 90
method depends
are:
the conception of
elective
will
"
universe of conceivable
represented by
by
selection or limitation.
l.r
This operation of
electing,
all
the
A"s,
is
represented by
or x;
the
s is
similarly represented
by
1y or y,
and
so on.
not distinguish between this operation of election represented by x, and the result of performing that operation an ambiguity common in mathe matics
.r
all
the A^
s.
Thus
x, y, z, etc.,
them from
coefficients.
This operation of election suggests arithmetical multiplication: the suggestion becomes stronger when we note that it is not confined to 1.
1
"
x y or
xy
universe",
all all
the
in the
the Fs.
The
result of these
A"s
and F
s.
will
class of the
common members
of x
and
to
y;
z,
xyz, the class of those things which belong at once to and so on. And for any x, lx = x.
x, to y,
and
The operation of "aggregating parts into a whole" is represented by + + y symbolizes the class formed by x combining the two distinct classes,
.
x and
y.
It is a distinctive feature of
Boole
The
"On
relation
may
be read,
"that
which
Camb.
See
De Morgan s note to
the article
Propositions Numerically
Definite",
Work
first
53
either x or y but not both". Although Boole does not remark it, x + y cannot be as completely analogous to the corresponding operation of ordinary algebra as xy is to the ordinary algebraic product. In numerical
number may be added to itself: but since Boole conceives the terms of any logical sum to be quite distinct 91 mutually exclusive classes, x + x cannot have a meaning in his system. As we shall see, this is very awkward, because such expressions still occur in his algebra and have to be
algebras a
"
",
But making the relation x + y completely disjunctive has one advantage it makes possible the inverse relation of "subtraction". The "separa
tion of a part, x, from a whole,
?/",
is
represented by y
x.
If
x+
z
]
have nothing
in
common, y
and y
= =
y,
x.
Hence
x +
of x or
and
y,
members
members
members of x and members of y. x y repre members of x which are not members of y the x s
two
classes
except the y
s.
i.
=
]
same members,
e.,
The
entity
(1
.r)
.r
is
of especial
all
s,
or
importance. This represents the It is, then, the things which are not x s.
supplement or negative of x.
of this
symbolism for the negative of a class, the sum of two which have members in common, can be represented by
xy + x(l
#)
(!
x)y.
s
The
first
term of
this
s
sum
is
and y
s;
the second,
but not y
but not x
s.
Thus the three terms represent classes which are all mutually exclusive, and the sum satisfies the meaning of + In a similar fashion, x + y may
.
be expanded to
x(l
y)
(l
x)y.
Consideration of the laws of thought and of the meaning of these bols will show us that the following principles hold
:
sym
x.
(1)
xy
yx
(2)
91
x +y
y+x
What What
is is
See L. of
54
A
(3) z(x
+ y)
zx + zy
That which
is
is
both
z
and and
(either
z
.r
or y)
y.
either both
is
and x or both
z
(a;
and
(4) z(.r
y)
zx
zy
That which
is
both
but not
2
y)
y.
both
and
(5)
If
.T
y,
then
= z + x = xz =
Z.T
.r
zy
z
+y
(6)
.r
= 
y +
This
last is
first
symbols"
that
if
the operation
.r,
upon
1,
is
=
(7)
lxx
2
=
X
x x x
Hence we have
.T
Boole
calls this
the
"index law".
92
All these
laws, except
logic,
"If
(7),
It
may
At
be
noted that, in
glance, this
.r
y,
then zx
=
zy"
is
not reversible.
first
may
in question.
But
"If
zx
zy,
then x
=
y"
Law (7) is, then, the distinguishing in numerical algebra when z = 0. The only finite numbers for which it holds are principle of this algebra. and 1. With 1. All the above laws hold for an algebra of the numbers and
this observation, Boole
2 modified by the law x
adopts the entire procedure of ordinary algebra, x, introduces numerical coefficients other than
and
1,
and makes
use,
properties of functions,
to serve his purpose. 93
and
on occasion, of the operation of division, of the of any algebraic transformations which happen
This borrowing of algebraic operations which often have no logical interpretation is at first confusing to the student of logic; and commen
tators have seemed to smile indulgently
upon
it.
An example
it,
will help:
"law
the
"law
from the
"index law".
92 In Mathematical Analysis of Logic he gives he avoids this, probably because the factors of x n
it
also in the
form x n
x) are
(e. g.,
z3
iriterpretable.
93 and 1, and the fractions which can This procedure characterizes L. of T. Only be formed from them appear in Math. An. of Logic, and the use of division and of fractional coefficients is not successfully explained in that book.
94
L. of T., p. 49.
55
.r,
a
0.
Hence, factoring,
.r(l
x)
0.
Whoever says
is
This transformation hardly represents any process of logical deduction. 2 "What is both x and a, a* is equivalent to x; therefore what
,
is nothing"
may well
if
.r 2
ing.
Nor should we be
satisfied
.r
0.
way
of using uninterpretable
thoroughly sound. Boole s algebra may be viewed as an abstract mathematical system, generated by the laws we have noted, which has two interpretations. On the one hand, the "logical" or "elec
tive"
cal
symbols may be interpreted as variables whose value is either numeri or numerical 1, although numerical coefficients other than and 1 are
be remembered that such coefficients do not obey which holds for "elective" symbols. All the usual alge
it
admissible, provided
the
"index law"
On
the
symbols may be interpreted as classes. For this interpretation, some of the algebraical processes logical of the system and some resultant expressions will not be expressible in terms
"logical"
or
"elective"
of logic.
But whenever they are interpretable, they will be valid conse quences of the premises, and even when they are not interpretable, any
further results, derived from them, which are interpretable, will also be
valid consequences of the premises.
It
proprieties
His consistent course would have been to develop this al without reference to logical meanings, and then to discuss in a thorough gebra
of his procedure.
fashion the interpretation, and the limits of that interpretation, for logical
such a method, he would have avoided, for example, the + .r. We should have + x = 2.r, the interpretation of difficulty about
classes.
By
.1
.1
which
classes
for the
numbers
and
1 is
obvious, and
its
which
will
would depend upon certain conventions which Boole made and be explained shortly. The point is that the two interpretations
should be kept separate, although the processes of the system need not be limited by the narrower interpretation that for logical classes. Instead
of
making this separation of the abstract algebra and its two interpretations, Boole takes himself to be developing a calculus of logic; he observes that
"axioms"
95
its
numbers
and
95
1;
56
hence he applies the whole machinery of that algebra, yet arbitrarily rejects from it any expressions which are not finally interpretable in terms of logical The retaining of noninterpretable expressions which can be relations.
transformed into interpretable expressions he compares to
ment
of the uninterpretable
"
of trigonometry.
96
It
"the employ symbol V 1 in the intermediate processes would be a pretty piece of research to take Boole s
it (his
it,
and systematically
investi
gate
its
interpretations.
of
Boole
symbolism
will
now be
entirely clear.
Classes
y),
be represented by
x, y, z, etc.;
their negatives,
by
(1
is
x), (1
That which
.r(l
is
both x and y
be
;r(l
i.
y), etc.
That which
is
y) + (1
x}y.
is
That which
is
x)y
e.,
that which
x or not x but y
xy + x(l
that which
"universe"
(1
x)y
x.
1
is
represents the
is
or
The
logical significance of
determined
Oy = 0: the only class which remains unaltered of electing from it whatever is the class "nothing". by any operation Since Boole s algebra is the basis of the classic algebra of which logic
by the
any
y,
the topic of the next chapter it will be unnecessary to comment upon those parts of Boole s procedure which were taken over into the classic
is
algebra.
These
will
be clear to any
who understand
in its current
II.
turn our attention chiefly to those parts of his method which are peculiar to him.
shall, then,
We
"x
is
included in
?/".
Conse
quently the only copula by which the relation of terms in a proposition can be represented is the relation =. And since all relations are taken in
extension, x
membership.
y symbolizes the fact that x and y are classes with identical Propositions must be represented by equations in which
or == 1, or else the predicate must be quantified. something is put = Boole uses both methods, but mainly relies upon quantification of the This involves an awkward procedure, though one which still predicate.
survives
96
L. of T., p. 69.
57
and symbolize
vy:
"Some
class"
"Some".
Thus
#",
"All
is
(some)
If
v,
i/"
is
represented by
"the
is
(some)
by wx =
this
vy.
or w. were here
indefinite
or
v,
"any
class",
method would be
that
less objectionable.
it
But
class
in
such cases
or w,
must be very
but
this,
definitely specified:
it
97
must be a
of
"indefinite
in all respects
contains some
members
it is prefixed".
The
universal affirmative
can also be expressed, without this symbol for the indeterminate, as .r(l y) = 0; "All xisy" means "That which is x but not y is nothing". Negative
cate.
propositions are treated as affirmative propositions with a negative predi So the four typical propositions of traditional logic are expressed as
98
follows:
All x
is is
y:
y:
is
x
.r
= =
vy,
v(l
or,
x(l
y)
0. 0.
Xo
y), y),
y),
or
or, or,
Some x
y:
v.r
= =
w(\
xy = v =
v
xy.
Some
Each
.r
is
not y:
vx
w(l
x(l
y).
which
may
be readily deter
of the algebra.
To
reason
by the
proper manner and then operate upon the resultant equation according to the laws of the algebra. Or, as Boole more explicitly
premises explicitly in the
puts
"
it,
"
1st,
That a
of the
symbols be correctly determined from that interpretation. "2nd, That the formal processes of solution or demonstration be con
ducted throughout in obedience to all the laws determined as above, with out regard to the question of the interpretation of the particular results
obtained.
"3rd,
That the
final result
it
be
data."
As we
97
shall
see,
Boole
methods
of solution
is
sometimes involve an
provided a machinery by
This translation of the arbitrary v by "Some" is unwarranted, and inconsistent with Boole s later treatment of the arbitrary coefficient.
null.
There
98 99
no reason why such an arbitrary coefficient may not be See Math. An. of Logic, pp. 2122; L. of T., Chap. iv.
L. of T., p. 68.
58
A
may
interpretable.
To
we must first understand the process known as the develop comprehend ment of a function. With regard to this, we can be brief, because Boole s method of development belongs also to the classic algebra and is essentially
the process explained in the next chapter. 100
Any
the form
(1
x)
may
it
be
called a function of x.
A
.r).
function of x
is
said to be developed
when
has
Ax + B(l
x be a
"logical symbol",
and
be
1.
But the
"logical
coefficients,
symbol"
not so limited: A, or B,
"law
may
such a
of
duality",
or
If
it
may
or involve
such a number.
Ax
+ B(1
.r), it
may
Let
/(.r)
Then
/(I) /(O)
And
Hence
f(x)
x)
Thus
if
f(x)
~,
2x +
;
Hence
f(x)
(1
x)
will
y)
And
for
any function,
/(.r, y),
f(*,y)
y)
+/((),
!)(! x)y
See Math. An. of Logic, pp. 6069; L. of T., pp. 7179; "The Calculus of Logic," Cambridge and Dublin Math. Jour., m, 18889. That this same method of development should belong both to Boole s algebra and to the remodeled algebra of logic, in which + is not completely But a completely developed function, disjunctive, is at first surprising. either algebra, is always a sum of terms any two of which have nothing in common. This accounts for the identity of form where there is a real and important difference in the meaning of the symbols.
[
59
Thus
if
f( x y)
,
= = =
ax + 2by,
a1 + 261
a1
/(I, 1) /(I, 0)
/((), 1)
=
=
a + 26
+ 260
261
a
26
/((),
0)
= a0 + = tt0 +
= =
260
Hence
/(.r,
y)
a + 21) xy + ay (I
y)
+ 26(1
.r)y
An
any number
of variables.
In the words of
Boole:
101
"The
consist of
two
parts, the
formation of the constituents of the expansion, the second to the determination of their respective coefficients. It is as
follows
:
first of
which
"1st.
the
symbols x,
y, z
Form
a series
first
constituent be the
product of the symbols: change in this product any symbol z into 1 z, for the second constituent. Then in both these change any other symbol Then in the four constituents y into 1 y, for two more constituents.
thus obtained change any other symbol .r into 1 .r, for four new constit uents, and so on until the number of possible changes has been exhausted.
"2ndly.
To find
the coefficient of
any
constituent
If
.r
that constituent
into
1
;
but
if it
x as a factor, change in the original function .r into 0. Apply the same rule with reference to the symbols y, z, etc.: the final calculated
value of the function thus transformed will be the coefficient
sought."
Two
solutions
is 0.
and interpretations,
are:
(1)
The product
of
If one constituent be, for example, xyz, any other constituent will have as a factor one or more of the negatives, 1 1 z. .r, y, 1 Thus the product of the two will have a factor of the form x(l And .r).
where
is
.r
is
a
0.
"logical
symbol
",
and
1,
x(l
x)
always
ficient 1, the
All information
may
by a
solution, to
.r,
"logical
symbol"
or
L.
o/T
.,
pp. 7576.
60
A
by an
(2)
elimination, to discover
what statements
independent of volve
x, or (3)
some term
x, are warranted
by a combination
/,
of x in terms of
other
result.
from equations which involve x, t, u, v, and some logical" symbol or symbols which must be eliminated in the desired Formal reasoning is accomplished by the elimination of "middle"
u,
v,
" "
terms.
The student
of equations
set
may
involving a term x
be combined into a single equation, that any equation = 0, and that may be given the form Ax + B(l x)
the result of eliminating x from such an equation is solution of any such equation, provided the condition
will
AB = AB =
0.
Also, the
be satisfied,
s
be x
= B
1,
+ v(l
A), where
is
undetermined.
Boole
methods
achieve these same results, but the presence of numerical coefficients other
than
and
and
division,
And
in which we should expect to find it but in a more complicated fashion which nevertheless gives equivalent results. We have
form
now
by which which have been mentioned. algebra The simplest form of equation is that in which a developed function,
difficulties of his
of
any number
of variables,
is
equated to
0, as:
Ax + B(l x) =
Axy +
It
is
0,
or,
.r)(l
B.r(l
y)
+ C(l
x)y + D(l
y)
0,
etc.
an important property of such equations that, since the product of in a developed function is 0, any such equation gives any one of its constituents, whose coefficient does not vanish in the develop ment, = 0. For example, if we multiply the second of the equations given
by
we
xy, all constituents after the first will vanish, giving shall have xy = 0.
Axy =
0.
Whence
Any equation in which a developed function is equated to reduced to the form in which one member is by the law;
1
may be If V = 1,
1
"

V =
0.
The more
symbol",
is
logical
w,
is
suppose x
x =
equated to some function of such symbols. For example, yz, and it be desired to interpret z as a function of x and y.
yz gives z
x/y
is
We
shall, then,
61
x)y
+/(0,0)(1 x)(l
y)
By
this law:
If z
3*
, then
= x+x(l
These fractional
7)
+ 0(1
system. Because any function can always be de and the difference between any two developed functions, of the veloped,
his
same
variables,
is
always confined
to
the
coefficients.
If,
then,
we can
interpret
difficulties of
system
is
removed.
The fraction 0/0 is indeterminate, and this suggests that a proper inter pretation of the coefficient 0/0 would be to regard it as indicating an unde termined portion of the class whose coefficient it is. This interpretation may be corroborated by considering the symbolic interpretation of "All
x
is
?/",
which
y)
is
x(l
0,
y)
0.
If
x(l
then x
xy
and x =
xy.
Whence y
x/x.
x + 
(1
x).
is
?/",
the class y
is
not.r.
made up Whence
an undetermined
0/0
is
equivalent to an arbitrary
parameter
or as
"All,
v,
"an
undetermined portion
of"
The
obey the
.r(l
x~ =
.r,
or
its
equivalent, the
of
duality",
x)
0.
At
system.
But
in
any
its logical
interpretation goes.
Any constituent of a
developed function which does not satisfy the index law must be separately
equated
to 0.
Suppose that
in
any equation
w = At + P
62
A
be a
"
logical
symbol",
whose
be the
law.
coefficient
be a constituent of a developed function does not satisfy the index law, A 2 = A. And let P
t
and
sum Then
whose
coefficients
do
satisfy this
w =
2
w,
t*
t,
and
P = P
2
is 0,
Hence
w = (At+P) = A w = A t+P
2 2 2
2 2
t
(A
Hence
since
 A )t = = A(l = A(l  A) 4= 0,
2
A)t
of the
form
w = P+OQ+
is
IR+IS
S =
w = P + vR
which together represent
It will
its
and
complete solution.
be noted that a fraction, in Boole s algebra, is always an am biguous function. Hence the division operation must never be performed: the value of a fraction is to be determined by the law of development only, except for the numerical coefficients, which are elsewhere discussed.
We
have already remarked that ax = bx does not give a = b, because .r may have the value 0. But we may transform ax = bx into a = bx/x and determine this fraction by the law
f(b,x)
/(!, l).fcr+/(l,0).&(l
*)+/&lt;)
!!
 bx
We
shall
then have
a= ==
and
this
is
bx
bx *
(1
~^
6.
coefficients, v
If
and w,
then
a
ax
bx,
x)
63
And
this result
is
always
valid.
rational
not give us
rational
married
but instead
married men + v married nonmen + w nonmarried nonmen our hypothesis is satisfied if the class "rational is, beings" consist of the married men together with any portion (which may be null) of the
rational

That
class
"married women"
class
"unmarried women".
If
we
consider Boole
system as an algebra of
and
1,
and the
fact that
for
any
fraction, x/y,
a*+ 6 *(l)+
we
shall find,
(1)
(l
*)(!)

(2)
=
x
1,
y
0,
0;
(3)
0,
0,
1;
and
that
it
(4)
Or, to speak
Since there are, in general, only four possible coefficients, 1, 0, 0/0, and such as do not obey the index law, of which 1/0 is a special case, this means
that any equation can be interpreted, and the difficulty due to the presence of an uninterpretable division operation in the has disappeared. system
"logical
symbol"
.r,
by trans
by subtraction
is
or division or both, and developing that side of the equation. Any equation may be put in the form in which one member
by
64
When
this is done,
expanded, all the coefficients which do not vanish may be changed to Boole calls this a "rule of unity, except such as already have that value.
102
interpretation".
(1)
Any
not
member
whose
coefficient does
vanish in development,
may
member
in
which each
coefficients
Negative
may
The
"
logical symbols" in
any function
x),
of the
form
(1
where
is
"logical
it
and
1.
Hence no constituent
altered.
And any
made
positive.
Xo new
effected
terms
will
of a developed function
always
null.
is
no consequence,
proposed the
"method
of indeterminate
each equation, after the first, is equations then added. But these indeterminate multipliers complicate the process of elimination, and the method is, as he afterward recognized, an
inferior one.
be reduced, by the methods just described, to the form in which one member is 0, and each coefficient is 1. They may then be simply added; the resulting equation will combine
simply, such equations
More
may
Any
logic.
"logical
symbol"
which
is is
may
be
eliminated by the
method which
all
students of symbolic
To
Ax + B(l The
result of elimination will be 104
AB =
102
L. of T., p. 90.
See Math. An. of Logic, pp. 7881; L. of T., pp. 115120. See L. of T., p. 101. We do not pause upon this or other matters which will be entirely clear to those who understand current theory.
104
103
65
By
reduced to a minimum.
classic algebra
Any
logical
Boole
may
also be got
by
solely
Two
s
which
principles are:
x,
2xy y). with coefficients other than unity, 2x may be replaced in the equation by x, and 2xy + x(l  y) by xy + x(l  y), which
algebra, the methods of reduction which have been discussed will always give x in place of x + x or of x + xy, in any equation in which these appear. The expansion of x + x gives 2x, the expansion of x + xy gives + x(l
x + xy These seem to be inconsistent with the Boolean meaning of + the first them does not hold for x = 1; the second does not hold for x = 1, y = But although they do not belong to Boole s system as an abstract
(2)
x +x
and
of
1.
By
the
method
for dealing
is
equal to x.
should
The methods of applying the algebra to the relations of logical classes now be sufficiently clear. The application to propositions is made
familiar device of correlating the
"
by the
times
logical
symbol",
x,
with the
when some
proposition,
is
true,
xy
X and
times
times
1"
when
when
of
A"
is
true
and
is false,
and so on.
x+y
the times
or
is
true.
or
we
xy + x(l
 y)+
and
(1
x)y.
the
"
universe",
will
times"
or
"always";
"A
will
be
"no
time"
or
==
"never",
will represent
is
is
always
true";
or
(i
.r)
1,
"A"
is
never true,
always
false".
Just as there
is,
classes, so there is
with Boole, no symbol for the inclusion relation of no symbol for the implication relation of propositions.
is
T"
For
classes,
"All
A"
or
"A"
is
contained in
Y"
becomes x
vy.
Cor
"If
respondingly,
"All
times
when
then
A"
"
or
"A
implies
F"
when
is
when Y is or = vy. x = y will mean, "The times when Y is true are the same" or implies
A"
is
true"
is
"A
Y and Y
The
follows:
105
implies
\
"secondary
propositions"
is
summarized as
L. of T., p. 178.
66
A
"Rule.
"Eliminate
Express symbolically the given propositions. separately from each equation in which
.
it
is
found the
indefinite
symbol
"Eliminate
it is
member
of
which
V =
0.
"Then
it is
desired
If in
not vanish.
2ndly.
If in
the
sum
of those constituents
If in
whose
coefficients vanish.
3rdly.
element, as x or
x,
the form of a conditional proposition having a com .r)(l pound expression, as xy, xy + (1 y), etc., for its antecedent,
4thly.
If in
t,
and determine
as a developed
. .
.
5thly. only be desired to ascertain whether a particular elementary proposition x is true or false, we must eliminate all the
If it
...
symbols but
tion
is
x\
true, x
that
it is false,
= 1 =
ficient to
determine whether
it is
true or
It is
and
The law
1
4=
1,
and
if
4= 0,
when
it
made
and
system completely inapplicable to logical classes. For propositions, this principle means, If x is not true, then x
"
is false,
if
is
not
false, it is
true".
But
interpre
tation for
"propositions"
shows that
system x
should be inter
67
is
"x
terpreted
"
true at
all times".
and x =
calls
propositions"
is,
are
that
statements which
may
call
",
upon what we now call "propositions" namely that they must be absolutely determinate, and hence simply true
limitation put
or false
sitional
under others.
The
system.
And
his
treatment of
"prepo
gives
them
in the application of the algebra to probability theory the character of prepositional functions" rather than of our
"
one hundred and seventyfive pages of the Laws of Thought are devoted to an application of the to the solution of problems in algebra 106 This application amounts to the probabilities. of a new
last
The
method
is
inyention involved in the problem logical analysis performed as automatically as the purely mathematical operations. a
We
can make this provisionally clear by a single illustration All the objects belonging to a certain collection are classified in three
:
ways
as
^4 s
or not, as
or not, and as
s
C
s
or not.
(2)
It is
then found
m/n
of the .4
are also
B
s
and
the
consist of the
s.
A
it
one of the
be taken at random,
By
premise
(2)
C = A(l  B)
Since A, B, and
Hence,
C are = A* (I  B) AC
"logical
+ .1(1
s
0.
The
that
s.
a given
is
also a
is
m/n, which is the required solution. In any problem concerning probabilities, there are usually two sorts of difficulties, the purely mathematical ones, and those involved in the logical
;
not a
or
by premise
(1 ), 1
The methods
106
of Boole
upon which the probability in question depends. algebra provide a means for expressing the relations
and then transforming these
logical
See also the Keith Prize essay "On the Application of the Theory of Probabilities to the Question of the Combination of Testimonies or Judgments", Trans. Roy. Soc. Edinburgh, xxi, 597 ff. Also a series of articles in Phil. Mag., 185154 (see An article on the related topic "Of Propositions Numerically Definite" appeared Bibl).
Chap. 16 ff.
posthumously;
396411.
68
equations so as to express the class which the quaesitum concerns as a func It thus affords a method for untangling tion of the other classes involved.
the problem
often the most difficult part of the solution.
The
Boole
is
striking.
the event
107
A".
Suppose x represent the class of cases (in a given total) in which the occurrence of occurs or those which "are favorable
to"
The
then
p be the probability, numerically expressed, of the event X. total class of cases will constitute the logical "universe", or 1; the
And
let
Thus,
if
1
is
if all
X
0.
p =
the probability of
X
=
"certainty".
0,
then p
X does
1.
not occur,
is
will
be expressed by
1
the numerical
p.
Also,
x+
x)
and p +
(1
This parallelism extends likewise to the combinations of two or more If x represent the class of cases in which events. occurs, and y the class
of cases in
which
which
x(l
occurs without Y;
(1
which
occurs without
y~)
X;
x)(l
y),
the
cases in
X or
and
which neither occurs; x(l occurs but not both, and so on.
events,
+ y(l
Suppose that
X and
of
Y are
X,
"
"
simple
"independent"
and
let
p be the probability
q the
prob
ability of Y.
Then we have:
of events
s
Combination
expressed in Boole
algebra
xy
pq
p(l
(1 (1
 y) x(l  x)y (1
(1
 q)  q)p
+
(1
*)(! y)
p)(l q)
x(l
y) +
(1
x)y
p(l
q)
p)q
Etc. etc.
is
formulated:
107
108
asserting
thing.
108
Boole prefers to consider x as representing the times when a certain proposition, an occurrence, will be true. But this interpretation comes to exactly the same
L. ofT., p. 258.
69
p, q,
r,
events, x, y,
z,
are the respective probabilities of unconditioned simple the probability of any event V will be [V], compound
,
formed by changing,
.
.
symbols
into p,
q, r,
"According
the event
V occur,
to the wellknown law of Pascal, the probability that if the event will occur with it, is expressed by a fraction
whose numerator
the probability of the joint occurrence of V and V, and whose denominator is the probability of the occurrence of V. We can then extend the rule just given to such cases:
is
"The
probability that
if
the event
occur,
will
[V

V]

where [V
V]
in the
into p,
q, r,
."
The inverse problem of finding the absolute probability of an event when its probability upon a given condition is known can also be solved. Given The probabilities of simple events x, y, z, are respectively when a certain condition V is satisfied. r, p, q,
:
To determine:
/,
m,
n,
of x, y,
z,
By
U f]
&gt;
[yV]
[zV]
W\
&gt;
And
I,
the
7i,
m,
number of such equations will be equal to the number of unknowns, ... to be determined. 109 The determination of any logical expres
form
sion of the
xV
is is
any
developed function
as a factor.
if
the
sum
of those constituents of
V which
*)(!
contain x
For example:
xyz + x(l
V =
xV = xyz + x(l y)z = xyz+(l x)y(l  z) yV zV = xyz + x(l  y)z + (1 Thus any equation
of the
y)z + (1
x)y(l
z)
(1
y)z,
.r)(l
y)z
form
109
On
xxm
ibid.,
certain difficulties in this connection, and their solution, see Cayley, "On a Probability" (with discussion by Boole), Phil. Mag., Ser. iv,
(1862), 35265,
and Boole,
"On
Probabilities",
xxv
(1863), 31317.
70
is
example
"
Suppose that in the drawings of balls from an urn, attention had only been paid to those cases in which the balls drawn were either of a particular
white/ or of a particular composition, marble/ or were marked by both of these characters, no record having been kept of those cases in which
color,
which was neither white nor of marble had been drawn. Let it then have been found, that whenever the supposed condition was satisfied, there
ball
was a probability p that a white ball would be drawn, and a probability q that a marble ball would be drawn: and from these data alone let it be
required to find the probability
ence at
all
that in the next drawing, without refer to the condition above mentioned, a white ball will be drawn;
also a probability
"Here
n that a marble
ball will
be drawn.
if
ball,
the condition
x represent the drawing of a white ball, y that of a marble F will be represented by the logical function
xy + x(l
y) + (1
x)y
Hence we have
xV = yV =
Whence
[xV]
xy + x(l y) + (1  x)y xy
= =
or
m,
[yV]
= n
and the
final
m
mn + m(l
n) + (1
m)n
m)n
+ =p
= P
=
q
mn + m(l
from which we find
n) + (1
m =
p+
q
q
... To meet a possible objection, I here remark that the above reasoning does not require that the drawings of a white and a marble ball should be independent, in virtue of the physical constitution of the balls.
"In
any system
any event
X may
+
form
x
= A+OB
^C
ID
by a change
of letters.
110
L. of T., p. 262.
have
71
where A, B, C, and D are functions of the symbols of the other events. As has already been shown, this is the general type of the logical equation, and its interpretation is given by
= A D =
x
+ vC,
where
v is arbitrary
and
By
we have
A+B+C+D
and, since
=
1
D =
0,
A+B + C
A+B+ C
abilities
universe
given in the data are to be interpreted as conditioned by A + B + C = 1, since = is the condition of the solution x = A + vC. If the given probability of some event S is p, of T is q, etc., then the supposed absolute
probabilities of S, T, etc.,
may
been described.
Let
V = A+B+C,
[sV]
then
W]
where
[sV],
[tV],
etc.
=P,
M
[IV]
?,
probabilities"
are the
"absolute
sought.
These,
being determined,
may
w =
[A + vC]
\yr~
which
"The
term vC
will
appear only
in cases
insufficient
Where
v
may
be determined by giving
and
1.
Thus
Lower
limit of Prob.
w =
\A] 
[A +
Upper
limit
cr
method, and with the theoretical difficulties of its application and interpretation, we need not here concern ourselves. Suffice it to say that, with certain modifications, it is an entirely workable
detail of this
With the
marked advantages over those more matter of surprise that this immediately useful application of symbolic logic has been so generally overlooked.
method and seems
to possess certain
generally in use.
It is a
72
JEVONS
s
"calculus
It
of
logic"
is
not so
of
much
the numbers
and
1,
some
whose ex
and some
of
of
fact,
and
and unnaturalness
of
some parts
s
of this
interpretation.
way
algebra, dis
and laying down a procedure which is just as general and is, in important In his first book on this subject, Jevons says: l12 respects, superior.
"So
of giving results
had
in this fact
an impregnable stronghold. Those who were not prepared to draw the same inferences in some other manner could not quarrel with the manner
of Professor Boole.
But
if
it
chapters
is
of equal
be true that the system of the foregoing s system, the case is altered.
of
There are now two systems of notation, giving the same formal results, one which gives them with selfevident force and meaning, the other by dark
processes.
and symbolic
The burden
of proof
is
shifted,
and
it
must be
in
it is
some
way
system."
He sums up
follows:
"1.
s,
as
113
Even process is of selfevident nature and force, and governed by laws as simple and primary as those of Euclid s axioms.
"2.
The The
process
is infallible,
alous results.
"3.
inferences
may
than in Professor
Boole
development
111
each
inference."
William Stanley Jevons (18351882), B.A., M.A. (London), logician and economist; professor of logic and mental and moral philosophy and Cobden professor of political economy in Owens College, Manchester, 186676; professor of political Uni
economy,
Pure Logic, or
Ibid., p. 74.
from Quantity,
p. 75.
73
The
unduly
his
third
of these
observations
is
Jevons
and methods
own procedure
be expeditious.
often cumbersome and tedious where Boole s would Yet the honor of first pointing out the simplifications
in the algebra of logic belongs to
He
the
discards Boole
of a
inverse operations, a
"
and
a/6,
and he interprets
sum
and
b as
either a or
shall
6,
where a and
exclusive
classes".
(We
114
A+B
is
or
J^.)
symbolize this relation by a + b: Jevons has Thus, for Jevons, a + a = a, whereas for Boole a + a
if it
be taken as
an expression
so that a + a
and
1, it
2a.
As has been
and
indicated, this
a source of
much awk
ward procedure
in Boole s system.
The law a + a = a
eliminates numerical
1, and this is a most important simplification. Jevons supposes that the fundamental difference between himself and Boole is that Boole s system, being mathematical, is a calculus of things taken in their logical extension, while his own system, being "pure
coefficients, other
than
logic",
is
It
is
it is
Since Jevons
we
The fundamental
(1)
a b denotes that
which
is
both a and
b,
sum
of the
meanings of the two terms combined. (2) a + b denotes that which is either a or
(in intension) a
as
b.
b
116
b,
symbolized
in
system by
6.
(o)
"excluded
114 115
indicates that which is contradictory or According to Jevons, from thought". He prefers it to appear as a factor rather than
\
A +B
it is
But
hardly correct; it makes no difference if it is known which, since the meaning of a + 6 does not depend on the state of our knowledge. Perhaps a better qualification would be "but it is not specified which". 113 Jevons uses capital roman letters for positive terms and the corresponding small
this is
italics for their negatives.
Following
De Morgan,
he
calls
and a
"contrary"
terms.
74
117
A
itself.
by
intension.
a proper interpretation of the symbol in Its meaning in extension is the nullclass or "nothing", as with
Boole.
Jevons does not use any symbol for the "universe", but writes out the This "logical alphabet", for any number n of ele "logical alphabet".
ments,
a, 6, c,
. . .
system, form
alphabet"
consists of a
z,
terms, x  Zf x
y, z, it
consists of x y
x y
z,
two elements, a and 6, Thus, ab, ab, and a b. For three x y z, x y z, x y z, x y z, and
for
Jevons usually writes these in a column instead of adding them y and putting the sum == 1. Thus the absence of 1 from his system is simply
a
real difference
from Boole
procedure.
of the
}
system
(8) (9)
If a = b and b = c then a = c. a b = b a. a a = a. a a = 0. a + b = b + a. a + a = a. = a. This law is made use of but is not stated. a+ = a b + a c and (a + b)(c + d) =ac + ad + bc + bd. a(b + c) = a. This law, since called the of absorption", a+ a b
"law
allows
is
Its
analogue for
a(a+
follows from
(8), (3),
b)
a
of absorption extends to
and
(9).
The law
also
.
any
number
of terms, so that
we have
a+
(10)
ab + ac + ab d+
This
is
=
a,
any term,
it
etc.
gives us
a(b + b)
is
(c
+ c)
= abc
This expansion
which appears
Boole
system, except
But the product any two terms of such an expansion will always have a factor of the form a a, and hence, by (4), Thus the terms of any expansion will always represent classes will be null.
for the different meaning of +
of
117
75
This accounts for the fact that, in spite of developed functions in Boole s system and in
Jevons
(11)
The
"logical
alphabet" is
made up
this
.
of
its
negative,
a + a.
It follows
immediately from
and law
.
.
a, b, c,
be
and
will
the expansion of
its
in Boole s
system because
it is
It
procedure by which Jevons sets great store is the "substitution of similars of a for b or b for a when a = b. Not only is this procedure valid
",
when the
expressions in which a and b occur belong to the system, but it holds good whatever the rational complex in which a and b stand. He
first
principle of reasoning,
et nullo. 11 *
is undoubtedly correct, another principle, which underlies Aristotle s dictum, which equally fundamental the substitution for variables of values of these is
dictum de omni
In this he
variables.
And
this procedure
is
equivalents.
= ];
No
is
is
a
c a
Some
"U"
is b:
"
= = =
a b a b
c
ab
or
Ua = Va
used to suggest
of
Unknown".
The methods
than Boole
s,
working with
in
are not.
and
"
extrinsic".
Intrinsic eliminations
member
get
of
Thus from a =
a
b c d,
we
a c d
= ab
a c
a d
a b
b a,
and
if
b,
ac =
If
be.
For example
aa
118
ad.
76
Also, in cases
form
a a,
is
involved, eliminations
0.
is
may
and a a =
which
Extrinsic elimination
that simplification or
or
"solution"
may
occur
when two
multiply such equations but uses them as a basis for striking out terms in the same "logical alphabet".
This method
is
first
forming the expansion of 1 (which contains the terms of the logical alphabet) and then putting any equations given in the form in which one member is and "subtracting" them from the expansion of 1. But Jevons did not
hit
His own
is
described thus:
119
Any
premises
term
into
involved therein.
its
of this
[E. g.,
b,
and
are involved,
form the
"logical
alphabet"
of all the
(a
"2.
Combine
term, as a be,]
When
tradiction [an expression having a factor of the form (a a)] with neither side of a premise, call it an included subject of the premise; when it forms a
it
an excluded subject
call it
of the premise;
when
it
a contradictory
com
bination or subject,
"We
and
strike it out.
an included or excluded subject a possible subject as distinguished from a contradictory combination or impossible subject. Perform the same process with each premise. Then a combination
may
call
"3.
is
an included subject of a
series of premises,
of of
any one;
any one;
it is it
a contradictory subject
is
an excluded subject
every premise.
consists of
any term [as a or b] involved in the premises the included and excluded subjects containing the term, treated as alternatives [in the relation + ].
"4.
The
expression of
all
"5.
Such expressions
may
be simplified by reducing
all
dual terms
[of
119
77
all
6. When it is observed that the expression of a term contains a com bination which would not occur in the expression of any contrary of that term, we may eliminate the part of the combination common to the term
and
in
its
"7.
expression.
Unless each term of the premises and the contrary of each appear one or other of the possible subjects, the premises must be deemed in
consistent or contradictory.
possible subjects.
least
two
Required by the above process the inferences of the premise a = b c. possible combinations of the terms a, b, c, and their contraries are as given [in the column at the left, which is, for this case, the logical
"
"The
alphabet
].
Each
of these being
we have the
following results:
abc
abc
ab
c
ale
abc
ab
c
= ab = ab
abc
included subject
contradiction contradiction
c c
ab
c
a b c
ab
c
= abbc = = a b b c c = =
a
b c
a b c a b c
contradiction
abc
abc
a be
a b c
"It
= = = =
a a b c
a a b c
abc
=
contradiction
a a b c
= ab c c = abb c
= ab
b
c c
a
b c
c
excluded subject
excluded subject
=0
=
a b
a a b c
a b c excluded subject
appears, then, that the four combinations abc to abc are to be struck out, and only the rest retained as possible subjects. Suppose we now require an expression for the term b as inferred from the premise
b c.
6,
namely a b
"Then
and a b
c.
b
= a
c
b
+ a b c, but as a
not with
6,
its
contrary,
we may, by Rule
0,
ab
hence b
a
+ a b
c."
This method resembles nothing so much as solution by means of the Venn diagrams (to be explained in Chapter III). The "logical alphabet"
is
compartments in such a diagram those marked "contradiction" are the ones which would be struck out in the diagram by transforming the equations given into the form in which one member is 0.
a
list
of the different
78
for his method, apart from its obvious namely, that the solutions for different terms do not require to be
is
separately performed,
all
also
an advantage
of the diagram,
which exhibits
of Boole,
alphabet"
this method of Jevons and also that be found that the comparison is as follows: The logical consists of the terms which when added give 1, or the universe.
"
"
"
contradiction
will,
by Boole
will
or 1/0;
"included subject"
ficient 1;
any marked
is
"excluded subject"
have the
coefficient 0/0, or v
where
arbitrary.
If,
then,
we remember
and thus eliminated, we two methods give substantially the same results. The single important difference is in Boole s favor: the method of Jevons does not distinguish decisively between the coefficients 1 and v. If, for example,
see that the
x y
z,
Boole
y
One
may
be mentioned
12
:
Any
tradictory subject, after combination with any unrelated terms. This means simply that, in any problem, the value of a term remains its value as a factor when the term is multiplied by any new terms which may be
introduced into the problem. In a problem involving a, 6, and c, let a be be a "contradictory" term. Then if x be introduced, abcx and a b c x will be "contradictory".
On
methods are
likely to be tedious
and have
little
of mathematical nicety
about them.
The
will
"logical
alphabet"
will
have to be investigated
much
remained for
others,
notably Mrs. LaddFranklin and Schroder, to accept Jevons s amended meaning of addition and its attendant advantages, yet retain
s
Boole
methods
solution.
development and similar methods of elimination and But Jevons should have credit for first noting the main clue to
of
this simplification
120
the laws a + a
a and a + a
a.
Pure Logic,
p. 48.
79
PEIRCE
contributions of C. S. Peirce 121 to symbolic logic are more numerous and varied than those of any other writer at least in the nineteenth
century.
The
Boole
He understood how to profit by the work of his predecessors, and De Morgan, and built upon their foundations, and he anticipated
when he did not
himself.
recent developments in
Again and again, one finds the clue to the most the writings of Peirce. These contributions may
(1)
summed up under
three heads:
He improved
more
characteristic of logical
more
and
division in Boole).
The
methods of develop ment, transformation, elimination, and solution, and certain advantages over the algebra of Boole because it distinguishes those operations and
retains the mathematical
relations
Also Peirce
into
introduced the
relation,
"is
contained
or
"implies",
Following the researches of De Morgan, he marked advance in the treatment of relations and relative terms.
symbolic
logic.
(2)
made The
method
is
made more
precise
and
"mathematical",
and the laws which govern them are related to those classes. Also the method of treating "some" and
of Boole s algebra of
"all"
propositions as
sums
(2) and products (II) respectively of "propositions" containing This is the historic origin of "formal variables was here first introduced.
implication"
and
all
it
in the
ment
He worked
out in detail the derivation of various multiple algebras from the calculus of relatives, and he improved Boole s method of applying symbolic logic to
problems
121
in probability.
Charles Saylnders Peirce (18391914), son of Benjamin Peirce, the celebrated mathematician, A.B. (Harvard, 1859), B.S. (Harvard, 1863), lecturer in logic at Johns Hopkins, ^596 ?. For a number of years, Peirce was engaged in statistical researches for the U. S. Coast Survey, and was at one time head of the Department of Weights and Measures. His writings cover a wide variety of topics in the history of science, meta According to William James, his physics, mathematics, astronomy, and chemistry.
articles
on
"Some
Logic",
source of pragmatism.
80
We
shall take
up these contributions
named.
mainly in the 122 and in Logic",
The improvement
brief article,
"On
set forth
an Improvement in Boole
the Algebra
.of Logic".
123
Calculus of
two papers,
It will
"On
tions
made up
[
two mutually
if
exclusive classes,
a and
is
] t
so that
x+b
a,
then x
b;
denotes the class of those things which are common to a and 6; and division is the strict inverse of multiplication, so that if x b = a, then
b or a b
ax
afb.
[ ]
and
These relations are not homogeneous in type. Boole s [+] have properties which approximate closely those of arithmetical
If [n]x indicate
the
number
of
members
of the
[n]a+ [n]b
[n](a + b)
is
because a and b are mutually exclusive classes, and every a member of (a + b) and every member of b is a member of
member
of a
a and b
larly,
This (a + b). from arithmetical addition only by the fact that are not necessarily to be regarded as numbers or Simi quantities.
[n]a
[n]b
s
[n](a
b or
b,
b)
But
Boole
a x
[n]ax
will
[n]b
[n] (a b)
and
this relation
is
is
its
arith
metical counterpart.
And
the same
true of
inverse, a/6.
Thus, in
Boole
calculus, addition
as arithmetical addition
are different in
and subtraction are relations of the same type and subtraction; but multiplication and division from arithmetical multiplication and division. type
by completing both sets of these adding multiplication and division of the arithmetical type, and
124 addition and subtraction of the nonarithmetical type. character of these relations is as follows
:
The
general
122
"Boole
Calculus
123
will
Amer. Jour. Math., in (1880), 1557, and vn These two papers (1885), 180202. be referred to hereafter as Alg. Log. 1880, and Alg. Log. 1885, respectively.
124
"Boole
s Calculus,"
pp. 25054.
81
The
"
a+
or 6
or
both.
125
(2)
The
\b.
\b,
is
such that
if
x+b
a,
then
Since x and
6,
in
x+
b,
an ambiguous function.
a
\
classes, a h b is
Then
x,
and
a h b
a
b
i.
Thus a
isx.
h b has
\
an upper
b coincides
limit, a.
=
e.,
a and no b
Then a
with a
x,
which
a
\
is
not a
6)
f
and
6,
b
it)
Thus a
a b.
f
or (as
we elsewhere symbolize
all
And
any
case, a
(
b is
b is a, the
We may
summarize
[0]
all
these facts
by
a h b
a 6 + v a b +
[0]
b
it is
where
v is
undetermined, and
null.
prefixed
(3)
must be
Boole
a
and
s.
This
is
(4)
The
is
126
b x
o,
then x
a/6.
This
is
Boole
a/b
an ambiguous function.
Its
upper limit
is
a + b;
i.
its
lower
limit, a.
e.,
a/b
a b + v a b +
[0]
a b
B.
(5)
The
"Arithmetical"
Relations
or 6
s,
where a and
Boole
a +
6.
a 6 + a 6 +
[0]
a 6
"a
(6)
is
The
6".
which
not
(7)
125
a
i
X
rce
pe
As has been mentioned, it coincides with the lower limit of a [6. 6 and a f 6 are strictly analogous to the corresponding relations indicates the logical relations by putting a comma underneath the sign of
is
both a and
b is a, b.
:
Peirce indicates the upper limit by a only in the paper Boole s Calculus".
6,
5
6.
These occur
82
of arithmetic.
"
logical"
They have no such connection with the corresponding b. relations as do a + b and a Peirce does not use them
relations, the following familiar laws are stated:
For the
"logical"
a+ a
= =
a
b
c
a a
a
b a
a+ b
(a +
+ a
ab =
a+
(b
b)
c)
(a b)c
= =
a(b
(a
c)
(a + b)c
= ac
+ bc
a b+ c
c) (b
c)
The
last
Peirce
two are derived from those which precede. s discussion of transformations and solutions
in this
system
is
inadequate.
Any sufficient account would carry us quite beyond what he has given or suggested, and require our report to be longer than the We shall be content to suggest ways in which the methods original paper.
As has been pointed
out,
if
any function
=/(!) x +/(0).z,
y}
=
&lt;p(l,
l)xy +
&lt;p(l,
Q)xy +
^(0, !)
xy +
&lt;p(Q,
0)
xy,
Etc., etc.,
the terms on the righthand side of these equations will always represent mutually exclusive classes. That is to say, the difference between the
"logical"
relation,
+, and the
"arithmetical"
relation,
+, here
vanishes.
Thus any
it
can be interpreted by developing to the above laws, provided that we can interpret these rela according
s
tions
in the coefficients.
And
Developing the
x+y
"logical"
sum, x +
y,
we have,
0)
(l
+ l)xy+(l +
Q)xy+(0+l) zy+(0 +
and (0+0) =
\b,
xy
(1
Comparing
=
this
==
1)
1,
(1
+ 0)
(0+1)
==
1,
0.
Developing the
x
\y
"logical"
difference, a
we have
(0 f0)
.
(1
x 
y
is
Comparing
127
this
above,
0;
we
equivalent to the
"Boole
undetermined
pp. 25053.
coefficient
that
hO)
1;
that
s Calculus,"
83
it is
equivalent to
null,
[0],
prefixed
must be
and that
0.
The
interpretation of the
"arithmetical"
relations,
and ^,
in coef
not to be attempted. These are of service only in probability theory, where the related symbols are numerical in their
significance.
ficients of classsymbols is
is
too complicated
to be entirely satisfactory.
Relatives", all
In the
"
logical"
in"
relations, a + b
and a
b.
The
relation of
"inclusion
or
"being
as small
as"
(which we shall
c) appears for the first time in the "Description of a Notation for the Logic of Relatives". 129 Aside from its treatment of relative terms and the use of the "arithmetical" relations, this monograph
symbolize by
gives the laws of the logic of classes almost identically as they stand in the
128
The
If
cy and
b,
(2) If (3)
=
a.
b.
If
then there
a,
(c
is
b.
(4) If b
(5)
=
6,
then a c
+ a) c
(c
If
a c
b)
(6)
(7) (8)
(9)
If
If
a cb,
a
c
a cb, a
cc b. cb c.
abca.
xc(x
y).
(y
z).
= xy +
x.
xz.
(14)
(xy)z
x(yz).
(15) x x
(16) (17)
128
= x. x x = O. 131 x + x = 1.
is
&lt;
Peirce s symbol
&lt;
pler to write.
129 130
Memoirs of
the
"Description
Amer. Acad., n. s., ix (1867), 31778. of a Notation for the Logic of Relatives," loc.
is
cit.,
342.
131
1 symbolized by n
"different
from every
x,
or
by
z
&lt;r~
84
(18) z +
(19)
A
= = = x.
1.
x+
(20)
(21)
&lt;p(x)
?(1)
[*(!)
*+*&gt;(())
a.
&lt;K.r)
(22)
If
&lt;p(x)
(23) If
= =
0,
1,
&lt;p(x)
equation of condition and the elimination re Boole had stated (22), which sultant for equations with one member 1. is the corresponding law for equations with one member 0, but not (23).
The
Most
by
of the
(9),
either
Jevons.
by Boole or c appears
notably,
x c y, then y c x.
And
relation. 132
In this
is
a;
is
true,
is
true,"
but he
well
aware
of
"a;
of the difference
cy and
is
usual significance
implies
is
y".
He
"It
means
if a: is
true, y
true
But
this
meaning
is
of things
referred
to.
... Now
is
true, y
is
true
is
that
it
goes
out beyond the actual state of things and declares what would happen were The utility of this is that it puts things other than they are or may be.
us in possession of a rule, say that if A we afterward learn something of which
is
true,
is
true
we
are
now
ignorant,
namely that
is
true, then,
by virtue
of this rule,
we
we know something
.
.
else,
namely, that
true
if
is
true.
if
the proposition,
is false.
ac
.
.
b, is
.
is
false or
b is true,
but
is false if
is
true while b
For example, we shall see that from (x cy) [the negation of x cy] we can infer z ex. This does not mean that because in the actual state of
things x
is
is
false or x true
but
it
does
mean
we
find x
true and y
false, in
is
or x
is
true [since,
ex hypothesi, x
132
true
133
2. Alg. Log. 1880, see esp. Alg. Log. 1885, pp. 18687.
85
"
cy,
material
implication,"
them.
They
will
be intelligible
is
the reader
"The
not one in
Peirce gives a number remember that x c y means, which x is true and y false".
"A
known.
(1)
arc (yea:).
This
is
true proposition is
implied by any
(2)
proposition".
If
"x
implies
y"
implies that x
is
true,
then x
is
true.
(3) [(a: cy) ca] ex, where a is used in such a sense that means that from x cy every proposition follows.
(ar
cy) c a
The
meaning
time. 134
difference between
of
"implies"
"material
implication"
is
But
it
is
interesting to
its
the
relation, understood
limitations as
have not.
c x.
[xc(ycz)] c [yc(xcz)].
(xcz)]..
This
is
the
"Principle
Syllogism".
Peirce
logic of relatives.
His interest
To
follow
on
this topic
would probably
shall
result in
complete con
Instead,
Peirce:
we
(1)
make
subject as
treated
by
De
Morgan
calculus of relatives
"mathe
matical"
symbolism for the most part contained in the early paper, "Description of a Notation for the Logic of Relatives"; (2) the calculus
and
in a
form which
makes
it
may
be seen at
best in
"The
Logic of
Relatives",
Note B
in the Studies
in Logic by members of Johns Hopkins University; consideration of the theory of relatives, which
papers, but has almost complete continuity.
134
and
is
But
i,
and Chap,
v, Sect. v.
86
A
The terms
may
"lover,"
names, they obey may be taken for granted without further discussion. But relative terms have additional properties which do not belong to nonrelatives; and it is to these that our attention must be given. If w signifies "woman" and s,
all
"servant,"
logic
is
"logical
product"
concerned not only with such relations as "servant woman", s + w, the sum" "logical
both)
",
iv,
the
"either
servant or
in the
and
relations
first
is
contained
in general of a
symbolized
"
by De Morgan,
of
U5
"servant
of every
woman,
and
"servant
none but
This
is
women".
We may
represent
"servant
of a woman"
by s\w.
a kind of
"multiplication"
relation.
It is associative,
s\(l\w)
"Servant
=
is
(s\l)\w
"
of a loverofawoman
"
servantofalover of a
woman".
Also,
it is
"addition"
symbol
ized
by +
s\(m + w)
"Servant
= s\m + s w
a
"
of either a
".
man
or a
woman" is "servant of
:
man
or servant of
"
a
is
woman
But
it is
not commutative
"lover
s\lis not
l\s,
servant of a lover
not equivalent to
of a
servant".
To
distinguish s
from
s w,
or s x
call s

w w
who
s
and women
w
,
we
shall
and w.
proposed
s
For
and
for
"servant
of none but
women"
w.
As we
by
as
W W W
ly
We may
and the
represent individual
class of all the
,
members
2,
3,
etc.,
Ws
as
W +W
l
+ ....
Remembering
the interpretation of +
we may
write
w =
and
that
this
is,
W W
1
W, +
means,
"The
classterm, w, denotes
or
or
s.
or
...,"
denotes an unspecified
member
of the class of PF
The servant
of a (some, any)
woman
is,
then, s w.
sw
"A
woman" is either
135
or
W
is s
or
3,
etc.;
s,
"servant
of a woman"
is
either
w, he uses
87
or servant of
Similarly,
z
"servant
and servant of
,
and servant of
z,
n represents the relative product, and x s\ n of the nonrelative logical product translated by "and". The represents above can be more briefly symbolized, following the obvious mathematical
"s
,"
where, of course,
analogies,
w = 2 W s\w = 2 w (s W) w = H s w (s\W)

Unless
we
shall
have
or
jc*w
"servants
The
class
"servants
of every woman"
is
of a woman".
lover of a servant of
all
women
is
a lover of a servant of a
(/]*)
woman.
Kc
A
woman
in the rela
woman
in the
From
The
proof of this
theorem
...
. . .
is
as follows
a
or a
= abc = abc
+ab
c ...
+a
b c ... + ...,
+ P, where
is
the
sum
of the
remaining terms.
+
,
Whence,
if
represent any
Similarly,
+mQP + mOQ
+
rnQR
Now
be respectively
the
left side,
/(.Ci), f(x*),
fM,
etc.,
all
On
we have
88
A
also,
we have
^c*,
If
/
or
8cs
woman
in the
We
w w
c sw c
I
s
and,
If s
c w, then
The
first of
these means:
woman is also a servant of every woman. The second means are women, then a lover of every woman is also a lover of
These laws
are, of course, general.
If all
servants
We
=
I
have
also:
every servant.
(l\s)\w
(s\w)
The
or a
who
is
either a servant
woman
is
woman.
An
interesting law
s) w
+ 2 q (l w
&lt;*
x s9) + s w
One who
is
eitherloverorservant of every
woman,
women,
is
is
woman
or, for
some portion
q of the class
lover of every
q, or, finally, is
woman
servant
member
of
woman. Peirce also gives this law in a form which approximates even more closely the binomial theorem. The corresponding law for the
product
is
simpler,
(I
xs)
xs w
which
is
On
we have
(raOP) + (mQQ) +
(mOabc
where
...) +
(mOR)
..,
or
(mQabc...)+K
is
is
sum
But
(mOabc
?nQ[f(x
x f(x2 ) x /(z3 )
.],
which
mOH x f(x)
Hence [mOHf(x)]+K = U x [mOf(x)]. Hence m O n /Or) c n x [m Of(x)
]
.
it
of a
Notation",
p. 346).
"Description of
a Notation, p. 334.
89
One who
woman
woman,
is
Peirce introduces a fourth term, and summarizes in a diagram the inclu sion relations obtained by extending the formulae already given. 138 The
number
is
hundred eighty.
He
challenges
the
formulation of these by means of ordinary language and formal logic. An s of none but members of w, Peirce symbolizes by 8 w. He calls this
operation
"
"backward involution",
relatives",
and
w he
refers to
as
infinitesimal
difficult
mathe
The laws
l
gous to those of
sw
If s
If all
c w, then
c w
l
servants are
women, then a
If
/
is
lover of none
but women.
c s, then
wc w
l
If all
women
is
a lover of
w) =
("OH;
The
lovers
of
loversof
women.
i+*w
= w
l
Those who are eitherloversorservants of none but women are those who
are lovers of none but
women and
women.
(w xv)
= w
s
The servants
those
of
violinists are
who
women and
)&
linists.
&lt;"&gt;MJC
Whoever
is
loverofaservant of
none but
women
is
a loverofevery
women.
l\
wc
&lt;*&gt;w
lover of one
who
is
women
is
a loverofnonebut
women.
l
wc
(s
w)
138
139
348 jf.
90
A
to a
Whoever stands
of hers
is
woman
women.
by the laws:
who
woman
is
the same as
woman
i
none but
/
 HO
It
Lover
from
s
of
is
appears
= O) =
(I
l\s,
l\s,
Not Not
nonlover of a servant.
is
lover of a non
servant.
l
s
~(l\s}
I s ,
A
,
lover of none
but nonservants
is
one who
is
not loverofaservant, a nonlover of every servant. i s = = l~ s A nonlover of none but servants (/ s)
is
one
who
is
not a nonloverofanonservant, a lover of every nonservant. We have the further laws governing negatives: 141
the Description of a Notation for the Logic negatives are treated in a curious fashion. A symbol is
"On
used for
"different
s".
from
"
of s
is
represented by
s
,
"
differ
Converses are barely mentioned in this study. In the paper of 1880, converses and negatives appear in their usual notation, "relative addition" is brought in to balance "relative multiplication", and
the two kinds of involution are retained.
in the
But
in
"
The Logic
of Relatives
"
are retained.
final
form of Peirce
calculus of relatives.
We
have here,
(1)
Relative terms,
a, b,
... x, y, z.
(2)
140
141
The negative
ibid., p. 353.
of x, x.
See
Notz
is
here symbolized by
(1
z)
91
The converse
a;
of x, ^x.
If
is
"lover",
^.r is
"beloved";
if
KC
is
"lover",
is
"beloved".
(4)
Nonrelative addition, a +
b,
"either
a or
6".
(5)
(6) (7)
b, "both
a and
6".
s,
a of every non6
".
(8)
(9)
The The
relations
= and c
as before.
1,
universal relation,
"consistent
with,"
term with
(10)
itself
other.
1.
(11)
itself.
The The
relation
"identical
with",
7,
(12)
The
relation
"different
from",
142
is
distinct.
(2)
(3) (4) (5)
6).
(6)
(7)
(x a)6 x\(a\b) x t (a t b) = (x t a) t b
(8)
(9)
x\(atb)c(x (a t b) x c a t
a)
t6
(6 x)
z(at&)c(3ta)(a:t&)
(13)
(14)
(15) (a 6)
1
(16) (17)
(18)
(19)
by comparison.
92
A
(20)
Ocx
= x + 1 = 1 ztl = 1 1 t.T = 1 sttf = x Art* = x + x = 1
x+
.r
xcl
= = z0 z0 = 0\x =
x
(23) xl
(25)
(27)
(29)
(31)
z/cg
(33) /z
(34)
(36)
or
(35) (37)
= x x x =
fcr
7c[a;1XaO]
is,
(*)]
ctf
This calculus
ciples of solution
down. 143
Not only
the variety
symmetry between
relative multiplication
and
But, as
chief value of
any calculus
s
of relatives
is
not in
any elimination
made
directly
formulae.
Peirce
is
fore, of
much
importance than
the theoretic foundation upon which It is this which has proved useful in later
made
is
of a
in the
or
of individual things.
etc.,
If
twodimensional array,
A:A,A\B,A:e,A:D,
B
:
A,
B,
C,
D,
Etc., etc.
itself is
be noted that any individual thing coupled with relative but that in general A B differs from B A
It will
: :
an individual
individual relatives
A
143
general relative
"Logic
is
sum
of such
of
Relatives"
p. 193.
93
represent
b
"benefactor",
then
= Z,S
y (6) t.,.(/
,/),
where (&)/
factor of J,
is
is
in case / is a
bene
and otherwise
That
is
to say, b
is
the logical
is
sum
of
all
the
This
the
first
formulation
in
in
extension",
in logistic,
though seldom
By
It is
b
the coefficient 0.
some expression
= (X: y)!+(Z:
logical
Y) 2
+(X
,
) 3
+...
If,
"b
meaning
of +
we
may
be read,
either
(X
F), or
(X
F) 2 or
(X
F) 8 or ...
is,
To say
that 6 repre
b repre
it
then, inexact:
should represent
"some"
in a sense
precluded by the
any two
mutually
calls
genera)
variable".
mathematical
is
founded on suppo
the required condi
case, his
assumes a
tions.
esis
is
fulfilling
hypoth
must belong to every such case. The ad vantage of his procedure lies in the fact that the logical laws of individual terms are simpler than those which relate to general terms, because indi
individual case but those which
viduals are either identical or mutually exclusive, and cannot intersect or be subordinated to one another as classes can.
. .
.
"The
indi
mduum vagum. Julius Caesar is an example of the former; man of the latter. The indimduum vagum, in the days when
,
a certain
such con
difficulty
from
its
"Description
94
If
A
we
,
include under
any individual
is
man
what
true of
any
individual
man
is
true of
it
men.
Such a term
is in
it
represents every
man.
is
But
represents each
man
not
.
itself
. .
terms.
so, though an individual term, it stands for any one of a class of such The letters which the mathematician uses (whether in algebra
individual; and
by second
intention.
All the
formal logical laws relating to individuals will hold good of such individuals
by second
intention,
may
be
The
It
is
relative
b,
denoting ambiguously any one of the benefactorbeneis such an individual by second intention.
of the
:
"
defined
by means
prepositional
function",
"I
"I
benefits
J"
J",
as the logical
sum
of the (7
J) couples for
which
benefits
is
true.
The compound
If
S,Sy(a)iy(Z
J),
and
SiSy(&)&lt;y(/
J),
then a +
S,Sy[(a),y +
(&),,](/
J)
:
That
"7
is,
if
"agent"
is
the logical
sum
of all the (7
is
is
agent of
J"
is
true,
and
"benefactor"
the
sum
"
J)
couples for
is
which
"7
benefits
J"
is
true, then
"either
agent or
benefactor"
is
the logical
sum
J"
of all the (7
is
Either 7
facts
agent of
or 7 benefits
true.
We
same
6) t,.
more simply
by
"prepositional function", (a
145
The
equation.
The
compound
(ax6)
or
"Both
(a) tyx(6) iy
:
a xb
J)
:
agent and
"7
benefactor" is
the logical
sum
J"
of the (7
is
J) couples for
which
is
agent of
J and
(a
I
is
benefactor of
true.
&)./
S*{(a)*x(6) Ay
&lt;A
or
145
a\b
",
S,Sy[S*{(a)
188.
(b) hi ]](I
J)
See
"Logic
of Relatives
loc. cit., p.
95
J) couples such
is
of a benefactor"
is
the logical
sum
//
is
of all the (7
that, for
some
//,
"I
is
agent of 77
and
benefactor of
J"
true.
difficulties in
"agent
last.
The
first
Peirce, like
De Morgan,
itself
or
is
"agent
of"
name
"agent".
Now
note
name denotes the first term in With this in mind, the compound
of a benefactor"
"
a\b, will
:
become
which
"Agent
names the 7
in the /
".
pairs
make up the field of the relation, agent of a benefactor of Any reference J s at the other end of the relation is gone, just as omits The second any reference to the J s in the field of the relation "agent concerns the operator, SA, which we have read, "For some difficulty
to the
"agent"
of".
H".
is
Consider any statement involving a "prepositional the variable representing the individual of which
function",
&lt;p
&lt;pz,
where
is
asserted.
That
is,
2z
3
&lt;pz
symbolizes
",
"Either
&lt;p
is
true of Zi or
&lt;p
is
true of
"For
or
z
(p
is
true of
or ...
and
this
is
some
(some
"/
2 or other),
&lt;pz".
is
(o)t*x(6) A y,
is
agent of // and //
the relation
"7
benefactor of
J".
The
is
agent of a benefactor of
is
J",
some
agent of
"
H and H
benefactor of J.
&lt;pz
Suppose we
ator n.
consider any
propositional
function",
Tl 2 (pz
=
&lt;f&gt;Zi
(pZ/z
&lt;pZs
x
&lt;p
...
is
That
of
3
is,
"
&lt;pz
symbolizes
",
is
&lt;p
true of Z\ and
z".
true of
and
&lt;p
is
true
Z and
or
b.
"
is
&lt;p
This operator
is
needed
in the
definition of a t
(ot&),/ n A {(a)
"
+(6) A
is
,}
"
J"
equivalent to
For every
77,
H ot6
J".
"Agent
of all nonbenefactors"
the logical
sum
of all the (7
is
J) couples
agent of
H or 77
benefactor of J.
The
of relatives
96
which are
first
applies
We
to be discussed, con
"prepositional
func
a t
b.
This step
The
of the converse
(6)./
is
(6)/ f
For example:
To
prove, v(a + b)
(a
^a + ^b
(a
&),/
&),,
(a) /t+
(6) yf
t
(&)/&lt;
But
(a)ji
= (a)
t /,
and
w o) (
(w6) tV
Hence
(a + &)/
:
/+
(W&gt;)v
of
Hence S&lt;S,Ka + 6),,} (7 J) S;Z y {(a) t7 + (&)/}(/ J) Q.E.D. For the complete development of this theory, there must be a discussion the laws which govern such expressions as (a) or in general, expressions
:
,,,
of the
is
form
(px,
where
&lt;px
is
(px
of the variable is
functions".
147
specified.
(a)*,
or in
propositional
is
&lt;px,
a propositional function of
one variable;
(a),/, y), may be regarded as a propositional function of two variables, or as a function of the single variable, the individual rela
tive (7
J), or
(X
Y).
is
On
the Algebra of
Logic".
It
is
4 he laws
The
time.
first
46 The formulae of the calculus of classes can also be derived from these, considered as themselves laws of the calculus of propositions (see below, Chap, vi, Sec. iv). 147 Peirce has no name for such expressions, though he discusses their properties acutely
2).
first
This assumption
97
of
this
law
is
"If
4= 0,
then x
is
1",
which gives
is
0"
"If
not
false,
then x
true,
Peirce uses v
is
instead of
and
0,
stated in
"true"
and
f)(vx)
=0
though derived from the
"x
of
propositional
is
functions,
it.
is
man" is
neither
A
"x
propositional function
a:
may
a
some
cases.
"If
is
a man, then x
is
mortal" is
or
true of any x;
of x.
is
man" is
some values
2x
"
&lt;px
represents
&lt;px
&lt;?x
is
2
true for
is
some value
&lt;px
of the variable, x
. .
."
that
is,
either
is
true or
&lt;px
true or
3 is
true or
Similarly,
TL x TL t
tf
&lt;px
=
&lt;px
(pxi
&lt;px
&lt;px
...
represents
true and
(a) xy ,
,
&lt;px
tpx
2
is
true for
all
is
that
is,
&lt;pxi
is
is
true and #r 3
If
or
more conveniently,
\f/(x,
&lt;p(x,
y), represent
"x
is
agent of
y",
y",
and
(b) xy or
more conveniently,
IL x
y),
mean
"x
is
benefactor of
then
2 y [&lt;p(x,y) *t(x,
y)]
y,
"x
will
mean
is
that for
all
is
agent of y
and x
benefactor of
is
true
that
is,
it
"Everyone is
if
someone".
This
will
appear
we expand n x 2 y
y) x^(.r, y)]:
2/0
2/j)
xiKzi,
,
2/i)]
+
+
[&lt;?Oi,
2/2)
x x
{[&lt;p(x
2,
x^(.r 2 yi)]
[&lt;f&gt;(x
z,
y z ) x^(z z y 2 )] +
,
..)
.
U&lt;/&gt;(&gt;3,
2/i)
x^(.r 3
2/i)]
[^(.r 3 , 2/2)
x^(.r 3 2/2)]+
,
{Either [xi
is
agent of y 2 and Xi
a: 2
agent of
2/1
and
. .
is
benefactor of
{either [x s is
.r 3
y\] or [x 2
is
agent of y 2
is
.
and x 2
is
bene
factor of y 2 ] or
2/i]
.}
and
agent of y\ and x 3
]
.
.
benefactor of
.
or [# 3 is agent of y 2
8
and
is
benefactor of y z or
and
Etc.,
etc".
98
A
Z,
The operator
it,
which
is
as Peirce calls
II,
within the
lines.
i.
The
e.,
outside operator,
in the
the subscript of
columns; and the operator nearer the Boolian indicates the letter which
x, between the lines
line to line.
little
varies within the lines, the subscript of the outside operator, the letter
With a
may
learn to interpret
any such expression directly from the meaning of simple logical sums and For example, with the same meanings of logical products. y) and
&lt;p(x,
t(x, y),
lUZj^Cr,
will
y) xifr(y, x)]
(y)
mean
"Everyone
(x)
is
agent of some
benefactor of
himself".
And
"
will
symbolize
There
is
is
is
agent of z or z
benefactor of
is
some x and some y such that, for every z, either or, more simply, "There is some
y";
pair,
x and
y,
such that x
agent of
all
nonbenefactors of
y".
The laws
for the
:
n and 2
operators
The
different premises
(the same index not being used in two propositions) are written together, and all the n s and Z s are to be brought to the left. This can evidently be
done, for
[Or in the
familiar, notation,
Xll y
&lt;py
= U xU y
(&lt;px
x
&lt;py)
x
&lt;py)]
"2d.
of the indices of
the
n
149
may
be
moved
relatively
left
to one another,
and as
pp.
19698.
99
We
have
n,ny.T ty
S.SyZiy
= n itey = SyZtf.y
;
[Or, [Or,
mix*,
S.S^fo
y)
y)
and
also
this
Silly&.yy
= n/Z^y
^) = n y S x (^r
But
and
We do
have, however,
SillyZiy
It will, therefore,
&lt;
IlyZtf.v
[Or,
ZJI^ar,
y)
c n,2 x ^(.r,
y)]
be well to begin by putting the 2 s to the left as far as because at a later stage of the work they can be carried to the possible,
right but not [always] to the left.
premises are
orders
II
For example, if the operators of two and 2 JI 2 z we can unite them in either of the two ^II*
tf
and
shall
often be
. . .
room
"oth.
usually obtain different conclusions accordingly. There will for skill in choosing the most suitable arrangement.
The next
in the Quantifier.
Thus,
for
Silly/,,
[Or, for
2JI 2JI
tf *&gt;(.r,
y),
we can write
"6th.
2 JI//^
tf (
&lt;p(x,
y) x ^(.r, x)}]
The next
adding to any part any term we like; 2d, from any part any factor we like, and 3d, in observing that
part, consisting, 1st, in
dropping
xx
so that
"7th.
= =
/,
x+x
(x
v,
xxy +
+ x + y)z
and S
in the Quantifier
in
fifth
and seventh.
Thus, from 2
!! ,/,/
we
shall at
once proceed to 2
/ lt if
we
like."
We may
&lt;pxi
in this
by treating 2 x v?.r
tpx\
&lt;^.r
as a sum,
&lt;^.r
&lt;.r
&lt;px
and
Il x
&lt;px
as
product,
x,
2 JI,,^(,r,
y) as a
100
Thus
this calculus
may
be derived from
But Peirce does not carry out any proofs and he notes that this method of proof
"It
is to be remarked that SjX, and would be theoretically unsound. 150 IliXi are only similar to a sum and a product; they are not strictly of that
may
be
innumerable."
Another way
of saying the
calculus of propositions
same thing would be this: The laws of the extend to 2iX and Ito, because the extension cannot

by the method which the mathemati suggest, would require the principle of
infinite.
The whole
that
may
of propositional functions
is,
by representing any
II
J),
relations, such as
"converse
"relativeproduct,"
We
of
of the elementary relatives. need not enter into the detail of this matter, since Sections II and III
will
and 2 functions
Chapter IV
modification of Peirce
develop the calculus of propositional functions by a s method, while Section IV of that chapter will show
will indicate the
how
functions, Section
tions
manner
s
in
of rela
may
VI
how, by a
mathematics
may
be obtained.
remains to consider briefly Peirce s studies toward the derivation of other mathematical relations, operations, and systems from symbolic logic.
in this connection,
is
"Upon
the Logic of
Boole
Calculus of
and
of the
Relatives",
The
+
,
firstmentioned of these
etc., of arithmetic
concerned to show
how
the relations
cal
can be defined in terms of the corresponding logi =, relations, and the properties of arithmetical relations deduced from
logical analogues.
152
...
calculus, in
which the
Alg. Log. 1885, p. 195. Proc. Amer. Acad., vn, 40212. Loc.
tit.,
152
pp. 41011.
101
no longer terms of
first
first
and
... Let
the letters
relate exclusively to
the extension of
things
intensions.
the letters signify only the differ In other words, the only logical comprehension which the letters considered as terms will have is the greater
ences of classes as wider or narrower.
or less divisibility of the class.
calculus
might, for
example, denote
all
it
New England
make
what
supposed,
neglected,
these states
what they
are, being
would
signify only
which
has the same relation to higher and lower classes which the class of
New
that
is,
a collection of
six.
if
In this case, the sign of identity will receive a special meaning. For, denotes what essentially belongs to a class of the rank of sides of a
,
cube
State
then [the
is
logical]
m=
will
New England
the side of a cube, and conversely, but that whatever essentially belongs to a class of the numerical rank of New England States essentially
and conversely.
Identity
may
be termed equality.
..."
etc. represent
a,
b,
c,
etc.,
then the arithmetical relations can be defined as logical relations. The logical relation a + 6, already defined, will represent arithmetical addition:
And from
will follow
is
+
is
is
commutative and
associative,
it
so also.
Arithmetical multiplication
153
more
a
difficult to deal
with but
may
be defined as follows:
b are events
b represents
only a
is
if
these events
is
which case a
X
3,
[where a b
the
is
By
two
meant that
B\,
it
possible to take
A A
2,
,
etc.,
and
B B
2,
3,
etc.,
satisfied.
, ,
A m A n Bm B n
(.4
two
series of terms,
and 2 A, 2 B, 2
B)
logical
of the
of the
A n s,
the
B n s,
and the (A n
BJ
respectively.)
Condition
1.
2. 3.
4.
153
No A m is A n No m is n x = 2 (.4 B) a = 2 A
Loc.
cit.,
p. 403.
102
Condition
= 2B
is
6.
Some A m
Bn
of the
This definition
a b
is
matter
is
that
tions of a
member
will
of a
member
of
b.
of a are
distinct (condition 1)
combinations
It is
b.
worthy
of
remark
that, in respect
of
upon the same fundamental ideas by means which arithmetical relations are defined in Principia Mathematical
"second
The
intention"
of a class
term
is,
in Principia,
Nc
b,
in
Peirce
sum"
discussion, corresponds to
what
is
"arithmetical
of
two
logical classes,
s
and a
b to
what
product".
But Peirce
meet
the difficulties
it
that
could hardly be expected in a short paper. In particular, define the arithmetical sum in case the classes summed have
does not
in
members
common, and
class,
it
it
though
does not indicate the manner of defining the number of a does suggest exactly the mode of attack adopted in Prin
cipia,
algebras from those of the logic of relatives is more comprehensive, though here it is only the order of the systems which is derived from the order of
is
We
individual in which
are here to take a closed system of elementary relatives, every is either a T or a P and none is both.
Let
c
s
= (T T) = (P:P)
:
p = (P:T)
t
= (T :P)
:
Suppose T here represent an Individual teacher, and P an individual pupil the system will then be comparable to a school in which every person is
either teacher or pupil,
pupil.
354 155
and none
is
The
relative
term, c, will
both and every teacher teaches every then be defined as the relation of one
103
P), the rela
is,
colleague".
is,
Similarly, s
is
(P
"
"schoolmate".
The
relative
term, p,
any pupil to any teacher, that is, pupil". And the relative term, t, is (T P), the relation of any teacher to any pupil, that is, "teacher". Thus from the two nonrelative terms, T and P, are generated the four elementary relatives, c, s, t, and p.
(P
:
:
The
if
we venture upon
certain
explanations of the properties of elementary relatives which Peirce does not give and to the form of which he might object. For any such relative
(7
:
J),
(1)
where the /
and the J
are distinct,
we
shall
(I:J)\J =
1
(7
:
J) relation to a
J must
must be a
teacher.
(7:./)l/
=
relation to a teacher (where teachers
and
(I
:J)\(II:K) =
[(I
J)\H]
K
(7
:
The
relation of those
:
the (77
K)
relation
is
J)relation
toan77 to a
K.
law which
is
It is this third
the system.
t\p
For example:
= (T :P)(P
of
fit
T)
s
[(T
P) P]
T = (T
s
1
:
The teachers
or his
any person
colleagues.
(Our
illustration, to
may
be his
own
colleague
own
schoolmate.)
c\c
= (T
= (T
T)\(T
s
T)
[(T
T)\ T]
s
T = (T
T)
The
colleagues of one
t
t
:
colleagues.
T]
:
P) (T
P)
[(T
P)
P = P =
(0
P)
= (P
T)
(P
P)
[(P
T) \P]
(0
P)
There are no pupils of anyone s schoolmates in the system. The results may be summarized in the following multiplication table,
in
in the
column at the
right
104
The symmetry
easily in
gives.
To
may
If f is
will
be a modifier
French teachers have only French pupils, and vice versa. Such modifiers are "scalars" of the system, and any expression of the form
a c +
bt+
p+d
s
a, b, c,
where
c,
t,
p,
and
s are
d are scalars,
Peirce calls a
is
"logical
quaternion".
The product
of a scalar with a
term
commutative,
bt = tb
since this relation
is
Inasmuch as
any (dyadic,
triads, etc.)
sum
of (pairs,
elementary relatives,
ever
is
If
is
of the
form
az +
where
i,
bj + ck + dl+
...
j, k,
I,
etc. are
mu + nv + oiv+
we
...
where m, n, o, etc. are scalars, and u, v, w, etc. are elementary relatives, shall have a more complex algebra. By such processes of complication, multiple algebras of various types can be generated. In fact, Peirce says: 157 can assert, upon reasonable inductive evidence, that all such
"I
[linear
on the principles of the present notation in the same way as those given above. In other words, all such algebras are complications and modifications of the algebra of (156) [for which the multiplication table has been It is very likely that this given].
155
157
36364.
105
true of
all
algebras whatever.
The algebra
of (156),
which
is
of such a
pure algebra and our logical nota tion, has been shown by Professor [Benjamin] Peirce to be the algebra of
in reference to
fundamental character
Hamilton
s quaternions."
and
of importance
is
is
Peirce
modification of Boole
calculus
This
in
set forth
s
an Improvement
identity of
Boole
with extreme brevity in the paper, "On Calculus of Logic". 158 For the expression of
we
shall
two
classes in extension
need to distinguish the logical relation of from the relation of numerical equality.
We
of
as the class
or
all
are
s,
by a
b,
and the
of
as the
number
members
by a +
b,
ab, a
b,
\b,
expressed by a
"Let
b,
b.
Peirce says:
159
every expression for a class have a second meaning, which is its meaning in a [numerical] equation. Namely, let it denote the proportion
of individuals of that class to be
in the long run.
"Then
found among
all
we have
If
"
as a class,
Let b a denote the frequency of the 6 s among the a s. Then considered if a and b are events b a denotes the fact that if a happens b happens.
X
"
ba
ab
It will
ties of
be convenient to set
.
the function b a
Xb =
a,
X
1
ab
C) a
&lt;p(b
Ca ) b) a
(1
= =
&lt;p(b,
ba
158 159
Proc.
Ibid., pp.
106
&
0(l_a)
The
abilities
(1)
between
prob
and the
in passing
equa
simply changes the relations involved from logical relations to the corresponding arithmetical relations, in accordance with the foregoing,
tions, Peirce
and
lets
the terms
a, b, etc.
s,
s,
etc.
system under discussion. Boole has no symbol for the frequency of the a s amongst the 6 s, which Peirce represents by a b As a result, Boole is led to treat the
in the
(2)
.
a pro
many
difficulties
and some
errors.
four
logical
operations,
and four
logical
Probably there is no one piece of work which would so immediately reward an investigator in symbolic logic as would the development of this calculus of probabilities in such shape as to make it simple and practicable.
Except
for a
the subject has lain almost untouched since Peirce wrote the above in 1867.
Peirce s contribution to our subject is the most considerable of any up to his time, with the doubtful exception of Boole s. His papers, however, are brief to the point of obscurity: results are given summarily with little
or no explanation
and only infrequent demonstrations. As a consequence, them make tremendously tough reading, and they
serves. 161
have never received onetenth the attention which their importance de If Peirce had been given to the pleasantly discursive style of
clearly accurate
manner
of Schroder, his
several volumes.
an unusually
logic.
161
Since the above was written, a paper by Couturat, posthumously published, gives clear presentation of the fundamental laws of probability in terms of symbolic
See Bibl.
Any who
work unduly
difficult or
107
VIII.
of Peirce
Contributions to symbolic logic which have been made since the time need be mentioned only briefly. These are all accessible and in a
sufficiently close to current notation to be readily intelligible.
form
Also,
they have not been superseded, as have most of the papers so far discussed; consequently they are worth studying quite apart from any relation to later work. And finally, much of the content and method of the most *
important of them
later chapters, or
will
is
is
substantially the
will
is
be set forth in
there set forth
such that
its
be pointed out.
But
and perspective, a
summary account may be given of these recent developments. We should first mention three important pieces of work contemporary
with Peirce
s later treatises. 162
containing
(3)
(1)
The
Begriffslehre
this
and as
would
indicate, the
development of
s
logic
is
entirely mathematical.
An important
character of Grassmann
procedure
is
laws of classes, or Begriffe, as he insists upon calling them, from the laws governing individuals. For example, the laws a + a = a and a a = a,
where a
where
it,
e,
is
e,
ee
e,
e\e&lt;i
0,
e\, e 2
represent individuals.
much
to
commend
but
it
as
an aggregate of individuals and the laws of such aggregates proved As Peirce has observed, this method generally by mathematical induction.
breaks
down when
the
number
of individuals
is
may
be
infinite.
Another
difference
language of intension.
and
illustrations in
which
in
"concepts"
162
are represented
by
Alexander MacFarlane, Principles of the Algebra of Logic, 1879, gives a masterly presentation of the Boolean algebra. There are some notable extensions of Boole s methods and one or two emendations, but in general it is the calculus of Boole unchanged. MacFarlane s paper "On a Calculus of Relationship" (Proc. Roy. Soc. Edin., x, 22432) re
sembles somewhat, in
its method, Peirce s treatment of "elementary relatives". But the development of it seems never to have been continued. 163 There are two editions, 1872 and 1890. The later is much expanded, but the plan and general character is the same.
108
calculus
is
Hugh MacColl
164
first
two papers on
paper
"On
"The
Calculus
of
Equivalent
165
in
printed 187880, present a calculus of propositions which has essentially the In others words, it is properties of Peirce s, without II and 2 operators.
Statements",
and
his first
Symbolical
Reasoning",
it
a calculus of propositions, like the TwoValued Algebra of Logic as we know today. And the date of these papers indicates that their content was
s
studies
this touic.
In fact, MacColl writes, in 1878, that he has not seen Boole. 166
The
calculus
167
set
forth
in
MacColl
book, Symbolic
Logic
and
its
Application*,
is
of
and instead
have
tain
77,
of the
two
"true"
and
"false",
we
"true",
"false",
"impossible"
and
"variable"
(not cer
8 respectively.
These are indicated by the exponents r, i, e, The result is a highly complex system, the fundamental
ideas
of Strict
The
"On
the
Algebra of
in the
168
differs
.
systems based on Boole by the use of the copula v or "Some a is classes, a v b represents ispartly
"a
Where a and
6",
6",
av
"
b,
represents
a iswhollynot6
b to
"
",
or
0.
No
is
".
lent to a b =h 0,
and a v
a b
cb
will
be a v6,
and a = b is represented by the pair, (a vb)(a v b). For propositions, a v b denotes that a and b are consistent a does not imply that b is false and b does not imply that a is false. And a v6 symbolizes and b are
"a
inconsistent"
if
is
true, b
is false;
if
6 is true, a is false.
The
is
use of the
terms
"consistent"
and
"inconsistent"
in this connection
possibly mis
leading:
164
166
(1) Proc. London Math. Soc., Mind, v (1880), 4560. Proc. London Math. Soc., ix,
920;
178.
Longmans, 1906. 38 The same volume contains an interesting and somewhat complicated system by O. H. Mitchell. Peirce acknowledged this paper as having shown us how to express uni versal and particular propositions as n and S functions. B. I. Oilman s study of relative number, also in that volume, belongs to the number of those papers which are important
in connecting symbolic logic with the theory of probabilities.
167
109
of
which
is
true
and
and
"
inconsistent
"
related to
what
is
usually
material implication a c b
a".
is
related to
meant by what
x
usually
meant by
"b
That a given
class, x, is
is false,
0,
may
be expressed by x v
1.
oo,
where
co
is
"everything"
represented by
That a
class, y,
has members,
is
where y is a proposition, since Mrs. system does not contain the assumption which is true for propositions but not for classes, usually x =(= 0, then expressed, x = 1, and if x == 1, then x = x v oo may be abbreviated to xv,
is
This last
of doubtful interpretation
s
LaddFranklin
"If
0".
a b v
to a b
etc., since it is
always
is
term
tion:
is
This convention leads to a very pretty and convenient opera v or v may be moved past its terms in either direction. Thus,
oo
.
(a
v 6) =
(a
by) =
va
vx
6)
and
(xvy) = (xyv) =
y)
But the forms (va6) and (vxy) are never used, being redundant both logically and psychologically.
Mrs. LaddFranklin
logic particularly well
:
All a
is b.
b,
or or
or or
a 6 a b a b
No
is b.
avb, avb,
6.
Some Some
Thus v
a a
is b.
is
not
a v 6,
a 6
characterizes a universal,
v a particular proposition.
"
And any
then
all
by the difference
between v and v
a
"
The
syllogism,
If all
is
and
all b is c,
is c,
will be represented
by v b) (b v c) v
(a
(a
v c)
is interpreted for classes, and v between the parentheses takes the propositional interpretation. This ex
may
also be read,
"
All a
is
and
is
all b is c
is
inconsistent with
Some
not
".
It is
equivalent to
b) (b
v c) (a v c) v
110
A
three propositions,
"The
All b
is c,
and
Some
is
not
are inconsistent
three syllogisms:
they cannot
all
three be
true".
(1)
"If
(a
v6)(6 vc) v
all
(a
vc)
all
is
and
all 6 is c,
then
(a
is
c";
(2)
"If
v b) (a v c) v
c,
(b
v c)
not
c";
all
is
and some a
(3)
is
not
(b
then some
b is
v c) (a v c) v
c,
(a
v 6)
is
"
If all b is c
and some a
is
not
then some a
not b
".
Also, this
method
v b) (b v c) (a v c) v
where the order of the parentheses, and their position relative to the sign v which stands outside the parentheses, may be altered at will. This single
rule covers all the
modes and
169
illicit
We
shall revert to
Chapter III. The copulas v and v have several advantages over their equivalents, = and =j= 0, or c and its negative: (1) v and v are symmetrical rela tions whose terms can always be interchanged; (2) the operation, mentioned above, of moving v and v with respect to their terms, accomplishes trans
formations which are
the copula;
(3) for
matter
in
less
various reasons,
it is
more
terms
v and v than
in
=
,
and
==
0.
as against
4=
and c
other algebras.
worse,
first is
In his Operationskreis des Logikkalkuls (1877), Schroder pointed out that the logical relations expressed in Boole s calculus by subtraction and divi
sion were all otherwise expressible,
of
meaning
has, first
as Peirce had already noted. The + given by Boole is abandoned in favor of that which it now introduced by Jevons. And the "law of duality", which con
,
or +
and
1,
with corresponding
0, is
theorems
169
in
x, or
x and
emphasized.
111
paper,
we know it today. This perfected and elaborated in Vorlesungen uber die Algebra der Logik (189095). Volume I of this work covers the algebra of classes;
resulting system
is
The
system
is
Volume
is
devoted to the
calculus of relations.
The algebra
is
of classes, or as
we
shall call
it,
170 the system developed in the next chapter. rated the theory of functions, but in all essential
We
respects,
we
There are two differences of some importance between Schroder s procedure and the one we have adopted. Schroder s assumptions are in terms of the relation of subsumption, c instead of the
it
as
appears
in Schroder.
and =; which appear in our postulates. And, second, Schroder gives and discusses the various methods of his predecessors,
The calculus of propositions (Aussagenkalkid) is the extension of the BooleSchroder algebra to propositions by a method which differs little from that adopted in Chapter IV, Section I, of this book. The
Peirce
s
discussion of relations
is
But
mental propositions which Schroder gives for the sake of comparison both with Peirce and with the procedure we shall adopt in Sections II and III
of
Chapter IV.
1)
A, B, C, D,
symbolize
"elements"
or individuals. 171
These
and from
0.
the
order.
j, k,
/,
3)
2,
m,
n, p, q
of
4)
70
Sri
For an excellent summary by Schroder, see Abriss der Algebra der Logik ; ed. Dr. Eugen Miillor, 190910. Parts i and n, covering Vols. i and n of Schroder s Vorlesungen, have so far appeared.
171
The
in, 342.
Many
propositions here noted will be found in Vorksungen uber die Algebra der Logik, others, and much discussion of theory, have been omitted.
112
A
5) i 6)
(i
:
j represents
and j,
of I 1 in a determined order.
j)
(i
i),
(i 4= j)
*j
for every
7)
and
j.
i:j*0
I
1
Pairs of elements of
may
be arranged in a
"block":
...
...
...
C,
D,
:A,
"individual
binary
:
relatives".
(A
+ (B + (C
A) + (A A) + (B A) + (C
:
B) + (A
B) + (B
:
C) + C) + C) +
.
.
B) + (C
S,S, (i
j)
= S S,t
(i
:j)
= S
iy
(i
j)
9)
and
10)
may
1
Ziji :j =
A :A+A :B + A :C+
+
...
B :A
C :A
+
+
B :B + B C+
:
11)
C :B
C+
...
+
12)
13)
i
........
"individual
:j
h will symbolize an
S^SyS,(i
:
ternary
relative".
A)
= 2 A t:j:A
A A
:
A,
B A
:
A,
A B
:
A,
A A
:
B,
A B C
:
:
It is
obvious that we
may
fourth, fifth,
...
or
any thinkable
113
The
a, is
= Sy an
is 1
(i
j) j) pairs in
where a i;
is
for those
(i
which
has
the relation a to
1
and
is
otherwise
0.
==
Ziii.j
/=
JV
(a b)ij
172
S, y
(i=t=
j)(i:j)
(a
aij bij
a./
6) i?
= a, =
&,/
(a),,
(a,y)
(a &);y
1
= 2A
a,*
6 A/
(a t 6),v
n* (a* +
fe*,)
general laws which govern propositional functions, or Aussagenschemata, such as (ab) iit 2 h a ih b hi U h (a ih + b hj ), II a a i} S a a,y, etc., are as
, ,
The
follows
Au
case,
and only
is
Au =
for every
there
"A
at least one
u"
for every
U U A U will have the value 1 in u; 2 U A U will have the value 1 if 1. That is to say, U U A U means
u".
a)
j8)
U u A u cA c2 u A u K U A U = A U V A U1
v V
2UA U
0,
= A
+ 2UA U
(The subscript
7)
d)
u, in
a and
/i u
,
represents
any value
]
[n u ,4 u
= 2U
If ^4 M is
independent of
)
u,
then
A.
e)
r)
U u (A u cB) =
(2 u ^l M
(2 u A u
cB)
cn 5
y
v)
77)
2 u (A u cB) = (n u A u cB), S
I
f
2 u (AcB u ) =
(H u A u
i),
(Ac2 u B u
= =
0)
or 2 u 2 v (A u
(.i
cB
v)
= =
c2,B
9)
nw
=
=
i)
(n u .4
n u (,i u =
S B (.4
==
0)
1)
;
= =
(a

(z M ,i M
(S
tt
o)
1)
172
= (n u /l u =
Schroder has
0),
1
;
.4
where he has
6)
for (a j 6);
for a.
114
A
H u (A u cB u )c]
\
(U U A U c
UUBU
K)
(S.4
tt
c ZU
U U (A U ^B U
o\f/x.)
is
"formal
implication",
in
2 UBU = SM
AB
= S
U)
A + U U B U = U U (A + B u
Au B
V)
//)
(2 U A U )(2 V B V )
yl
ttf
II U
A U + U B = n,
V
V
(A U +
)
v)
nu
= nu A Bu
A
,
+ 2UBU
,
= ? U (A
Bu
(U U A U )(U V B V )
= UU
A U B = UU A U BU
V
o)
2 u n^4 M
Ii v
Su
Au
From
is
developed.
much
new items
of
Schroder
calculus. Perhaps the most inter treatment are the use of "matrices"
relatives,
Dedekind
theory of
as contained in
Was
Anton Poretsky
logiques
des egalites logiques (1899), Quelques lois ulterieures de la theorie des egalites
(1901),
nonegalites logiques
(1904).
(With
will
his
we
Poret
sky
Law
of
Forms,
Law
of Consequences,
and Law
Causes
be
II. As Couturat notes, Schroder had been influenced overmuch by the analogies of the algebra of logic to other algebras, and these papers by Poretsky outline an entirely different procedure which, though based on the same fundamental principles, is somewhat more
given in Chapter
"natural" to logic. Poretsky s method is the perfection of that type of procedure adopted by Jevons and characteristic of the use of the Venn
diagrams.
The work
of Frege,
though
its
interest largely
through
influence
Russell.
Although
the Begrifsschrift (1879) and the Grundlagen der Arithmetik (1884) both
115
hardly more than mentioned his influence upon Peano and other contributors to the Formusurprisingly small when one considers how closely their task is re
Vorlesungen, Frege
is
lated to his.
Frege
is
but, in thorough
German
concerned explicitly with the logic of mathematics fashion, he pursues his analyses more and more
a development of arithmetic of unprecedented but a more or less complete treatise of the logicometaphysical problems rigor
deeply until
and
objects,
subtleties.
In a sense,
fundamental problem is the Kantian one of the nature of the judgments involved in mathematical demonstration. Judgments are analytic, de pending solely upon logical principles and definitions, or they are synthetic.
His
thesis, that
Principles of Mathematics.
metik, like
Grundgesetee der ArithFrege Principia Mathematica, undertakes to establish this thesis for
s
And
arithmetic
work
is
important
is
suf
of defined
it is notation, Frege almost diagrammatic, occupying unnecessary space and carrying the eye here and there in a way which militates against easy understanding. It is probably this forbidding character of his medium, combined with the
equivalents.
it
must be admitted
against him:
unprecedented demands upon the reader s logical subtlety, which accounts for the neglect which his writings so long suffered. But for this, the revival
of logistic proper
Frege
might have taken place ten years earlier, and dated from Grundlagen rather than Peano s Formulaire.
publication, beginning in 1894, of
The
Peano
Formulaire de Mathelogic.
matiques marks a
new epoch
in the history of
symbolic
Heretofore,
and
as Schroder remarks,
we had an elaborated
his collaborators, the
to do.
is reversed symbolic logic is investigated only as the instrument mathematical proof. As Peano puts it: 173 The laws of logic contained in what follows have generally been found
"
173
Formulaire,
(1901), 9.
116
in the
by formulating,
in
form of
mathematical
demonstrations."
The immediate
rate
view
is
new
logic,
no
destined, in fact, to
but with
its
siderations or
of
in
by any mathematical prettiness, but solely by the criterion De Morgan had said that algebraists and geometers live application.
"a
seems to have required the mathe matical intent to complete the rescue of logic from its traditional inanities. The outstanding differences of the logic of Peano from that of Peirce
higher realm of
syllogism":
it
follows:
alternative
The prominence
new
relation,
e,
the relation of a
member
of a
The prominence
implication"
and
of
"formal
and
formal
equivalence",
as against
"material
implication"
and
"material equivalence".
"existence"
(4)
and
of the properties
of classes,
(5)
members of classes, and so on, with reference to their "existence". The properties of relations in general are not studied, and "relative
addition"
in
does not appear at all, but various special relations, prominent mathematics, are treated of.
The disappearance
gain.
(6)
its
The
some value
in
We
The
The Principia is con cerned with the same topics and from the same point of view. But we see here a recognition of difficulties not suggested in the Formulaire, a deeper
of Frege
work
and more lengthy analysis of concepts and a corresponding complexity of procedure. There is also more attention to the details of a rigorous
method
174
of proof.
117
is
The method by which the mathematical logic of Principia Mathematica developed will be discussed, so far as we can discuss it, in the concluding
We shall be especially concerned to point out the sometimes lost sight of, between it and the older logic of Peirce connection, and Schroder. And the use of this logic as an instrument of mathematical
section of Chapter IV.
CHAPTER
THE
I.
II
CLASSIC,
The algebra of logic, in its generally accepted form, is hardly old enough warrant the epithet "classic". It was founded by Boole and given its to present form by Schroder, who incorporated into it certain emendations
which Jevons had proposed and certain additions particularly the relation or implies" which Peirce had made to Boole s system. contained
"
"is
in"
It is
sound judgment that the result is still an algebra, simpler yet more powerful than Boole s calculus. Jevons, in simplifying Boole s system, destroyed its mathematical form; Peirce, retaining the
due to Schroder
mathematical form, complicated instead of simplifying the original calculus. Since the publication of Schroder s Vorlesungen uber die Algebra der Logik
certain additions
of
and improved methods have been offered, the most notable which are contained in the studies of Poretsky and in Whitehead s Uni
versal Algebra. 1
But
this
if
the term
it
"classic"
is
inappropriate at present,
still
we may
out,
all
venture to use
by way
of prophecy.
system
is
differing
from
nonnumerical character.
It is
certainly the simplest mathematical system with any wide range of useful
applications,
it
parent stem
an important type will grow. Already sev The term "classic" will also serve to distinguish
Some
of these, like the system of Mrs. LaddFranklin, differ through the use of other relations than +
1
,
and =
versal Algebra,
See Whitehead s Uni studies, see Bibliography; also p. 114 above. Whitehead introduced a theory of "discriminants" and a treatment
of existential propositions by means of umbral letters. This last, though most ingenious interesting, seems to me rather too complicated for use; and I have not made use of "discriminants preferring to accomplish similar results by a somewhat extended study of
and
",
118
119
to say, with a
"
of these systems
may
dictionary" of equivalent expressions, any theorem be translated into a theorem of the BooleSchroder
Others are mathematically equivalent as far as some, like the calculus of classes in Principia logically but not mathematically equivalent. And,
And
systems such as that of Mr. MacColl s Symbolic Logic which are neither mathematically nor logically equivalent. Postulates for the classic algebra have been given by Huntington,
the Abriss}, by Del Re, by Sheffer and by Bernstein. 2 set here adopted represents a modification of Huntington s third
(in
by Schroder
The
set.*
It
much
for
economy of assumption
as for
"
natural
ness"
and obviousness.
Postulated:
A
11
class
If
K of elements
b are
b.
a, b, c, etc.,
and a
relation
b is
x such that:
in
a and
elements in K, then a x
an element
K, uniquely
determined by a and
1
a,
a xa
=
6,
a.
1 1
1
a x
c,
x
x
a.
4
5
a, b,
and
a x
(b
c)
(a x 6)
x c.
There
a.
is
a unique element,
0, in
such that a x
ment
1
a,
there
is
an element,
x xa y xa
0,
then x xa
and 162
If
y and y x a
y,
then y
0.
The element
These
17 18
and the
relations +
:
in the above.
may 1 = 0
be defined as follows
Def.
Def.
o+ 6 a cb
It
= (ax6)
is
19
equivalent to a x6
a
is
Def.
which
is
by a and
2 3
See Bibl. See "Sets of Independent Postulates for the Algebra of Logic", Trans. Amer. Math. Our set is got by replacing + in Huntington s set by x and Soc., v (1904), 288309. Thus 1 can be replacing the second half of G, which involves 1, by its analogue with 0. omitted. Postulate J is not strictly necessary. defined, and postulates E and
,
120
A
The
above has
if
its
j.
e .,
=
j j
is
y and
and
&lt;p(y)
if
$(x)
is
an ambiguous function
terms of
that a,
x,
is
every determined value of ifr(x), expressible in Suppose, for example, similarly expressible in terms of y.
y,
and x
"negative
of
a",
is
an ambiguous function
"negative
of a.
Still
we may
write
of
a"
itself,
any one
of the negatives of a
and
if
a
b,
then
is
&lt;p(a)
and
&lt;p(b)
will
This principle
We
terms,
shall
important in the early theorems which involve negatives. develop the algebra as an abstract mathematical system: the
a, b, c, etc.,
,
and x
and c
may be any entities which have the postulated properties, may be any relations consistent with the postulates.
we
give
But
two
possible applications:
(1) to
the system of
nullregion included,
and
(2) to
(1)
For the
first
common
to a
and
a
b (their
is
and a +
b will
denote that
"Region
which
is
either a or 6 or both,
(with or without
remainder)".
1
will
itself,
the plane
a,
a
that
is
nota.
The
as follows:
1
1
b,
If
is
common
to a
is
and
b,
a x
the null
region, 0.
1
2
3
a,
the region
to a
common
to a
and
a,
a x
a, is
itself.
1 1
The
The
c
is
region region
common
and
b is the region
is
common
to b
and
a.
and
15
1
the
to a and b x c
the region
common
to a x b
to all three.
The
region
to
nullregion, 0,
is 0.
6
4
there
more
121
and
If
this region
is
such that
a and any region x have only the nullregion in common to x and a is x itself, or x is contained in a;
the region
common
to y
and a
is y,
or y
is
contained
a, then
and the region common to y and a is y, or y is contained in must be the nullregion which is contained in every region.
That the
and
(2)
a, b, c, etc., will
that
is,
b will
mean
and
identically the
same members,
members
of a
a+
b,
contained in every
class.
the
"universe
of
or the class
entity
in the system.
of all
For any class a, a represents the negative of a, or the class The postulates will hold as fol things which are not members of a.
lows
11
If
members com
mon
in
to a
and
common,
this class
the nullclass,
to a
0.
12
1
class a itself.
to a
b
1
and
4
The members common to a, b, and c, all three, are the same, whether we first find the members common to b and c and then those common to a and this class, or w hether we first find the common members of a and b
r
this class
and
c.
to
any
class a
and the
the nullclass.
1
class a, there
is its
all
members
of the
universe of
discourse"
not contained in
161
If
common,
122
then
A
all
members
and 162
of
x are
common
to x
of
and
a,
or x
is
contained in a;
a,
If all
members
and common
17
also to
y and a,
of
The
"universe
discourse",
is
nullclass,
1
"nothing".
That which
is
"a
is
either a or 6 or both
is
that which
19
That
is
contained in b
of a
is
equivalent to
b
".
"The
class a is identical
with the
common members
and
proved by these interpretations. In the form given, they are not independent, but they may easily be made so by certain alternations in the form of statement. 5
is
The
ax6
will generally
be abbreviated to a b or a b,
ax(bxc)
to a (be),
ax(6xc)
2.
In proofs,
we
shall
and thereafter
shall
lemma by
this
number.
Also,
we
sometimes write
instead of repeating
The
by which
any
step in proof is
If
taken
x
will usually
brackets, thus:
0,
whose use
is
more
Theorems
insertion
will
we
of
The non
decimal part of the number will indicate some major division of theorems, as 1 indicates a postulate or definition. Theorems which have this digit
and the one immediately following the decimal point in common different forms of the same principle or otherwise closely related.
II.
will
be
ELEMENTARY THEOREMS
b c
21
If a
6,
then a
and
c b.
= and
1.
2 2
c b and
a.
b.
6,
then [12]
ab =
a and b a
s set.
On
this point,
The
123
And
But
if
a b
=
a b
a and b a
b,
then
[1 3]
a b
[1 9]
is,
is
equivalent to a c b and b
= b a = b. a = b to 6 c a.
Equality
23
aca.
a
a,
Every element
24
a a
itself.
=
[12]
= a a. a a = a. = (a a) a. = (a a) a. (a a) (a a) a = (a a) a = a a, then a a = 0. a = a a. Hence also, a a = 0.
a
(a a)
Hence
a
a (a a)
Also [12] a a
a.
Hence a a = a
But
[162]
[13]
if
And
a
is
0,
and
is
the
25
a b
=
If
is
equivalent to a b
a and to a c
b.
=
if
a,
then
[145,. 214]
a b
(a b) 6
(b b)
And
[161]
a 6
=
is
0,
then
(2),
a b a
=
a c
ab = a and ab = a
c6
(2)
are equivalent.
6.
We
later.
The above
"145"
is
required
"214"
immediately.
instead of
In this proof,
15"
we have
"21,
written
and
"14,
and
24".
26
If
acO, then a =
If
0.
=
c
a.
But
c b.
[15]
a0
0.
27
If
a c
then a
c,
and
a c
[19] a b
b) c
a and [21]
(ac)
b) c
=
Hence, by
(ac)(cb)
= (b a) c = b = (ac)(bc)
a c
c b
6,
= (ac)
c)
= ac = [a (c c)
(1)
b]
= [(ac)
a
c
c]
(2)
(1)
and
(2), if
then (a c)(b
b c.
= ac and
also c a
[1 9]
b.
b c.
And
28
(a)
[13] c a
= ac and

Hence
cc
a.
0.
Hence
Hence
[25] (a)
ca
caa.
(1)
By
(1),
[()] ca.
[27] a [(a)]
124
A
But
[24] a
a
0.
Hence
[2G] a [(a)]
and
[25] a
c(a)
(2)
[22] (1)
and
(2) are
equivalent to (a)
a.
31
a cb
is
equivalent to b c a.
[25]
ac6
is
equivalent to a 6
0.
And And
If
[28]
6 (a)
of
is
equivalent to b ca.
be transposed by negating both. then the portion of the plane not in b is contained in the portion of the plane not in a: if all a s are 6 s, all non6 s are nona s. This theorem gives immediately, by 28, the two corollaries:
The terms
is
any
relation
may
region a
contained in region
b,
312 313
32
a
a c b
is
a c
is
b is
equivalent to a
6.
[22] a [31]
=
is
b is
equivalent to (a
cb and bed).
cb
ca
to a c6.
is
Hence a =
lent to
b is
equivalent to (a
6.
equiva
a
The negatives
322
a
By
28,
we have
also
b
is
equivalent to a
b.
"negative
of"
be
There might be more than one element in the system having the properties postulated of a. Hence in the preceding theorems, a
unambiguous.
must be read
proved of a,
in
"any
negative of a
",
(6)
6,
and
the
33
a
is
uniquely determined by
a.
By
such.
16, there
is
is
Suppose there
Then
Since
all
[28] ( Gl )
(a 2 ).
Hence
a 2
the relation x
functions in the algebra are expressible in terms of the negative, and 0, while is unique and a x b
,
a, b, c, etc.,
is
uniquely
125
determined by a and b, it follows from 33 that all functions in the algebra are unambiguously determined when the elements involved are specified.
(This would not be true
"division"
if
"subtraction"
and
were admitted.)
1 is
333
The element
[15]
is
unique.
is
0.
334
1
0.
[17] 1
0.
Hence
[32]
Q.E.D.
is
335
If
an element
in
uniquely deter
mined by a and
The theorem
337
If
and
c
8.
6,
then a +
and
+a
b.
The theorem
,
of
=.
34
(a +
6)
a
b. b
[18] a +
Hence
.341
(a b). + b)
[(ab)}
a b.
(a
6)
a +
6.
[18, 28]
a + 6
[(a) (6)]
(a
6).
De Morgan
sum
is
product is the sum of the negatives of its factors. The definition 18 is a form of this theorem. Still other forms follow at once from 34 and 341,
by 28:
342
343
(a + b)
(a + 6)
344
345
(a +
b)
(ab)
(ab)
= =
346
a + b.
s
From De Morgan
334, 1
principle, 32,
1
"The
0,
and theorem
x there
is
0, it
corresponding theorem
replaced by
its
terms of +
If in
negative,
valid theorem.
by
positive,
126
since
A
we can suppose x =
system there
is
b, etc.
Thus
theorem
in the
another got by interchanging the negatives and 1 and the symbols x and + This principle is called the Law of Duality. This law is to be illustrated immediately by deriving from the postulates
. .
The
correlate of
1 is
42
a+a
a.
[12]
aa =
a.
Hence
[18,
32,
28] a + a
(a a)
(a)
a.
43
a + b
a.
[13]
a b
b a.
Hence
[32]
(a 6)
(6 a).
Hence
44
a+
(b
\1
[18]
c)
4.1
Q.E.D.
(a
(
=
n
+
A
b)
^\
c.
h\
Hence
[(a 6)
c].
But
And
45
a+1 =
Hence
[15]
a0
0.
Hence
[32] (a0)
[346] a +
0
0,
and
x.
[17]
= a+1 = = =
0.
1.
461
If
x + a
If
1,
then x a
1,
x + a
[25]
then [323444] x a
is
(z + a)
x.
=1=0.
And
4612
x a
equivalent to x a
If a;
+a
1,
then x + a
= =
y,
a.
[461] If
a + x
if
1,
then a x
1,
a,
and
a.
1.
[32]
a + x
a
(1)
By
462
If
(1)
and 28,
x + a
x+a
= =
a)
y+a
If
y and y + a
y, [32]
= =
then y
And
But
48
a + a
y+a = if y + a
if
y a
y,
y a
=
y. y,
y.
[162]
y a
y and y a
= =
y
and y = 0 =
1.
l=a + a.
a a
(Correlate of 24)
[24]
0.
Hence
[32] a
+ a
is 1.
(a a)
=0 =
all
1.
of the operation +
1,
a+ b
=
0,
b,
a 6
=
a,
0,
a,
and a c b are
equivalent.
[2 5]
a b
6
a 6
[32]
a +
1 is
equivalent to a b
(a +
b)
=1=0.
M,V)
127
=
b,
1,
a+ b
b.
And
if
a+b
= =
[3
37] a + b
a +
(a
b)
(a + a) + 6
= 1+6
1.
Hence a +
b is equivalent to
a + b
1.
We
51
If
ace.
cb
a
is
is
equivalent to a b
c
a,
and
a
cc
to b c
=
a.
b.
a b
=
c
a and b
b,
(a b) c
c.
(b c)
a b
But a
equivalent to a c
the Syllogism.
This law of the transitivity of the relation c is called the Principle of It is usually included in any set of postulates for the algebra which are expressed in terms of the relation c
.
52
a b
cb.
521
and
ca +
b.
a b c a and a b c b.
Hence
[312]
But (a
b)
(a
b).
correlates
by the Law
of Duality.
In
,
we
theorems
such pairs.
A
522
ca + b.
[5
12 21]
c
53
If
a c b and
[19] If
d,
then a
c
cb
d.
a c b and
cd, then a
b
c,
a and
c
d
d.
c.
Hence
531
If a
(a c) (b d)
b
=
d,
(a b) (c d)
=
c
b
and a
cb
and
then a +
d.
If
a c 6 and
[53] [18]
Hence Hence
bdcac, and
Q.E.D.
a and a + a
By
532
the laws,
If
aa =
a,
a c
and
b cc,
then
ab
cc.
128
A
If
533
a c c and b c
c,
then a + 6 c
c.
If a If If
c.
c6
c.
(Correlate of 27)
[23]cce.
54
a + ab =
a.
Hence
[531]
Q.E.D.
[521]
[23]
aca + ab
a,
(1)
a c
and
[52]
(2),
ab ca.
Hence
[5
33] a + a
ca
(2)
[22] If (1)
and
a.
then Q.E.D.
541
a (a +
b)
[54]
a + ab
a.
But
[34] (a +
&) =
Hence
(a)
a.
Law of Absorption. We have next to prove the Distributive Law, which requires several lemmas.
54 and 541 are the two forms of the
55
a
(b
c)
= ab +
ac.
Lemma
1:
[52] [52] a b
Hence
[533] a b + a c
ca
cb+
c.
(1)
and a
c
b
c
c
c.
But
6
[521] b
c.
and
Hence Hence
[51] a b
[5
and ac cb +
c
+
c
33] a b + a c
(2)
[534] If (1)
and
q
(2),
then a
+a
ca
is
(b
c).
Lemma
If
pc
q.
is false,
then there
an element
x, 4= 0,
such that
x c p and x c
pq
[49]
if
is
pqcp
q
(This
pc
is false,
Lemma
3.)
Lemma
4=
(6
c)
+a
c.
false.
Then, by lemma
2,
there
is
an element
x,
xca(b + c)
.TC(6 + ac) But [312] if xc(b + ac), then b + accx [51] If (1), then since [52] o (6 + oc) co, z
(1)
and
(2)
co
(3) (4)
(5)
and
Also
+ a
c)
c, .r
[51] If (2),
[5
1] if
then since [521] b cb + a c, b c x and [312] xcb (2), then since [5 21] a c c b + a c, a c c x and [312]
zc(oc)
(6)
The
129
From
(6)
and
x c
(3), it
a,
follows that
[5
xcc must
But
if
be
c
false;
for
if
.TCC
c),
and
(3)
then
34]
x ca
c.
x ca
and
(6)
x c (a
then [162] x
0,
But
if
x c
be
false,
0,
such that
ycx
and
[51] If (7)
If (8)
If (7)
(7)
?/cc,
(5),
or [312]
and
(9),
and
and
(4),
then y cb and [312] then [533] b + ccy and [312] yc(b + then [51] ycb + c
ccy bcy
(8) (9)
c)
(10)
(11)
[19] If (11),
But
y
if
(12),
(10),
y~(b +
c)
(12)
*0.
false
is
Hence the supposition that a(b + c)cb + ac be supposition, and the lemma is established.
a false
Lemma
+ a
c.
(b
+ a
c)
(a c
b).
But by lemma
3,
a (a c + b) c a c + a
c.
b.
And
Proof of
a c + a b
a b + a
Hence a
1
(b
c)
a b + a
c.
the theorem:
[22]
Lemma
c.
lent to a (b + c)
a b + a
Law
is
The proof of Algebra of Logic due to Peirce, who worked it out for his paper of 1880 but mislaid the sheets, and it was printed for the first time in
of
Independent Postulates
difficult
for the
lemma
is
Huntington
551
(a
paper.
+ b)(c + d) = (a c + [5 5] (a + b)(c + d)
b c)
(a
d+
b d).
(a +
b) c
(a
b)
(a c
+b
c)
(a
d + b d).
552
a + b c
(a + b)(a
c).
(Correlate of 5 5)
[551] (a +
b) (a +
c)
(a a + b a) + (a c + b c)
=
But
[54] (a + a
b)
[(a
+ a
b)
+ a
c]
c.
+ a
a + a c
a.
Hence Q.E.D.
follows:
300, footnote.
Further theorems which are often useful in working the algebra and
See
"Sets
of
Independent Postulates,
etc.",
loc. rit., p.
10
130
A
a1
56
1 a.
[15]
a0
if
0.
Hence al =
0.
But
561
[101]
al
= =
0,
then a1
a.
acl.
[19] Since a1
a,
acl.
a0
563
a1
a.
Hence
[3 2]
a+
(a0)
(a)
a.
Oca.
0a
a0
0.
Hence
[19]
Q.E.D.
564
ca
is
equivalent to a
1.
[22] a
1 is 1
acl and
ca.
But
565
a c
[561] a c
is
holds always.
Hence Q.E.D.
equivalent to a
0.
[2 2]
is
and
a.
But
57
If
[563]
a;
ca
holds always.
Hence Q.E.D.
x.
a + 6
If
= and a = 0, then 6 = a = 0, a + 6 = + 6 = 6. =
x and a
571
If a 6
= =
1,
then 6
x.
If a
= =
1,
a 6
16
6.
572
a + 6
If
=
a
is
and
0.
and
=
=
0,
then a + 6
And
But
if
if
a+ 6
a 6
if
= =
0,
1,
a 6
a
(a +
6)
= =
= =
1.
0.
0
a1
a(a 6)
(a a)
6
06
0.
And
573
a 6
[57]
1 is
a+ 6
and a
0,
then 6
0.
=
If
and
1.
And
if
a 6
[572]
a
57 and 572 are important theorems of the 57, "Any null algebra. term of a sum may be dropped", would hold in almost any system; but a sum is null, each of its summands is 572, is a special law null",
"If
It is
due to the fact that the system con and 0. a and a are inverses with
131
x and
1.
correlates of 57
58
(b
+ b)
[55]
=
a
+ 6)
=
1.
a b + a b.
And
585
a+
b
[48] b
+ b
b.
Hence a
(b
+ b)
al
a.
a + a
b)
+a
flb.
be convenient to have certain principles, already proved for two terms or three, in the more general form which they can be given by the
use of mathematical induction.
of such extension is
Since both
x and
we can
5901
a+ b+
c=(a + b)+c
Def.
5902
591
+ c) (d + d)
592
=
a
(a
593
a+
ab + ac + ad+...
(a + d)
[54]
5 93 1
=
+
a (a + b) (a +
[541]
c)
594
(b
+ d + ...)
[55]
= ab + ac
ad+
.
...
5941
a + bcd.,..
[552]
=
(a
+ b)(a + c)(a + d)
595
(a + b +
If
= a
b c ...
+ a n)
ai
a z
a n
then
it will
hold for n
And
[34] the
theorem holds
for
two terms.
Hence
it
holds for
any
number
of terms.
132
A
(a
5951
bed...) = a + b +
a+ b + c +
d+
596
+ab
c
[48, 5951]
597
a + b+
is
equivalent to the
set,
0, 6
0, c
0,
[572]
5971
abed...
[573]
=
1 is
equivalent to the
set,
==
1,
1, c
I,
598
ab
[12] a a a a
=abacad... = a.
.
5981
a +
(b
+d+
)
.
=
. .
(a
b)
(a
c)
+ (a + d) +
[42] a + a + a +
a.
s
The
extension of
De Morgan
is
especially
important.
by
may
be expanded into a
sum and any elements not introduced into it. Thus any expression
be regarded as a function of any given elements, even though in the expression, a peculiarity of the algebra. 592, the expression of the universe of discourse in any desired terms, or expansion
may
of
1, is
many important
procedures.
if
the
number
of elements
We may
only
members
denote any expression which involves and the relations x and + The further
.
requirement that the expression represented by /(.r) should involve x or its negative, x, that $0, y) should involve x or .r and y or y, is unnecessary, for if .r and x do not appear in a given expression, there is an equivalent
expression in which they do appear.
a
By
591,
?/)
a (x +
.r)
a x + a
x
(a
x + a x) (y +
a x y, etc.
reference
reference
may be called the expansion, or development, of a with And any or all terms of a function may be expanded with
and
so
of elements.
133
ments
in
and the
relations
x and +
may
be treated as a function of
as the
"value
of x
",
then a value
as in
any
other.
any function of x is determined, in this algebra But functions of .r in this system are of two types: (1)
may vary, and those such that any value of the function being assigned, the value of x (2)
thereby determined, within limits or completely. Any function which symmetrical with respect to x and x will belong to the first of these
in general, a function
is
is
classes;
which
is
respect to x
and x
will
But
it
must be remembered,
unless
it
symmetrical function may not look symmetrical be completely expanded with reference to each of the elements
involved.
For example,
a + a b + b
is
symmetrical with respect to a and a and with respect to panding the first and last terms, we have
a
(b
and
b.
Ex
+ b) + 
(a
+ a) b
b.
it
a b + a b + a b + a b
Any
is
.r,
symmetrical with respect to x and x. The decision what elements a given expression shall be considered a
function of
is,
it
is
deter
of result desired.
"unknowns"
The
or
is
not fundamental.
In fact, we shall
frequently find
it
say
of x
and
y,
convenient to treat a given expression first as a functionthen as a function of z, or of x alone. In general, coef
ficients will
a Function.
Any
/(.r),
A
where
x+
x
is
and
are independent of x.
This
of one variable.
Any function
of
is
and some
/(.r)
= A
x +
x
134
A
Any
the relations
and expression which involves only elements in the class x and + will consist either of a single term a single
or of a
sum
of such terms.
Only
x,
(1)
those which involve x, (3) those which involve both, and (4)
and
of (x x), the
an expres
sion
is
px + q x + r
where
p,
q, r,
and
s are
But
[24]
And
[59]
Hence
px
x) + s
q +
=
s,
(p + s) x + (q + s) x.
Therefore,
A = p + s, B =
of a function of
C2,
+
Xn
,
variables,
+!
may
Law
of
X2
...
X n ) X n +i +/
(.Tl, .T 2 ,
X n ) X n+ i
x*&gt;,
where/ and/
in the
are each
some function
is
...
x n and
,
normal form.
This
"step
by
of a function of
two variables
is
of a function of
two
found by expanding
y
(A x +
It will be,
x)
y+(Cx + D x)
Axy + B x y + C x y + D x y
of a function of three variables, ty(x, y, z), will be
G x y
z + II x y 2
And
7
so on.
Any
By
a term which
is
either is x or has
is
factor".
But
"factor"
seems inappropriate
an algebra
in
which h x
x a always contained
"as
in x, h
8
x ex.
This definition alters somewhat the usual order of terms in the normal form of func tions. But it enables us to apply mathematical induction and thus prove theorems of a generality not otherwise to be attained.
135
that
is,
each variable, or
its
negative, will
appear
in
every term.
be given the normal form.
of
611
Any
function
may
1
,
(a)
By
If
(&gt;
any function
one variable
may
be given the
normal form.
(6)
then functions of n
1
,
Let $(1,
Xz, ...
variables.
By
definition, its
normal form
be equivalent to
(Xi,
X n ) Xn+l +f
XZ
,
...
X n ) Xn+l
wLere / and /
form.
...
By
x n x n +i)
,
may
be re
Hence, by
i
61, for
some
and some
of x n ^
,
X n Xn+l)
= A
X n+ i +
X n +l
7
Also,
by the
A = f(xi,
and
Hence, for some / and /
&(Xi, X 2
,
. . .
x2
...
.
.
B =
(xi, Xz,
xn)
of x n +\
X n Xn+l)
,
X n ) X n +l +/
(.Ti,
X Zt
X n )Xn+l
Therefore,
if
x n x n +\) can be given the normal form. normal form, then $(0*1, Xz, be given the normal form, (c) Since functions of one variable can
and
since
if
functions of n
functions of any
number
The second
is
above proof
may seem
,
.
arbitrary.
That
it
valid, is
in this algebra.
. 
x n ), the normal form will n variables, &(xi, x 2 all the combinations of x it positive or be a sum of 2 terms, representing with .r n positive or negative, with x z positive or negative, with negative, each term having its coefficient.
612
For a function
n
136
A
(a)
by
will
definition of the
normal form function of one variable has two terms, and normal form of functions of n + 1 variables, if
variables
have 2 k
(b)
2 k or 2 A+1
,
terms.
normal form function of one variable has the further character described in the theorem; and if normal form functions
of k variables
will
have
f
variables
have
1
it,
l)st
may
1.
be
0,
may
con
which are
is,
null.
of course,
of terms in
function will vary as the order of the variables in the argument of the function is varied. For example, the normal form of y) is, by defini
3&gt;(x,
tion,
A
and the normal form
xy+
B x y + C x y + D
is
x y
of V(y, x)
only in the order of the terms and And since + and x are both associa
is
not material.
differ materially
615
Any two
functions of the
only
The theorem
that, for
In consequence of 615,
we
assume
in the
the same.
,
And
is
also,
any function
of
variables,
3&gt;(
Xl
x2
. .
xn
.r
n+1 ),
which
(X,,
1
.To,
equated to
.
/Ol,
x n+l
it
.12,
Xn)
X n+1 +f
X n ) X n+l
may
variables.
is
always the
617
null.
The product
of
137
By
there will be
some
variable, x n
such that x n
is
positive in one of
them and negative in the other; since otherwise the two terms would represent the same combination of x\ positive or negative,
t
with
0*2,
Consequently, the product of involve a factor of the form x n or n and will
,
therefore be null.
it
will
mentioned
is
in the
normal form.
The
Coefficients in a Function.
The
itself.
coefficients in
621
If f(x)
= A = A
y.
11 .r,
then /(I)
= A.
+ B0
For /(I)
622
If /(.r)
= A1
x+
+ U1
= A
A.
For/(0)
623
/(.r)
==
B.
=/(!)
a;
The theorem
These laws,
first
and 022.
stated
by Boole,
For example,
x + d x) +
(c
c (d
.r)
method would be
tedious.
But we have
d
*(1)
and
*(0)
+ (c+l) d +
(c
= acd+cd+d =
+ 0) d
a c d + c d
given by
c
\F(.r)
d x + (a
d +
c d)
x
for functions
Laws analogous to 623, also stated by Boole, may be given of more than one variable. For example,
/(.r,
y)
/(I, l)x
y +/(0,
?/
1)
a y +/(!,
1, 1)
.)*
0)
.r
y
+/((), 0)
0,
.*
//
and
$(.r, y, z)
$(1,
1, l).r
+ $(0,
+
&lt;I&gt;(1,
l)xijz
.r
yz
0) x y z + $(0,
0, 0)
.r //
z
We
functions of anv
number
of variables.
138
A
If
624
x n
f J
be any term of
^(.TI, xz,
x3
...
#)&gt;
then
j
[
will
be the coefficient, C.
I 0, 0, 0,
...
(a)
(6)
By
If
variables,
it
will
hold
for functions of k
variables, for,
By
is
611,
such that,
.T*,
0, 0,
...
Therefore,
if
{1,1, I 0, 0,
...
...
ll
j
f
I
.n
xz...
. . .
xk
\
}
x, x 2
is
x k
in
which x k+
r
fi,i,...ii
I
0, 0, ...
J
x,
x^.. xk
. . .
\
J
.n x 2
xk
\
(^
and the
coefficient of
will
be /
r U, U,
... U J
which,
And
similarly,
if
every term of /
...
...
be of the form
.,f 1,1,
I 0, 0,
11
j
.rj ,r,
x t
*
1,
in
...
which
x/,+i is
f
l
Tl
x,...
.rt l
Mo.o,. .ol
i.,,.,,.
.,,
139
and the
coefficient of
will
be /
which,
of
&lt;
will be of the
form
x2
...
.
1,1,.
..MI
j
0, 0, ... 0,
I .?! .r 2
x k xk+l
(c)
since
of k
if it
it
number
of
variables.
For functions
but
less useful If f(x)
of
Ax + Bx:
= f(A
+ B).
625
/(I)
/(())
= f(A
626
= AB =
/(4)
= f(AB) = f(A
+ B)
The
will
form
$(.?!,
x2
x n ) by A\,
z,
3,
A^,
or
by
Ci,
C2 C3
,
.,
etc.
This notation
is
fixed.
If
the argument
,
x n ), then any one of the variables, x k any function be (a*i, .r 2 be positive in the term of which Cm is the coefficient in case
.
will
p2
~l
&lt;
m^
(p
l)2
~l
.r^
where p
= any
in the term.
Otherwise
for
will
be negative
in the
is
function, whether
coefficient,
term
of
which Cm
the
We make
no use
be convenient to distinguish the coefficients of those Occasionally in which some one of the variables, say x k is positive terms in a function from the coefficients of terms in which x k is negative. We shall do this
,
140
terms in
for coefficients of
terms
in
which Xk
of terms,
is
negative.
This notation
is
number
it is
is
for a function of
is
n variables,
always equal to the number of those in which and the distribution of the terms in which x k is positive, or
positive
is
negative,
is
negative,
+
2
.,
3
will frequently
...
be indi
h
^A
is
or
^A
h
the product,
AiA A
by JJ/l or YL^h
number
tion which
in question, it
The Limits
the coefficients.
of a Function.
The lower
limit of
any function
is
is
the prod
the
sum
of
63
A B
cAx + BxcA + B.
(A B)(A x +
x)
= AB
x +
AB x
= A B
Hence
[19]
ABcAx + Bx.
B x)(A
+ B)
And (A
But
x +
Ax + AB x + ABx + B x
=
(A
B + A)
and
x +
(AB + B)
A
x+
x.
[19]
x
cA + B.
631
632
then
;i),
be
C C2 C
lt
,
3,
(a)
(b)
By
3,
Let $(xi, x z
for
.
. .
variables.
By (Ml,
$(.ri, *2,
f(x lt x z
xk )
iCfc+i
(.TI,
+/
Since this last expression
x2
x k ) xk+ i
(1)
may
x2
which the
i,
/ and /
$(.ri,
[63]
,
. .
x2
x k ) x/
(#,,
xh ) c
x2
x k x k +i)
,
141
{3&gt;},
etc.,
Ai{f\,
^2 {/I, A*\f},
etc.,
etc.,
\,
{/
[,
{f
},
the coefficients in /
If
c/
then [03]
and, by
(1),
EU
{/J
JL
,
!/
&lt;.
But
in
any
/
coefficient in
will
be either a coefficient
/ or a
coefficient in
and hence
!
lUf/! xILi{/
Hence
if
==IU!*i
x2 x k ),
II A
then
{/}
c/Cd, * 2
xk )
...
and
,
EU
l,
{/
c/
(x lt
II^l^!
if
c$(ai, x z
x k x k+} ).
Hence
/c
.!{/}
and / c
[531]
*&lt;=Z^{/1
Z^{/
But
since
any
,
coefficient in
is
either a coefficient in
==
ficient in
z^m
1
/ or a coef
z^j.
it
will
variables.
since
if it
it
of k f
holds generally.
As we
Functions of Functions.
be given the same normal form, the operations of the algebra be performed simply by operating upon the coefficients.
64
If f(x)
= A
x+
x, then
4/0)] 
.1
.r
B
x.
[34]
(A x +
x)
= (A
= =
x)(B .r)
.r)
.1
B
+ .1
.T
B x
(A B
+ A) x + (A
+
B
+ B) x
But
[54] .1
B
+ .1
x)
Hence (,1 x +
B = B.
142
A
The negative
of
641
any function,
normal form,
its
is
found by
re
negative.
By
If
is
will
variables.
Xk, Xk+i)
2, for
be any function of k
,
variables.
+f
But
f(xi, Xz,
.
. .
(xi,
x2
x k )x k+l
x k )x k+ i
+f
Oi&gt;
^,
Zk)
Xk+i
may
be regarded
as a function of a^+i.
Hence, by 64,
,
ao,
x k ) x k+l +/
(xi,
xz
x k ) xk+ i]
Hence
if
the negative of /
its
is found by replacing each of the coefficients in / by and the negative of / is found by replacing each of the negative
coefficients in
by
its
will
be
found by replacing each of the coefficients as has just been shown, any term of
F by
its
negative, for,
x
in
l}
xz,
x kj
its coefficient is
which x k +i
is
positive
is
such that
X2
,
a coefficient in
[/(&gt;!,
...
X k )]
of
[F(x lt
in
xz,
xk
which x k +\
is
negative
is
such that
its coefficient is
a coefficient in
[/ (*
(c)
X*
...
.T,; )]
(b)
for
is
com
expanded
so that every
term
an expanded function is missing, its coefficient is 0, and negative of the function that term will appear with the coefficient 1.
of
143
642
The sum
#2,
.
.
of
.
of the
same
variables, $(0:1, T 2
xn)
and V(xi,
x n ),
such that the coefficient of any term in the corresponding terms in $ and ^.
is
the
sum
of the coefficients of
By
615,
$(0:1,0:2,
...
x n ) and V(xi, x 2
...
xn)
cannot
differ
3,
etc.,
be the coefficients in $;
BI,
B
TI
2,
B
x*
3,
etc.,
the
terms in ^.
. .
r responding terms,
f
Ak J 1
.!
TI
TO
Tn
L 0*1 .1 2
r ~.l n J
and B k
Tn
1
f
xn
.
.
]
}
l
I .Ti .r 2
.
X n
*1
*2
...
Ah
f
nk
f
j
T!
0: 2
...
.
. .
Ti 1 2
Xn
Xi X 2
Xn
Ti
T2
Xn
x n
of the
TI T 2
And
since addition
is
two functions
and commutative, the sum equivalent to the sum of the sums of such
is
associative
corre
pair.
The product
^f(xi,
0*2,
of
two functions
is
of the
same
variables,
$(1,
o: 2 ,
...
xn)
and
...
x n ),
Xn),
F
^.
f
is
coeffi
$ and
Xi
X2
Xn
TI T 2
x n
and Bk
^.
Tl
0*2
Ti
T2
...
.
. .
Xn
}
r }
T! T 2
x n
be any two
corresponding terms in
$ and
J
I
A k \[
,
TI
T2
n 1
...
.
. .
Xn
Ti T^
& H X n Jfx5
Ti T 2
X n
T n
By
615,
$ and ^ do not
differ
product of any two terms is null. Hence all the crossproducts of terms in $ and ^ will be null, and the product of the functions will
144
A
be equivalent to the
pair
sum
by
pair.
Since in this algebra two functions in which the variables are not the same may be so expanded as to become functions of the same variables,
these theorems concerning functions of functions are very useful.
TV.
We
may
of
have now to consider the methods by which any given element be eliminated from an equation, and the methods by which the value
"
an
unknown
"
may
The
is
and solutions
the
member
Equivalent Equations of Different Forms. If an equation be not in the form in which one member is 0, it may be given that form by multiplying
each side into the negative of the other and adding these two products.
71 a
b is
equivalent to a b + a b
0.
[22] a
[49] a
b is
0.
And
[5
72] a b
and a
+ a b
0.
The transformation
712
a
of
1 is
obvious:
1 is
equivalent to a
0.
[32]
By
in
641,
in
form
xn) = 1 any equation of the form f(xi, ,r 2 which one member is simply by replacing each
,
. . .
is
reduced to the
of the coefficients
/ by
its
negative.
is
Of especial interest
members
713
If
same
variables.
,
.T 2 ,
Xn)
and
ty(x 1} x 2
...
same
variables, then
is
if
equivalent to F(XI, x 2
xn)
0,
where
&lt;,
A A A
i,
2,
3,
etc.,
be the coefficients in
and
that
coef
ficients of the
corresponding terms in ty, then the coefficients of the corre sponding terms in F will be (Ai Bi + A l #0, (A 2 B 2 + A 2 B 2 ), (A 3 J5 3 + .4 3 # 3 ), etc.
145
By
By
7
1,
&lt;
$ = ^
is
equivalent to ($ x\F) +
of the
(&lt;!&gt;
x ^)
0.
&lt;
641,
same variables as
and ^.
be a
Hence, by 643, $
x^ and $
(i42,
will
same
variables,
and by
^,
(3&gt;
x^) + ($ x
will
also
Hence
3&gt;,
^,
&lt;,
&lt;J&gt;
(&lt;
x ^) are
all
functions of the same variables and, by 615, will not differ except in the coefficients of the terms.
If Ate
be any coefficient in
&lt;,
and
Bk
$
will
be
A k
be
.
be
Bk
in
$x^
be
will
fy will
(&lt;
A k Bk
(&lt;
will
coefficient in
x^) +
x ^)
is
By
in the
7
1,
is
an equivalent equation
be
form in which one member is 0, and by 7 13 the reduction can usually made by inspection. One of the most important additions to the general methods of the
s
work
Poretsky
Law
of
Forms.
By
t
lent equation of
may
715
is
equivalent to
a t + a
If a
And
Since
if t
t.
(a t
t
+ a
t)
t
+ (a
+ a
t)
a t + a
may
:
The more
general form
716
b is
1]
equivalent to
a
(a b
+ a 6)
+ (a b + a
b) t.
[7
b is equivalent to a b
+ a b
0.
And
[64] (a b
+ ab) = ab
+ a b.
Hence
[715]
Q.E.D.
The number
9
and expressible
i.
See Sept
11
lois
146
in
A
will
terms of n elements
~ can be formed from n elements and their negatives, that is, 2 /2. The sixteen distinct functions expressible in terms of two elements,
a and
b,
are:
b,
f
a, a,
b,
(i.
e.,
etc.), 1
6,
(i.
e.,
b, b.
^
_a i
_ rt
^
a +
6,
a +
b,
ab
+ a b, and a b + a
b are:
=&;
=
a 6 + a &
a 6 + a 6
b.
ab =
a + b;
a b
ab; a b
a + b: and a + 6
a +
Each
side of
an
equation, and none appears twice. For any equation, there is such a
set of equivalents in
terms of the
elements which appear in the given equation. And every such set has = a b + a b) what may be called its "zero member" (in the above,
its
"whole
member"
(in
the above,
a &+
&).
If
we observe
member"
we
"zero
member"
chosen
is
is
determined.
Any
the
=
{
0}
=
{
1
t,
The
validity of the
i.
law depends
It is
Octfcl,
e.,
1/ + 0/.
Solution of Equations in
sible
One Unknown.
is
pos
according to the laws of the system has a solution for each of the
un
to
knowns involved.
equations
in
This
is
We
turn
first
one unknown.
Every equation
in x,
.
if it
be possible in the
72
x +
x
is
equivalent to
c x c A.
[572]
x +
x
is
Ax
and
B x = 0. [49] B x = And A x =
721
is
equivalent to
ex.
is
equivalent to x (A)
A"
0,
hence to x cA.
is
x+
x
0.
The But
solution then
if
A
1.
x+
x
A ex cA. = 0, then A A
=
(x
+ x)
= A
x+
x
0,
and
A =
147
Hence the
satisfied
solution
is
equivalent to
x.
is
by every value of
In general, any equation symmetrical with respect to x and x which gives the solution, // c x c K, will give also // c x c K. c x and // c x, then [4 9] // x = But if and // x = H.
Hence
[162] II
0.
And =
if
xcK
[533] x +
xcK, and
1.
[48, 563]
I.
Hence
H ex c K will
be equivalent to
if
c a: c
It follows directly
may
722
ever
An A =
x +
x
B,
B = A. [322] A = B
is
and
.1
= B
are equivalent;
hence either of
these conditions
[7
21
equivalent to both.
is
it is
x+
if
Hence
B x = B = A,
[22] to
equivalent to
equivalent
B c x c A. to B ex cB and
x+
and hence
= B = A.
x
=
is
determines x be
tween the
if,
limits
B and
A.
unique
if,
and only
when they are respectively and 1, the limiting values of variables generally. 7221 The condition that an equation of the form A x + B x = be pos
sible in the algebra,
its
solution be possible,
[565]
if
is
AB =
0,
0.
By 63, A B A B = 0. Hence if A B
=(=
cAx + Bx.
0,
Hence
x +
B x =
then
then
x+
B x =
must be
of x.
equation of
it is
B x =
the solution
Bc.xc.A
false, or
equivalent to
x +
x
=
will
0,
whether
x +
x
be true,
impossible.
solution of
The
sense in which
AB =
made
x+
B x =
be
clear in
what
follows.
The equation
of condition
is
148
(In this connection,
A
it
AB =
follows
from
x +
x
0.)
For example,
a b x + (a +
6)
x
=
of condition
is
then (a +
b) c.r
c(a
b).
a b (a + b)
= ab =
0,
is
or,
(a
b)
may
be written a+ b
ex
is
However,
this simplified
satisfied
and a
0.
Again suppose
ax
+ b x +
=
we have
Expanding
with reference to
(a
x,
and
(b
collecting coefficients,
c)
x +
c)
x
of condition is
+ c)(b +
c)
a b + a
b
+b
= ab + =
The
solution
is
c x c a c
c
0,
and hence c =
1.
c x c a
is
But here
ex ca
only on the assumption, contained in the equation of condition, that c = 0. This example may also serve to illustrate the fact that in any equation
is 0, any terms which do not involve x or x may be dropped without affecting the solution for x. If a x + bx + c = 0, then by 572, a x + b x = 0, and any addition to the solution by retaining c will
be indeterminate.
All
its
of the equation
to be
explained shortly.
Poretsky
is
Law
x
of
Forms
7 23
x+
B
=
=
x +
is
equivalent to x
= A
x +
x.
[715]
x
is
equivalent to
(A x +
x) x + (A x +
B
r)
r B
x +
Ax
149
of Jevons. 10
is
by the method
is
Although
it
is
this
in reality a useful
and
form
is
of the solution.
It follows
= A
x +
B x
equivalent to
B
is
cx c
.1.
Many
This
is
the most
"mathematical"
724
+
If
AE =
is
0,
as
the equation
x+
B x =
= B
requires,
then
Ax
is
x
=
x
satisfied
by x
= B u + A
is
u, or x
+ u A, where u
arbitrary.
x +
B
.
solution
is
some value
+u
A
(a)
By
if
64,
if
u + A
?/,
u,
then x
= B u + A
u.
Hence
x +
B x
if
= B u + A then = A (B u + B u) + B
= B u
(B u +
A u) = AB
?/
AB
= A
Hence
of u,
x+
A B = and B x = 0.
.r
A
u,
x
0.
(B u +
A
if)
x + (B u +
n) x
= (A
The condition that
x +
A
x) u + (B x +
B
is,
x)
u =
some value
x + .1
of
by 7 221,
(A x + A x) (B x + B x)
This condition
is
= A B
x
satisfied
if
A
(c)
(B + B) x + (A + A)
and by 572,
If If
A B x + A B
0,
0.
AB =
then
B u + A = 0.
= B
uA,
for:
A B u = B + A B
u.
[585]
B + A Bu =
B+n A.
150
of this solution, x
= B+
u A,
will
be used hereafter.
The above
We
then have
A
And
if
(B u + A
0,
?/)
(B u +
is
AB =
the solution
That the
solution, x
B ex cA,
parameter,
If If
will
u, is
be clear
we
u =
AB
Hence x =
stated
0,
B + A = A B + A =
A.
fact,
B u + A u = B + uA
of condition of the
by
B ex
1,
simply expresses the cA, that the limits of x are B and A.
otherwise
The equation
form
C x + D x =
7 25
and
form
x+
B x = C x + D x,
1 is
follow readily
The equation
of condition that
C x + D x =
C.
C+D =
1,
and the
solution of
Cx
D x
63,
1 is
D c x c
(a)
By
if
Cx + DxcC + D.
any value
1.
==
Hence
there be
of x for
which C x +
D x =
0,
1,
then
necessarily
If
C+
D =
1,
then [64]
Cx + Dx =
1
and
[72]
C+
D =
1,
Cx + Dx =
equivalent to
is
satisfied
by
= D
uC, where u
Since [64]
arbitrary.
1
is
Cx + Dx =
is
Cx + Dx =
0,
and C +
D =
equivalent to
C D =
0,
from 724.
727
If
x+
B x = C x
B D
is
(A C +
A C)(B D
BD) =0
C, or
and the
solution
is
B
cx c
A C + A
+
+
= B D +B D By
7 13,
+ u (A
C + A
C), where
is
arbitrary.
x +
x
= Cx
x
is
equivalent to
(AC + AC)x
(BD
Bdx =
0.
151
Hence, by 7221, the equation of condition And by 72 and 724, the solution is
is
as given above.
or
is
arbitrary.
The
is
clearest notation
we have been
able to devise
in
somewhat cumbersome.
73
one unknown,
x+
x
0,
is
A*x +
x
0,
x+
x
h
0,
may
)
be regarded as simultaneous,
(A
fi,
Bk =
is
And
^B
k
cxc
is
HA
k
or x
^B
k
u&gt;
II ^* where u
k
arbitrary.
By
642
and
572,
x+
5
1
x
0,
A*x +
x
0,
x+
x
0,
1 (A +
A n )x +
,
(B + B* +
B")
x
or
k
z +
k
*
=
is
By
A* x
A:
B =
k
A
But
ZM* x Z
A
fc
=
1
(.I
+
. .
,I
+...+.4
= A B
1
+
l
B +
+
.
.
A Bn + A B
1
+ A*
+
A"
A*B
+ A*
+ A* B +
B B
+ +
A B + An B
+
2
Z(^*).
h,k
is
or
*=
A
A
And by
It
595,
{^M
k
==
II ~A
k
k
.
may
?i
partial solu
152
tions of the form
A
B
h
c.r
cA&gt;
Wc^B*
k
2 Similarly, 2
"
and
H A*
k
cA&gt;
partial solutions
members
of
k ^Z H
an d II
A k
Symmetrical and Unsymmetrical Constituents of Equations. Some of the most important properties of equations of the form A x + B x = are
by dividing the equation into two constituents the most comprehensive constituent which is symmetrical with respect to .r and x, and a completely unsymmetrical constituent. For brevity, these may
clear
made
"symmetrical constituent"
and the
unsymmetrical
In order to get the symmetrical constituent complete, it necessary to expand each term with reference to every element in the
Thus
in .1
x+
x
it is
necessary to
expand the
first
A
By
(B + B) x + (A + A)
5 72, this
.1
is
x
= A Bx
AB x + A B x + A B x
.1
(x
.1
B =
and
B
x + .4
x
is
=
the
The
first of
these
unsym
metrical constituent.
The symmetrical
tion of condition, while the unsymmetrical constituent will give the solution. But the form of the solution will most frequently be simplified con
by
The unsymmetrical
is
constituent
will always be such that its equation of condition the equation of condition of
satisfied a priori.
Thus
A B
is (.1
x +
A B x =
B)(A B) = 0, which is an identity. By this method of considering symmetrical and unsymmetrical con stituents, equations which are indeterminate reveal that fact by having no unsymmetrical constituent for the solution. Also, the method enables
any term
us to treat even complicated equations by inspection. Remembering that in which neither x nor x appears belongs to the symmetrical
and
.1,
the
separation can be
made
(c
directly.
For example,
+ x) d + c d + (a + x) b
153
have as
its
equation of condition
c
d + c d + a
will
be
b
c .r c d
symmetrical constituent
is
Also, as
we
always the
The method does not readily apply to equations which do not have member 0. But these can always be reduced to that form. Plow it
unknown
will
The problem
of elimination
is
.r.
favorable circumstances, be eliminated from two equations, two terms from three, n terms from n 1 equations. But in this algebra any number
terms (and their negatives) can be eliminated from a single equation; and the terms to be eliminated may be chosen at will. The principles whereby such eliminations are performed have already been provided in
of
A B =
By
contains
all
.r
or
.r
which can
be derived from
x+
x
0.
x+
B .r =
is
= B
?/
A u
we have
(B u +
A
11)
B
is
(B u +
u)
= A B u
AB
= A B =
.r.
Hence
It is at
AB =
with the equation of condition for solution and with the symmetrical con
stituent of the equation.
741
If
,
n elements,
,T 3 ,
.
. .
x\,
0,
.r 2 ,
.r 3
xn
F(XI, # 2
xn )
the equation to
of the
F(x it x 2
.r 3
x n ).
is
By
If
one element,
(b)
= 0. .r, from any equation, /(.r) the theorem hold for the elimination
of k elements, x\,
X2
154 xk
A
. . .
$(.TI,
x Zt
...
.TI,
xk )
.T L
&gt;,
=
.
.
0,
.
then
,
it will
hold
elements,
equation, V(xi, x z
x ki xk +i)
...
,
0, for:
/(.ri,
By (Ml,
V(xi,
.12,
xk x k+ i) =
xz
x k )x k+i xz
,
.
+f
(xi,
x k )Xk+i.
And
the coefficients in
will
By
from
i
X k ) Xk+i + /
. .
.
(Xi,
XZ
Xk)~X k+ =
a *)
f(xi,
.T 2 ,
x k ) x/ Oi,
.
.r 2 ,
=
=
is
And by
to
&lt;(.TI,
643,
2*2.,
/(.TI,
a;*;)
.T 2 ,
x k ) X/
(.TI, or 2 ,
a**)
equivalent
if
...
0,
2,
where $
3,
is
the
coefficients
in/ be PI,
P P
etc.,
coefficients
&lt;
in/
will then the corresponding coefficients in be PiQi, P 2 2, PsQa, etc. Hence if the theorem hold for the elimina x k ) = 0, this tion of k elements, .TI, x Z) ... x k from $(.TI, x z
be Qi,
Qz Q
,
s,
etc.,
(f iftftis
.ddQ..
0,
where PiP^Pzor in /
QiQzQsi.
3&gt;,
and /
if
e.,
Hence
.
.
xz
x k from
,
xz
xk )
,
0, it will
elements,
i
Xi,
xz
...
x k x k+ i, from
xz
x k x k+ i)
,
0,
provided x k+ be the first eliminated. But since the order of terms in a function
is
any order of elements in the argument of a function, there is a normal form of the function, x k+ in the above may be any of the k + 1 elements in \F, and the order of elimination is immaterial.
i
(c)
(b)
theorem holds
for the of
elimination of any
number
any function
of these elements.
it is
By this theorem,
form;
Equate to
the product of
This
will
155
form of an equation with one member gives the resultant have that form.
742
especially
if
The complete
resultant of eliminating
,
n elements,
x\
x2
...
xn
...
...
etc.,
xn) x n ).
==
1,
is
the equation to
of the
sum
x n ).
2,
A
=
3,
1 is
,
equivalent to [F(xi,
x n )]
,
0.
And
$
by 641, [F(xi, x 2
that
will
if
...
x n )]
is
a function, $(xi, x 2
...
x n ), such
any
A/c.
coefficient in
be
Hence, by 741, the complete resultant of eliminating x\ x 2 from F(XI, Xz, ... x n ) = 1 is
9
...
xn
UA
But
[595] {
=0,
=
or
{ 11^1
=
1
114}
Z A.
Hence Q.E.D.
For purposes
tion
is
terms.
For
example
If all b is x,
c x,
x
ax
and no a
then
is x,
= =
0,
ax
+ b x
0.
Whence, by elimination, a
0,
or no a
is b.
The complete solu unknown may be accomplished by tion of any equation in more than one eliminating all the unknowns except one and solving for that one, repeating
Solution of Equations in
Such solution
will
be complete
because the elimination, in each case, will give the complete resultant which is independent of the unknowns eliminated, and each solution will be a
unknown, and complete, by previous theorems. How ever, general formulae of the solution of any equation in n unknowns, for each of the unknowns, can be proved.
solution for one
any equation in n unknowns is identical with the resultant of the elimination of all the unknowns; and this resultant is the condition of the solution with respect to each of the unknowns sepa
7 5
The equation
of condition of
rately.
(a)
If the
xn )
0:
156
x n ) be AI,
2,
3,
etc.
Then,
by 6 32,
II A c F(xi, x 2
and
[565]
...
xn)
JJ
A =
=
0*2,
is
F(x lt
02,
z) Xi,
And
.
[741]
,
UA
.
is
.
x2
x n from F(x\,
(6)
xn)
0.
If
.
the equation in n
.
F(XI, x 2)
of that
Xn)
0,
then by 71,
is
form, and
its
elimination
equation
of
condition and
elimination resultant of
its
F(x lf X 2
(c)
...
X n)
=0
the
The
unknowns
is
the
,
equation of condition
because
(1)
:
ing
all
(2)
The equation to be solved for Xk will be the result of eliminat the unknowns but Xk from the original equation; and The condition that this equation, in which Xk is the only
is,
by
(a)
and
(6),
same as
all
the
is
of the order in
it
will
they are
751
Any
its
x n)
0,
provided
of the
each
unknowns
and
any one
of the
unknowns;
,
. . .
let P],
2,
x n ) in which Xk
positive,
Qi,
Q2 Q
,
etc.,
is
negative.
The
solution then
II Q ex* c
2 P,
or
Xk
= II Q
+ u
2 P,
where u
is
arbitrary.
157
(a)
By (Ml,
to
for
F(x\,
.
.
.TO,
.
x, n )
=
=
is
equivalent
/(.ri,
.r n _i).r n
+/
3,
(.ri, .T 2 ,
.r n _i).r n
0.
in
/ be PI,
P P
2,
etc., in
be ^i,
z,
Qs, etc.
Then
in
PI,
P
Xk
is
2,
3,
etc., will
F
in
which
Xk
is
Q3
which
negative.
.T n _]).r n
If /(.TI,
.TI, .TO,
.
.T 2 ,
n _i, its
.r 2
,
. .
2 .r n ,
3 .T n
etc.
.TI,
o* 2 ,
And
. .
.
Hence, by 642,
/(.?!,
X2
X n l) X n + / (1,
^(.TI,
.TO,
. .
.
.T 2 ,
Xni) X n
=
a function in
is
equivalent to
x n i)
0,
where
is
which the
+
coefficients
etc.
are
(PiX n + Qix n ),
is
(P 2 x n +
Q2
.
.r n ),
(P 3 ^n
3^ n ),
^(iCi,
And
T2
...
.r n _i)
equivalent to F(XI,
.r 2 ,
xn)
,
. .
0.
.r n _i 741, the complete resultant of the elimination of 0*1, .r 2 = will be the equation to of the product x n i) from ty(xi, x 2
By
of its coefficients,
of the
form
P nr n + Q
x n
is
Hence, by 6 43,
By
H P^n+
Qr ^
is
And
for
[5951]
{II P]

T.Pis
(b)
immaterial, and
is a any order of the variables in the argument normal form of the function, x n in the above may be any one of the .r n _ ) and / (.n, .r 2 .T n ), and /(.TI, .i 2 variables in F(x^ .T 2
of a function there
.T n
of the
holds for
That a
is
any number
of
unknowns
another peculiarity of the algebra, due to the fact that from a single equation any number of unknowns may be eliminated.
158
As an example of the last theorem, we give the solution of the exemplar equation in two unknowns, first directly from the theorem, then by elimina
tion
and solution
for
(1)
x y+
x y +
ABCD
Provided this be
satisfied,
BDcxcA+C,
CDcycA+B,
(2)
x
y
= BD
+ u (A + C)
or
= CD
is
+u
(A+B)
x y+
x y +
(a)
C x y + D x y =
equivalent to
x) y
.r
(Ax
B x) y+(Cx + D
=
=
and
to (b)
(a),
(A y + C y) x + (B y +
D y)
Eliminating y from
we have
+
(Ax
The equation
Bx)(Cx
Dx) =
ACx + BDx
is,
of condition
with respect to x
then,
(AC)(BD) =
And
the solution for x
C),
is
ABCD
C).
BDcxc(A
or
(6),
BD + u(A
And (A
C)
= A
C
Eliminating x from
we have
y+
is,
(Ay+Cy)(By + Dy) = A B
The equation
of condition with respect to y
is
C D y =
then,
BCD
0.
And
CDcyc(AB),
Another method
or
CD + v(AB).
y
in
would be to solve
for y
and
for
terms of the
by substituting
and solve
for x.
By
a similar substitution,
then be eliminated and the resulting equation solved for This y. method may inspire more confidence on the part of those unfamiliar with
may
in other
algebras
is
required.
The
y
x y+
x y +
(C x +
D x)
+ u(A x +
159
The
solution for y
is
y
= (A
x +
B x)
v(C x
D x) =
(A +
C) x + (B +
D) x
(A x +
B
=
(C + u A) x +
7?
=
Hence
AC x + D x
B (D = 0.
+ u B) x +
C (A +
B D ex cA
it is
+ C.
to equations in
Theoretically, this
any number
of
unknowns:
7 52
practically,
all.
Any
...
xn)
f(xi,
x2
...
xn)
unknowns
Qi, Qi,
3,
as follows:
Q.a,
unknowns;
If 2
,
let PI,
.
. .
P P
2,
3,
...
be the coefficients in F,
and
J/i,
3/3,
AI,
A2 A
,
terms
in /, so that
and
and Q r and
negative.
The
solu
II
or
Xk
(Qr
N
(Qr
+ Q,
r)
Xk
C
+ U
r
(P r
M
^
+ ~P,  J/ r )
= II
r
N
i,
+ Qr
...
r)
(Pr
By
i,
713,
I
7 X.i
n)
;i 2 ,
.r n )
=
is
#2,
0,
where $
be coefficients of any two corresponding terms in coefficient of the corresponding term in $ will be
will
be
II
Or
(Qr
"N r
+ Qr
r)
C Xk C
(P r
M
P
3/ r )
= II
[04]
(Qr
N
Q r
r)
r)
+ H
(P r  J/ + P,
r
J/r)
And
753
(P r J/ r + P J/
r
= (P
J/ r +
P J/
r
r ).
equations in n unknowns, each of the form be regarded as simultaneous, is as follows: xn) 0, may F(XI, x, x n ) = 0, Let the coefficients of the terms in F 1 in the equation F (x lf .r 2 1 1 let the coefficients of the correbe Pi 1 , Pa 1 P3 1 QJ, Qs Qi
160
A
in
spending terms
F
Qs
2
,
in the
.
equation
(.ri, ,r 2 ,
.r n )
0,
be Pi 2
2 2
,
Pa
2
,
2
,
.
Q\~, Qz~,
Fm
z
in the
,
equation
F m (x
lt
....
is
Or
if
*
r
or Q, in
P\
the condition
is
And
the solution which n such equations give, on this condition, for any
Xk, is
as follows:
P
h
7i
...
0, in
be the coef
any one
,
of the equations
F =
is
which
xk
is
positive,
and
let
Qi
h
,
Qi
h
3
,
...
F =
h
0, in
which Xk
is
negative.
The
nizvic^cznw]
r
ft
or
**
= ![ tZ
r
Qr
wZ
r
III
h
Pr"]
By
.
.
642,
x n)
xn)
0, 0,
xz
coefficients in
2
&lt;
is
the
sum
of the
l
P
.
P, Fm
2
,
&lt;
will
h
be
P, +
1
P
Q
r
2 r
+ P, .,
or
2 Pr
h
and
1
,
if
2
,
Qr
.
Qr
2
,
.
.
F F
will
Fm
be
h
The equation
of condition for
0,
&lt;&gt;
0,
F =
1
0,
F =
2
...
Fm =
may
be regarded as simultaneous,
that
is,
the equation to
or
nizjvixiiizcr io h
r
And by
#n)
for Xk
is
II
S Qr
h
]
c TA c
161
%
595, [
h
And by
7 54
P\ =
h
P/.
m
,
F(XI, x 2
x n)
f(xi,
x2
...
xn)
l
,
may
be regarded as simultaneous, is as follows: Let the coefficients in F l 1 in the equation F 1 =f be Px 1 P, Pj 1 and let the QJ, Q Q. in the equation F = f l be coefficients of the corresponding terms in /
1 1
,
, ,
,
. .
.
Mi MJ, Ms
1
,
1
,
.
. .
Ni As AY,
,
let
,
terms in
2
,
in the
equation
F =f
2
be P^,
2 3
,
.
.
and
.
.
let
2
,
QS, be
M M
2
i
, ,
.
Q 22 Q 32
,
the coefficients of the corresponding terms 2 , 3 Ni m = m P m P m m be m m W in the equation F Pi 2 3 Qi Qz $3 f m let the coefficients of the corresponding terms in / m be Mi m z T m m AT m .... The condition then is A
.
A A
2
let
Ms 2 in F m
,
.,
,
and
Ms m
A
h
II
r
H (Pr
MS + ~Prh
MS)] X JJ
r
E
h
(Qr*
N, +
(? r
A/)]
=
represent
is
or
if
h
r
Fh
+
whether
P
=
or Q,
and
h
r
whether
M or A, the condition
B
h
r
II
IZ
m
(Ar
BS
A
h
r
)}
And
.TA,
such equations give, on this condition, for any h h be the coefficients is as follows: Let P r and r
Fh = f h
in
which a^
is
positive,
and
is
let
h
r
and
h
r
F = fh
h
in
which x k
negative.
The
solution then
or
a*
 II
By
Z
.
(Qr
A7/ +
9,"
AV)] +
.
E [II
a? 2 ,
713, ^(.TI,
. .
#2,
0,
ar n )
= / A (TI
is
...
n)
is
equivalent to
if
h Q and
r
i, .r 2 ,
xn)
where
h
r
Ph
r
and Jf r A be
h
coefficients of
and / h the coefficient * h Q r Nr + Q r AV, and if h h the corresponding terms in F and /
,
Fh
term
h
r
in
h
r
^ will
+
be
P h M h + P h
r r
r r
h
r
.
P
h
r
h
r )
= P
P/
M \
162 x n)
A
F(XI,
.1*2,
. .
f(xi,
xz
a n )
is
F
of
in the algebra,
developed as functions of
of simultaneous equations, in
any number
is
of
unknowns,
for
itself to
generaliza
Boole
General Problem.
problem of the algebra of logic. Given any equation connecting the symbols x, y, ... w, z, .... Re quired to determine the logical expression of any class expressed in any
way by
.
. .
y,
... in
t
z,
....
z,
We may
)
;
Given
f(x, y,
z,
.)
and
is
$&gt;(.r,
y,
.)
ty(w,
to determine
. .
.
in
.
terms of w,
.
....
This
if
is,
x, y,
for solution
=
V(w,
2,
by any number of equations, there The rule single equation equivalent to them all. = f(x, y, .) and .) may be stated: Reduce both t y, to the form of equations with one member 0, combine them .)
.
are connected
3&gt;(.r,
by addition into a single equation, eliminate x, 71, the form of equation with one member
form.
y,
is
and solve
for
t.
By
And by
572, the
sum
of
two equations with one member is Hence the single equation resulting
will contain all the data.
is
The
independent
of these,
of
t
for
will
in
terms of w,
Consequences
of
Equations in General.
word
ence to the manipulation of equations in this algebra may not be out of As compared with other algebras, the algebra of logic gives more place.
room
The
choice of
method must,
But
0,
possible expression
1.
a+b+
c+
gives
0, 6
0, c
0,
a+ 6
0,
a+
0,
of Thought, p. 140.
163
But a
will
. .
a+ b
1,
a+
==
not generally give a + b = 0, etc. Also, a == == 1 will not 1, but a + b generally give a == 1.
of equations, with
gives
(2)
which
it is
necessary to deal,
by first reducing each to the form in which and then adding. The various constituent equations can always be recovered if that be desirable, and the single equation gives
to a single equivalent equation,
one
member
Do
not forget
that
it is
possible so to
If
we have a =
combine equations that the result is less general and 6 = 0, we have also a = b, or a b = 0,
0,
according to the
mode
of combination.
But a +
is
two are
less
comprehensive.
A
fact,
general
desired
method by which consequences of a given equation, in any 12 and is, in terms, may be derived, was formulated by Poretsky,
a corollary of his
Law
of
We
this
law
may
ab + ab =
be formulated as the principle that if a = b, and therefore and a b + a b = 1 then any t is such that a b + a b c t
,
the
"zero
member"
"
b,
t,
and
is
contained in the
whole
Now
Law
if
x c
t,
ux c
for
member" of
the
of
tcy, then
tcy + v,
by the
and the
"whole
member" in
the
Law
v.
of
Forms may be
increased
Law
of Consequences.
76
If
b,
then
= (ab + a
b + u)
v (a
b + a
b) /,
where u and
are arbitrary.
[7
1
12]
If a
b,
then a b + ab
+
v (a
= =
and ab + ab =
(1
1.

Hence
(a b +
a b +
all
it) t
b + a
b) t
?/) t
+ v0t
t.
the possible consequences of the given equation. more general than the previous formulae of elimina
tion
for
t.
and
solution.
.r
B x =
0,
and choosing
.1
we should
ItAx + B x =
then
AB
(A x + B x + u)
AB
Since u and
fore
12
and
may assume
the value
0,
there
AB =
0.
Sept
lois, etc.,
Chap. xn.
164
A
this
is
But
AB
gives.
Letting
u =
0,
and
1,
we have
x+
A B = A B
Letting u
A B x
= A and
B,
we have
AB
And
so on.
AB + ABx
AB
be found that every one of the equivalents of which the law gives will be null.
it will
But
t,
we should
x+
x
0,
then x
= =
(A x +
B x)
x
Since u and v
may
= A
x,
or
x c
A
(1)
And
since
u and
But
if
B x c x,
(1)
then
(2),
1,
Hence,
and
v
B x = (B B cxcA.
Law
x
of
.T)
0,
or
cx
(2)
When u =
of
and
the
Law
Forms.
in the above,
= A
x+
B x
in the
which
is
The introduction
all
u and
of
v,
Law
of
Consequences
it
Law
Forms
so that
all
covers not
only
the nonequivalent
As the explanation which precedes the proof suggests, this is accomplished by allowing the limits of the function equated to t to be expressed in all possible ways. If a = b, and therefore, by the Law of
inferences.
Forms,
t
(a b
+ a b)
(a
6 + a
6,
b) t
t,
0, is
expressed as a b + a
b.
limit of
/,
is
expressed as
ab + a
its
In the
Law
limit, 0, is expressed as v
expression as
is,
expressed as a b + a 6 + u, that
ways which ab + ab; and the upper limit, 1, is all possible ways which can be derived
in all possible
from
its
expression as a b + a b.
t
(a b
+ a b)
(a
b + a
b) t
165
(a b
+ a b + u)
v (a
b + a
b) t
determines
its limits,
of
Consequences covers
all
possible
ways
covers
possible inferences
is,
The number
of such
inferences
of course, unlimited.
The number
expressible in
terms of n
member and the other member expanded with reference to n elements. The number of constituent terms of this expanded member will be 2 n and the number of combinations formed from them will be 2 2 Therefore,
,
".
since pi + pz + p$ +
=0
number
of
of consequences of
0,
n elements.
Poretsky gives the sixteen determinations of a in terms of the three elements, a, b, and c, which can be derived from
As one
13
is b,
ab =
b c
0, 0,
and
all 6 is c,
then ab +
a
b c
0,
and hence,
=
=
a
5 b
+ c)
(b
c)
)
of
Consequences.
Just as the
Law
of
Conse
inference from a
0,
1
;
of the fact
1,
then a b + a b + u
the inference
fl( a 5
will give
+ _a a b +  a
_fr)
= =
b
i}
can be expressed by taking advantage of the fact that if then ab + ab = 1, and if ab + ab + u = 0, then
thus get Poretsky
s
l&gt;
o.
We
Law
of v
of Causes, or as
14
it
would
Law
of Sufficient Conditions.
77
If for
some value
t
of
(ab + a b)
+ (a b + a b + u)
t,
then a
b.
[v
then
[71, 572]
a ^ +  a
1)
+ u ) t
= (ab + a
b) t
+ u i
13
14
98 /. Chap. xxm.
166
A
Hence
(a
b + a
b) t
t
=
+ a b)
t
t,
(1)
Hence
also [57]
v (a b
and
[49]
b)
t[v (a b
+ a b)]
(a
=
b)
t
(v
+ a b + a
tv + (a b + ab)
Hence
[5
72]
b + a
(
=
(t
(2)
By
(1)
and
(2),
a b + a
b.
b)
t)
a 6 + a
6.
Hence
[71] a
Consequences and the Law of Sufficient Conditions are more general than the Law of Forms, which may be derived from either.
of
equation can be found automatically. The only sense in which these laws give results automatically is the sense in
which they make it possible to exhaust the list of consequences or conditions And since this process is expressible in terms of a given set of elements.
ordinarily too lengthy for practical purposes, these laws are of assistance
principally for testing results suggested
"intuition
".
One has
u and
v will
V.
The
In this algebra, the assertory or copulative relations are = and c denial of a = b may conveniently be symbolized in the customary way:
801
4= b is
equivalent to
"a
"a
b is false
".
Def.
We
or
"a
lent to a b
b is
false.".
But
since a
b is
its
negative
may
b"
be represented by a b
equiva 4= a
by a b
c
It is less
b is false
since
"a
is
necessary to have a separate symbolism for not contained in is seldom met with in logic
in
true,
is
which case a
then Q
is
=
",
0.
"If
P
it.
is
true
there
is
another,
If
".
This
of
and
==,
(1)
(2)
"If
= = =
b,
then then
c
6,
= gives = d&ndh =
d",
".
also,
k",
"If
c 4= d,
then a
=(=
&lt;Z,
4= b
".
gives also,
"Ifc
then a
4=
b",
and
then
"If
4= k,
then a
b
4= b
(3)
"If
and
"If
c c
c 4=
d",
and
d,
then h
=
k",
gives also,
4= b
".
"If
and h
4= k,
d and h
4= k,
then a
167
= =
b is equivalent to c
d"
gives also,
"a
4=
b is equivalent
toe
4=
d".
(5)
"a
"a
=
&lt;7,
k,
".
..,"
gives also,
4= 6 is
equivalent to
Either
c 4=
&lt;/
or h
4= A
or ...
16
The
But the
from
"calculus
this algebra
without a
circle in
demon
every step.
We
by ordinary logic, which mathematics generally presupposes. In later chapters, 16 we shall discuss another mode of developing mathematical logic the logistic
method
which avoids the paradox of assuming the principles of logic in order to prove them. For the present, our procedure may be viewed simply
as an application of the reductio ad
absurdum
in
ways
in
matician
feels free to
make
Since the propositions concerning inequations follow immediately, for the most part, from those concerning equations, proof will ordinarily be
unnecessary.
Elementary Theorems.
sitions are as follows:
81
If
ac
4=
be, then a
4= b.
[21]
812
If
a+c
=t=
c,
then a
4= b.
[337]
813
4= b is
equivalent to a
4"
b.
[32]
814
a + b
4= b,
ab
4= a,
a + b
4= 1,
and a b
4=
are
all
equivalent.
[49]
815
If
a+b
[57]
x and b
4= x,
then a
4=
8151
If
=
=
and
b 4= x,
then a + b
4= x.
[57]
816
If
ab
x and b
4=
.r,
then a
4=
1.
[571]
15
"Either
...
or
..." is
Chap,
and Chap.
168
A
If
8161
and
b 4= x,
then a b
4= x.
[571]
817
If a
+ b
4=
and a =
and a
0,
then b
4= 0.
[572]
818
If
a b
4= 1
1,
then
b 4= 1.
[573]
4= 0
In
this, it gives
may
be combined.
Suppose,
a+ b
=
if
(a + b)(x
Hence ax + bx
But
x
[8
0,
then a x
and
0.
Hence
82
If
17]
ax +
x
4= 0.
4= 0,
then a + b
4= 0.
[572]
821
If
4= 1,
then
064=!.
4=
[573]
822
If
a 6
4= 0,
then a
and
b 4= 0.
[15]
823
If
+ 6
[45]
4=
1,
then a
4= 1
and
b 4= 1.
824
If a b 4
x and a
x,
then
b 4= x.
[12]
825
If a 4=
[19] If
ac6, then a
4=
a.
4= 0.
Hence Hence
826
a+ 6
4
if
and a c
6,
then a 6
[822] 6
is
0.
"Either
equivalent to
4=
or 6
4=
".
[572]
8261
as
4=
ai
+ a 2 + a3 +
".
4=
is
equivalent to
"Either
ax
4=
or a 2
4=
or
0, or ...
827
a 6
4= 1 is
equivalent to
"Either
4=
or 6
4= 1
".
[573]
8271
3
aia 2
1
+1
=
is
equivalent to
"Either
ai 4= 1
or
a2
4= 1
or
or ...
".
The
tions
difference
b
572 a +
and
0,
and
The
573 a
b
169
1 is
==
and
points to a neces
sary difference
inequations.
between the treatment of equations and the treatment of Two or more equations may always be combined into an
equivalent equation;
two
or
into
But, by 82,
a+
b 4=
is
a consequence of the
The laws
of the equiva
83
4= b is
equivalent to a b + a b 4
0.
[71]
831
4=
1 is
equivalent to a
=h 0
[712]
832
If
&(xi,
2,
...
x n ) and V(xi,
.r 2 ,
X2
.
Xn)
4=
is
equivalent to F(XI,
.T 2 ,
xn)
Ai,
4= 0?
where
3,
is
variables
f
and such
,
that,
if
z,
etc.,
be the coefficients in
$ and
Bi 2 3 etc., be the coefficients of the corresponding terms in ^, then the coefficients of the corresponding terms in F will be AiBi + AiBi,
,
A, B 2 + A z
B A B
2,
3
[713]
Poretsky
833
a
4=
Law
of
Forms
is
equivalent to
3=
a t+ at.
[715]
Or
in
834
equivalent to
=h
(ab + a b)
(a
b + a
b) t.
[716]
Elimination.
by the
reductio ad absurdum.
same theorems
84
It
x +
B x
4= 0,
then
[63\Ax + B x
841
If
A + B + 0. cA+B. Hence
[8
25]
Q.E.D.
,
.
. .
x n ),
170
be Ci,
A
C
2,
To, etc.,
and
if
F(x 1}
.r 2 ,
.r n )
=f=
0,
then
ZC=NO
[032]
F(x lf
.r,
a n )
cC.
Hence
[825]
Q.E.D.
Thus, to eliminate any number of elements from an inequation with one member 0, reduce the other member to the form of a normal function
of the elements to be eliminated.
The
elimination
is
then secured by
putting
=f=
the
sum
of the coefficients.
The form
of elimination resultants
It is
obvious that they will be analogous to the elimination resultants of equa tions as follows: To get the elimination resultant of any inequation, take
the elimination resultant of the corresponding equation and replace
= by
=f=
and x by +
A
a
is
is
is
represented by an equation:
"All
by a b
"
0,
No
"
is b
by
a b
0.
always the contradictory of some universal, any particular proposition may be represented by an inequation: "Some a is by a b =f= 0, "Some
b"
by a b =j= 0. The elimination of the "middle" term from an which represents the combination of two universal premises equation But elimina gives the equation which represents the universal conclusion.
is
not
b"
an analogous
logical
Two
particulars
give
no conclusion:
a particular conclusion
is
The drawing
of a particular conclusion
by 822.
. .
For example,
c
aisb,
a
is c,
ab =
a
c
=J=
0. 0.
a b
a b
c
0.
Some
+ a b
4= 0.
c =h 0.
abc
[817] [822]
Some
"
fr
is c.
6c4=0.
Solution
"
of
an Inequation.
of
An
inequation
may
be said to have a
inequation one
member
4=
is
85
x +
x
equivalent to x
A
x +
x.
[723]
851
i.
A
to
x+
x
4=
is
equivalent to
or
"Either
x
or
",
e.,
"Either
B ex
is false
xcA
is false
".
[721
171
"
"solutions"
"
B ex
is
is
false
does not
x".
mean
"
is
excluded from
is
it
means only
is
false"
"
not
wholly within
Either
Bex
false
or
xcA
does not
determine either an upper or a lower limit of x; and limits x only by ex cluding B + uA from the range of its possible values. Thus "solutions"
of inequations are of small significance.
Consequences and
Sufficient Conditions of
an Inequation.
By
Poret
sky s method, the formula for any consequence of a given inequation follows from the Law of Sufficient Conditions for equations. 17 If for some value
of
of
v,
x (a b
+ a 6)
+ (a b + a b + u) t
then a 852
=
If
b.
then
^= v (a b
+ a b)
+ (a b + a
+ u)
t,
where
and
v are arbitrary.
[77]
The formula
lows from the
an inequation similarly
If
fol
Law
t
of
b,
then
+ a b + u)
v (a
b + a
b) t
where u and
853
If for
are arbitrary.
of
some value
t
3=
(a b
+ a b + u)
v (a
b + a
b) t
then a
==
b.
[76]
an Equation and an Inequation. If we have an equation in one unknown, x, and an inequation which involves x, these may be combined in either of two ways: (1) each may be reduced to the form in which one
System
of
member
either.
and expanded with reference to all the elements involved in Then all the terms which are common to the two may, by 8 17,
is
be dropped from the inequation; (2) the equation and this value substituted for x in the inequation.
may
C
be solved for
x,
86
If
0,
then .1
x +
BD .r
4= 0.
.r
172
A
[572] If
Ax + Bx =
and
0,
then
0.
and
B x =
0,
and hence
A Cx =
Hence
[8
BD
x
+
=
B
17] .4
Cx
x
4= 0.
The
for
one reason
x +
the equation
A Cx+BD
mode
x
is
not equivalent to
is
B x =
0.
Nevertheless this
of combination
861
x+
B x =
is,
C x + D x + B D 4=
may
x
4=
be regarded as simultaneous
of x
AB =
is
0,
AD)x + (BC + B D)
is
[723]
A x + B x =
equivalent to x
= A
x+
x.
Substi
C (A x + B
or (A
[84]
x) +
(A x +
4= 0.
is
B
x)
4=
C + A D) x+(B C + BD) x
condition of this inequation
4=0,
or
(A+B) C+(A
+ B) D 4= 0. x + B x = +A =
requires that
^45 =
0,
and hence
A + B = A and B
if
B.
Hence
AB =
0,
inequation reduces to
A C + B D
4= 0.
this condition is
C+
4= 0.
But
and
0, since
A C c C
and hence
hence
to
"
[825]
if
B
4= 0,
BDcD
is
equivalent
Either
"Either
C 4= A C 4=
or
4=
",
and A C + B
".
4=
is
equivalent to
or
B D
4=
Hence the
system
is
expressed by
AB =
And
[8
5]
and
A C + B
4=
(AC + AD)x
(BC + BD)x
x, in
the form of
173
"SUBTRACTION"
AND
It is possible to define
"subtraction"
and
a.
"division"
in the
algebra.
Let a
a.
b be
x such that
+x
And
let
be y such
that
by =
they are worth, and should not be admitted to the system. In the first place, it is not possible to give these relations a general meaning. We cannot have in the algebra: (1) If a and 6 are elements
in
A",
then a
b is
b is
an element
in
in
K; nor
If
(2) If
then a
it
an element
b
if
K.
a.
by
Thus
y a and
But
b
then for some y a, then, by 22, a cb y and, y be so chosen that a cb is false, then a b
a
:
b is b
an element,
if
cannot be any element in K. To give a b a general meaning, be required that every element be contained in every element
:
it
would
is,
that
that
x, in
all
elements in
for
be identical.
Similarly,
6
if
b
a.
be an element,
A, then
some
x, it
+x
=
if
But
if
+x
a,
then,
by
22, b
xca
and, by 521,
b
:
be a.
Thus
It
a and b be so chosen
in
A".
that b c a
is false,
then a
Again, a since a + a
1
and a
are ambiguous.
1
=
=
1,
the value of
is
satisfied
1,
lent to x + a
which
equivalent to
(x + a)
1
=
x.
a x
it
And a x =
that, since a a
:
is
equivalent to a c
:
Similarly,
might be expected
y,
or a
= 0, the value of a would be unambiguously a. = y = 0, is satisfied by any y such that y c a. a y
But
and
y c a are equivalent.
Finally, these relations can always be otherwise expressed.
of a
:
The value
b is
a.
is
equivalent to
a b y + a b + a y
=
0.
The equation
condition,
is
of condition here is a b
And
+u
(a
+ b)
= ab + uab, where u
is
undetermined.
b
The value
is
of a
b is the value of
x in the equation,
+x
a.
+x
equivalent to
a
+ a x + a b x
174
The equation
condition,
is
of condition here
is,
a
0.
And
a b + v a
a b + v a
6,
where
v is
undetermined.
of the expression,
In each case, the equation of condition gives the limitation of the meaning and the solution expresses the range of its possible values.
CHAPTER
There are four applications
III
which are
commonly
relations.
considered:
The
may
be
made
to continuous
and
discontinuous segments of a
or to continuous
in a plane, or to continuous
dimensions.
Segments
of a line
and discontinuous regions in space of any and regions in a plane have both been
used as diagrams for the relations of classes and of propositions, but the
application to regions in a plane gives the
for
obvious reasons.
And
since
it
is
we
I.
denote continuous or discontinuous regions in a given plane, or in a 1 represents the plane (or circumscribed circumscribed portion of a plane.
portion)
itself.
is
every region.
a,
i.
e.,
nota.
b.
is
com
0.
mon
The
over.
to a and
"sum",
a and 6 do not
"overlap
then a b
the nullregion,
a+
determining a +
denotes the region which is either a or b (or both). In the common region, a b, is not, of course, counted twice
a + b
a b + a b + a
b.
This
is
= b, signifies that a and b denote the same equation, a ac6 signifies that a lies wholly within b, that a is included or region. contained in b. It should be noted that whenever a = b, a c b and b c a.
of arithmetic.
The
Also, a
c a holds always.
.
Thus the
relation
is
analogous not to
&lt;
in
arithmetic but to
175
176
A
While the laws of
however
those regions
may
their distribution
classes.
with Euler
"All
"
earlier
than Boole
algebra.
"
is
6"
wholly within a
intersect;
circle b;
No
is b
by two
a
is
circles,
which nowhere
"Some
is
6"
and
"Some
not
6"
by
intersecting circles,
sometimes with an asterisk to indicate that division of the diagram which The defects of this style of diagram are obvious represents the proposition.
:
All a
is
No
is
cuto;
Some
a
is &
x*&gt;c^\
^^^\
Some
a
is
not 6
FIG. 1
the representation goes beyond the relation of classes indicated by the propo
sition.
In the case of
"All
is
&",
the circle a
b is
falls
a",
as to suggest that
we may
infer
"Some
"No
not
b"
is
not valid.
The representation
b",
of
a
is
is
similarly suggests
"Some
which
also
unwarranted.
With these
is
dia
grams, there
is
no way
of indicating
null/
But
Yet
mentioned,
and
to
others
which are
Euler
still
diagrams were in general use until the invention of Venn, and are
specifically to represent the relations
The
principle
diagrams is that classes be represented by regions in such relation to one another that all the possible logical relations of these classes can be
indicated in the same diagram.
for
any
then be specified
null.
the diagram initially leaves room and the actual or given relation, can by indicating that some particular region is null or is notis,
That
Initially the
or
1.
diagram represents simply the "universe of discourse", For one element, a, 1 = a + a. 2 For two elements, a and b,
1
(a
+ a) (b + b)
a b + a b + a b + a b
See Venn, Symbolic Logic, Chap. v. The first edition of this book appeared before Schroder s Algebra der Logik, but Venn adopts the most important alteration of Boole s the nonexclusive interpretation of a + b. original algebra 2 See above, Chap, n, propositions 48 and 592.
177
a, b,
and
c,
(a
a b c + a b c + a b
+ a b
b c
+ a b c
c
+ a
+ a b
+ a 6 c
Thus the
"
universe of
discourse"
r(
for
any number
of elements, n,
must
correspond to a diagram of 2
expansion of
1.
If
divisions, each representing a term in the the area within the square in the diagram represent
a
FIG. 2
ab
the universe, and the area within the circle represent the element a, then the remainder of the square will represent its negative, a. If another element,
6, is
may
be repre
sented by another circle whose periphery cuts the first. The divisions, (1) into a and, a, (2) into b and b, will thus be crossdivisions in the uni
verse.
If
all
the possible
ab, those things which are a but not b; ab, those things which are 6 but not a; a 6, those things which are neither a nor b. The area which
represents the product, a
b,
will readily
be located.
We
have enclosed
by a broken line, in figure 2, the area which represents a + b. The negative of any entity is always the plane exclusive of that
For example, (a b + a 6), in the above, two divisions of the diagram, a b + a 6.
If it
entity.
will
be the
sum
of the other
c,
it is
two
one
part which shall be in c and one part which shall be outside c. This can be be accomplished by introducing a third circle, thus It is not really necessary to draw the square, 1, since the area given to the
figure, or the
whole page,
is
may
omitted,
178
area outside
A
all
number
of elements
abc
FIG. 3
If
a fourth element,
d,
be introduced,
it is
by a
circle, since
each previous subdivision into two parts one part in d and one part outside d. But this can be done with ellipses. 3 Each
in figure 3 so as to cut
abcd
FIG. 4
3
of junction.
This helps somewhat in drawing the diagram, which is most easily done as follows: First, draw the upright ellipse, a. Mark a point at the base of it and one on the left. Next,
179
one of the subdivisions in figure 4 can be "named" by noting whether it is in or outside of each of the ellipses in turn. Thus the area indicated by
6
is
is
a b
c d.
With a diagram
of
c,
four elements,
requires care, at
first,
These can always be determined with certainty by developing each term of the expression with reference to the missing ele
a
c
d,
+b
+ d.
ments. 4
Thus
ac +
= ac
a b
c
(b
+ b) (d + d) +
b c
b
c
d (a + a) (c + c)
d + a b
c
d+ a
d + a b
d + a
b c
d + a
b c
d + a b c d 4 by the
The terms
divisions
of this
sum,
numbered
10,
9, 5,
14,
11,
15.
Hence ac + bd
one
is
the region
With a
it
little
practice,
may
identify such
b
Such an area as
2, 3, 6, 7,
+ d
is
more
and
by inspection:
comprises
it is
we
let e
The
principle
all
of the
"square
(figure 6) is the
same
as
Venn
s:
represents
FIG. 5
draw the horizontal ellipse, d, from one of these points to the other, so that the line con necting the two points is common to a and d. Then, draw ellipse 6 from and returning to the base point, and ellipse c from and returning to the point on the left. If not done in
this
way, the
first
180
number
of elements.
No
diagram
is
a
a
ab
a b
FIG. 6
frequently needed.
slightly the square
are those
made by modifying
Figure 7
in figure 6. 5
b&lt;
b
h
e
e
h
FIG. 7
gives,
by
this
for five
and
We
give
also the
diagram
(figure 8)
since this
not easy to
in the algebra
may
be specified in a
diagram of the proper number of divisions, has already been explained. We must now consider how any asserted relation of elements any incluwe
See Lewis Carroll, Symbolic Logic, for the particular form of the square diagram which adopt. Mr. Dodgson is able, by this method, to give diagrams for as many as 10 terms, 1024 subdivisions (p. 176).
6
181
b,
or
any equation, a =
relation, or
b,
or inequation, a
=)=
may
be repre
sented.
such relations, can be completely in these diagrams by taking advantage of the fact that they specified
Any such
any
set of
or to the can always be reduced either to the form of an expression = form of an expression 4= 0 Any inclusion, a c b, is equivalent to an equa
tion,
a 6
one of the
is
equivalent to
equation can be
any inequation of the Thus any of the form ab + ab =H O. 8 whatever can be specified by indicating that some region
Similarly,
is null,
= 0.
=
{
Oj, or
is
notnull,
0}.
in
We
can illustrate
this,
useful,
is 6,
is c,
a cb,
b cc,
and
All b
= =
0.
0.
6 7 8
182
A
have here indicated
striking
it
We
by
that
(figure 9) that a b
lines).
is
not
is
null
Similarly,
w e have
indicated
by striking out b c (with vertical lines). Together, the two have eliminated the whole of a c, thus indicating that a c = 0, operations
all b is c
or
"All
is c
".
FIG. 9
For purposes
of comparison,
9
we may
derive this
same conclusion by
algebraic processes.
Since a b
and and
b c
= =
= =
therefore, a b c
[572] a b c + a b c
a c
(b
+ b)
a c.
The equation in the third line, which combines the two premises, states The last exactly the same facts which are represented in the diagram.
equation gives the conclusion, which results from eliminating the middle term, b. Since a diagram will not perform an elimination, we must there
"look for"
the conclusion.
illustration of this kind
is b,
:
One more
Given: All a
a b
b c
and No
b is c,
is
= =
0.
0.
The
first
premise
by
striking out
be (with
vertical lines).
c,
To
clusion
9
ac =
0,
or
"No
is
c".
orem
in
Throughout this chapter, references in square brackets give the number Chap, u by which any unobvious step in proof is taken.
of the the
183
method
null or
of
diagramming,
we
is
is
notnull.
The
may
require that
null
some regions be
not.
null, or
that some
be notnull, or that
some be
and others
Consequently, even
we cannot presume
nullregions,
we can
say,
struck out
may
is
be notnull.
If,
we wish
given region
there
is
definitely notnull
mark
a
in
b
we must
this purpose,
by some
b.
distinctive
That a
0,
may
But here
say, a,
a
it
b c.
the diagram involve more than two elements, a b c and b, and c, the region a b will be divided into two parts, a b 4= 0, does not tell us that a b c =}= 0, and Now the inequation,
tell
does not
then,
us that a 6 c
4= 04=
it
It tells us
only that a b
+a
b c
==
0.
If,
we
wish to indicate a b
by an asterisk
that
is all
in the region a b,
we
c.
shall not be
warranted
in putting
we know.
Hence
it is
convenient to indicate a b
of a b
4=
by placing an
by a broken
line,
these regions
ab =0
FIG. 11
We
shall
show
is
best interpreted
by
0.
an inequation;
"Some
is
6",
members, by ab
184
A
we have:
is b, is c,
Suppose, then,
Given:
All a
ab =
a c
b is
c",
0.
and Some a
H= 0.
is
The
conclusion,
"Some
one of the two connected asterisks must remain the whole region a b c + a be cannot be null. But one of them, in a b c, is struck out in indi
cating the other premise, a 6
0.
Thus a
b c
=}=
0,
and hence a
c 4= 0.
FIG. 12
The
by
may
be represented
by and remembering that any region which is neither struck out nor occupied by an asterisk is in doubt. Also, the separate subdivisions of a region occupied connected asterisks are in doubt unless all but one by
are notnull,
of these connected asterisks
left in
made
specifically
occupy regions which are struck out. And doubt by a given set of premises might, of null or notnull by an additional premise.
In complicated problems, the use of the diagram is often simpler and more illuminating than the use of transformations, eliminations, and solu
tions in the algebra.
tions, the
diagram gives
"see"
it)
at a glance.
we
shall give
interpretation of the algebra for logical classes has already been 10 explained. a, b, c, etc., are to denote classes taken in extension; that is 10 Chap, n, pp. 12122.
The
185
the objects
denoted by some classconcept. Thus if a b, the concept of the class a not be a synonym for the concept of the class b, but the classes a and b may
must
nifies
consist of the
acb
sig
b.
that every
product",
member
b,
also a
member
of the class
The
"
bers of a
and members
denotes the class of those things which are both mem of b. The "sum a + 6, denotes the class of those
",
members
of a or
members
of b (or
members
of both).
Various concepts
may
an empty class "immortal men", "feathered invertebrates", "Julius Caesar s twin," etc. but all such terms have the same extension; they denote nothing existent. Thus, since classes are taken in extension,
denote
there
is
but one
nullclass, 0.
Since
it
is
every
x,
the nullclass
bers of
ex, we must accept, in this connection, the convention that is contained in every class. All the immortal men are mem
class, since there are
"universe
any
the
no such.
",
"every
thing
is
",
of discourse
or simply the
This term
if
But
A"
it
may
be defined as follows:
a n be any
member
then the
"a
and
of
discourse" is
the class of
If
all
the classes,
x,
such that
blind"
is
an
is
"The
is
stars"
and the
"blind"
do not
belong to the
same universe
of a, a,
is
of discourse.
The negative
of discourse:
members
+ a
==
in
in the universe
0,
"Nothing is
a".
both a and
nota",
and
1,
"Everything is
either a or
6,
Since inclusions, a c
equations, a
b,
and inequations, a
4= b,
repre
sent relations which are asserted to hold between classes, they are capable
of being interpreted as logical propositions.
And
would
algebra
transformations,
eliminations,
and
It
solutions
capable
of
hardly be correct to
say that the operations of the algebra represent the processes of reasoning
from given premises to conclusions: they do indeed represent processes of reasoning, but they seldom attain the result by just those operations which are supposed to characterize the customary processes of thinking.
In fact,
it is
makes
186
A
The representation
of propositions
by
inclusions, equations,
and
in
that the algebra represents relations of extension only, while ordinary logical
The
(1) (2)
"
universal affirmative,
a
All a
is
6",
= =
b,
flc6,
(3)
v b,
where
v is
undetermined,
(4) a b
0.
The only
its
possible
doubt concerns
(3)
v b,
where
v is
undetermined.
But
may
be
demon
strated as follows:
[7 1]
v b is
equivalent to a(v
b)
+ a
v b
v b
0.
v b.
But a
Hence
(v b) + a v b
[572]
if
a (v + b) + a
a v + a b + a
vb, then a b
0. v
(i.
And
if
b}
e., v
a),
v b.
These equivalents
(1)
is
6"
The a
is
and
6 s both.
b.
(2) a
(3)
contained in
class a
is
every
member
of a
is
also a
member
of
The The
identical with
of the
class
6.
(4)
members
of 6
is null.
If
"All
is
6",
we
not only may it hold when a 0, but it always = 06, c 6, and 06 = 0, will be true for every when a = 0. = b for some value of element b. And is always true for v = 0. = means that a has no members, it is thus clear that the algebra Since a
shall discover that
holds
"0
v"
requires that
"All
is
6"
exist.
The
that
"All
ambiguous on this point. We should hardly say sea serpents have red wings, because there aren t any sea ser
is
pents";
11
yet
of the novel
who
asserts "Whoever
187
my
dead
body".
who
The
difference of the
two cases
concept
portal"
"sea
"having
red
wings",
who
enter the
as conceived
by the hero
We
an
empty
class in
of the other
when
the relation
is
one of intension.
empty
class is
class,
In
"For
every
we must
(1)
that ordinary language frequently concerns relations of intension, and is usually confined to relations of intension where a null The law does not accord with the ordinary use of language. class is involved.
and
(2)
This
is,
its
truth, for
is
it is
a necessary law
It
an immediate consequence of
"All
the principle,
are also
"For
every
y,
cl",
that
is,
members
of
any
class, y,
members
things".
this last
every
class.
is
The
involved
is
both because
in
relations predicated are frequently thought of in extension and because the relation of classes in extension is entirely analogous to their relation in
or the class
all cases,
is
involved.
But the
be confined to extension.
in the algebra as stating
In brief:
"All
is
6"
whenever a
is
an empty
"No
"All
is
6"
should be true.
The
do not
proposition,
6",
is
is
represented by
ab =
of a
"Nothing is
both a and
or
"Those
"No
members
and
of b both,
exist".
Since
is
6"
is
equivalent to
"All
is not6", it
may
v is
also be represented
by
a b
b, a
c6, 6ca,
or a
=
is
v b,
where
undetermined.
no discrepancy
has been a problem to representation of particular propositions not clearly conceived the symbolic logicians, partly because they have
The
188
and have
which hold
in intension only.
is
6"
that
it is false
when the
"All
class a
6",
"Some
be so interpreted a is will
6"
a
"
is
for
is
is
6"
is
true whenever a
0,
0.
But
diffi
6"
if
Some a
be true
when a =
we have two
does not accord with ordinary usage, and (2) "Some a is will not, in that case, contradict "No a is For whenever there are no
6".
members
"Some
of a
is
6"
is
6"
(a b
0) will be true.
Hence
a
is
if
0,
then
"Some
is
b"
and
"No
6"
is
b"
solution of the difficulty lies in observing as a relation of extension requires that there be some a
of the class a exist.
"Some
The
member
is
6"
precisely because
whenever a =
"All
0,
"All
will be true.
But
"Some
is
b"
is 6,
and members
a
of a
exist".
To
is
interpret properly
"Some
a
6".
is
6",
it
the contradictory of
"No
is
Since
0,
is
6"
is
interpreted
by
a b
are
0, "Some
is
6"
will
be a
=(=
that
is,
"The
class of things
which
members
of a
and
of b
both
is notnull".
It is surprising
of particular propositions.
is
y"
where
v is
undetermined, and by u x
v y,
where u and
v are
undetermined.
Both
and
for the
same reason: An
1
"undetermined"
element
may have
the value
or the value
or
any other
value.
all.
Conse
are
quently, both these equations assert precisely nothing at both of them true a priori, true of every x and y and in
They
But
all cases.
For
them
to be significant,
u and
0.
in that
case they are equivalent to x y =J= 0, which is much simpler defined laws which are consonant with its meaning.
and obeys
well
Since
we
are to symbolize
"Some
"All
is
b"
by a b =
==
0, it is clear
that
its
contradictory,
is
not
6",
will
be a b
0.
To sum
follows:
up, then:
A. All a
is b,
a b
E.
I.
No
is b, is b, is
= 0. = 0. ab
a b
b,
==
O.
Some a Some a
0.
=}=
not
a b
0.
12
Each
12
of these four has various equivalents: See Chap. 11, 4 9 and 8 14.
189
0,
b,
a +
= =
1
1,
a + b
a, a
b,
equivalent.
E. a b
all
=
4
1
0,
=
4=
a b, a + b
1,
a +
a,
ac6, and be a
s
are
equivalent.
I.
a b
0, 0,
a &, a + 6 4
1, 1,
and a +
4 a are
4=
all all
equivalent. equivalent.
O. o 6
=}=
4=
6,
a +
and a + b
a are
The reader will easily translate these equivalent forms for himself. With these symbolic representations of A, E, I and O, let us investigate
the relation of propositions traditionally referred to under the topics,
"The
Square of
Opposition",
and
"Immediate Inference".
That the
is
two
at once obvious.
a b
is
=
is
is
true, then a 6
is
false;
if
a b
is false,
then a b
=)=
true.
and a
b 4= 0.
The
relation of contraries
Two
may
"contraries".
This relation
is
A
a
and E.
ex
tension:
it fails
a nullclass.
If
0,
That
is
to say,
"All
if
no members of a
6"
extension,
is
and
"No
is
6"
exist, then from the point of view of But if it be assumed are both true.
members
(a
==
0),
(b
if
+ b) a
a b + a b.
Hence
4= 0,
then a
+ a b
4= 0.
[817] If a b + a b 4=
and a b =
0,
then a
b 4= 0.
4= 0.
(1) (2)
And
if
a b + a b
=h
and a
b
:
0,
then a b
We may
(1)
is false.
If
members
of the class a
and
all
is b,
then
"No
is
6"
(2)
is false.
If
there are
members
of the class a
and no a
is b,
then
"All
is
6"
By
As has
been pointed out, this relation fails to hold when a = T ever a 4= 0. e can read a b 4= 0, in (1) above, as
But
a
it
holds
6"
when
"Some
is
instead
"
of of
"
No
13
is
is
b b
is
is
false",
and
a b 4= 0, in (2), as
"Some
is
not
6"
instead
All a
false
".
We
then have
190
If If
A
(1) (2)
"
members members
"
and
all
is b,
is b,
then some a
is b. is
and no a
then some a
not
b.
Subcontraries
both
may
be true.
Traditionally
"Some
is
6"
and
"Some
=j=
is
not
6"
are subcontraries.
false,
"
But whenever a =
and a b
0,
a b
a
"
4=
and a 6
0, it
are both
and the
is
is
Some a
(1)
When = is
holds.
is
Since a b
1
=
we
Some
a
not b
is
false",
can read
(1)
is
and
(2)
above:
of a,
If there are
b.
members
and
"Some
is
6"
is false,
then some a
not
(2)
If there are
members
of a
and
"Some
is
not
6"
is false,
then some
is 6.
To sum
class
"square
of
opposition"
hold in the algebra whenever the subject of the four propositions denotes a
When
The two
universal propositions
both true, and the two particular propositions both false. The subject of immediate inference is not so well crystallized by tradi
tion,
and
class
for the good reason that it runs against this very difficulty of the without members. For instance, the following principles would all
:
is
6"
gives gives
"Xo
b is b
a".
"No
b is
b is
a"
"All
is nota".
"All
nota"
nota"
gives gives
gives
(6)
"Some
"Some
b is
nota".
"Some
b is
"Xo
nota
is
6".
Hence
is
6"
"Some
"
nota
is
6".
Xo cows
implies
gasteropods
implies
(a)
"Some
noncows are
"
inflexed gasteropods":
"Some
"Xo
mathematician
(b)
These
infer
an empty
"Some
class;
b is
and because
that
"All
b is
nota"
gives
nota".
The
inference of any
proposition
from
the
class
proposition or the class denoted by its predicate ("not6" regarded as the is a class which has members, predicate of "Some a is not
6")
191
The
tive
is
"
conversion"
and
is
ab =
a
is
"
b a.
6",
Xo
4= 0,
a b
0,
gives b a
0,
"No
b is
And
a
is
"Some
a 6
gives b a
4= 0,
is
4= 0, "Some b is
"Some
a".
Also,
"Some
not
of
6",
a 6 4
0,
gives 6 a
not6
is
a".
The
"converse"
simply the
"converse"
What
are called
"obverses"
i.
e.,
same equation,
is
14
.
Since x y
=
And
"
is
Xo x
a
",
a b
is
0,
is not6".
since a b
equiva
lent to a (b)
0,
"Xo
is
6"
is
equivalent to
"All
is not6".
A
S
made by
putting
(subject)
and
and P between
its
two constituent
which the dia
The
ways
in
(Mverl Prop,
Converse
SP
FIG. 13
and thus suggest all the immediate inferences which are valid. For example, the arrow marked "converse" indicates the two terms which will appear in the converse of the given proposition and the
gram may be
read,
In
this diagram,
we must
and
192
wish
If
add the
"No
by the given proposition. And we may qualification that the classes, S and P, have members. is and S and P have members:
P",
we
SP
=
5.
4=
4.
SP
7.
6.
FIG. 14
Reading the diagram of figure 14 in the various possible ways, we have: 1. No S is P, and 1. Some S is not P. (According as we read what
is
is null,
or
what
is
that S
2.
3.
P
All
is
notnull.)
is
notP,
and
2.
Some S
is
notP.
All
4. 5.
6. 7.
Wanting.
8.
193
Similarly,
"All
is
P",
and S and
0,
f.
P
0,
have members:
S P =
P
5.
4=
FIG. 15
(figure 15),
we have:
All
is
P,
and
1.
Some S
is
P.
2.
No S
is
notP.
3.
Wanting.
4.
5. 6. 7.
notP.
S.
8.
notP
notS,
and
8.
Some notP
is
notS.
The whole
trivial.
so simple as to be almost subject of immediate inference is Yet in the clearing of certain difficulties concerning nullclasses
The
14
194
The premises
of
this information as
is
term.
This
And
it
elimination
is
very simple.
O.
15
eliminating x from
of a syllogism
is
A x + B x
may
of
AB =
is 0,
universal,
in a single equation
one
member
which
"middle"
term.
For example:
No
is y,
x y
z
0. 0.
All z
is x,
x
Combining
these, x y + z x
x,
is
"No
=
=
or
0.
0.
Eliminating
y y
is
Hence the
valid conclusion
z",
"No
z is
y".
Any
may
a
b,
also be symbolized so
b cc,
c6 and
then
ace".
By
this
a and
"If
a c
then b
ca",
are
some
For example:
No
is y, is x,
x c y.
z
All z
ex.
"No
Hence
There
is
cy, or
"No
z is
y",
and y cz, or
is
z".
no need to treat further examples of syllogisms with universal conclusions they are all alike, as far as the algebra is concerned. Of course,
:
and of getting the con but the above are the simplest. When a syllogism has a particular premise, and therefore a particular conclusion, the process is somewhat different. Here we have given one
ways
equation
(1)
1=0) and
one inequation
{+0j.
We
proceed as follows:
expand the inequation by introducing the third element; (2) multiply the equation by the element not appearing in it; (3) make use of the prin If a + b =f= and a = 0, then b 4= ciple, to obtain an inequation with
"
0",
literal
member;
term"
from
this inequation.
1 1 in
195
x z x y
z.
0.
=
=}=
0.
Some x
xy
is y,
0.
[15]
[8
+ z) = x y z + x y Since x z = 0. x y z =
x y
(z
Hence, x y
+ x y z
=f=
0.
17] Since
xy z + xy z
yz
4= 0,
4=
and
y
is
.r
0,
therefore x y z
=f=
0.
Hence
[822]
or
"Some
a".
An
exactly similar process gives the conclusion for every syllogism with a
particular premise.
have omitted, so far, any consideration of syllogisms with both premises universal and a particular conclusion those with "weakened" in the third and fourth figures. These conclusions, and A A I and E A
are
all
We
They
If
which
now
familiar:
we add
to the premises
class
term
is
example,
AA
All x
xy =
x z
0, 0.
==
0.
All x
is z,
Since x
=k Q,
xy + x
y
4= 0,
and
since x y
0,
x y 4
0.
Hence x y
Since x z
+ x y z 4= 0, x y z
(2),
0.
(1)
0.
(2)
z
=f=
By
(1)
and
xy
0,
and hence y
0,
or
"Some
is
z".
Syllogisms of this form are generally considered valid because of a tacit assumption that we are dealing with things which exist. In symbolic
reasoning, or
is
rigorous,
made
If
explicit.
An alternative treatment
we take
of the syllogism
the two premises of any syllogism and the contradictory of its three conclusion, we have what may be called an "inconsistent triad" that if any two of them be true, the third must be false. propositions such For if the two premises be true, the conclusion must be true and its con17
See
"On
the Algebra of
Logic",
in Studies in Logic by
University, ed.
by
Peirce;
We
196
A
And
false.
if
tradictory false.
if
e.,
the conclusion be
false,
and
premise must be
As a consequence, every inconsistent triad corre Any two members of the triad give the
For example:
Inconsistent Triad
1.
All
a:
is is
y
z
is
2. 3.
All y
Some x
not
z.
Valid Syllogisms
1.
Allan s?/
All
1.
M\
is
y
is is
2.
All y
is 2 is is
2.
.
.
is
3.
. .
All x
is 2.
Some x Some y
not
z
z.
3.
.
not
Some x Some x
not 2 not
y.
in
give a particular conclusion, since these really have three premises and
are not syllogisms, the inconsistent triad formed from any valid syllogism
will consist of
particular.
will
one premise be particular, the conclusion will be particular, and its contradictory will be the second universal. Representing universals and
if
particulars as
sistent triad,
we have done, this means that if we symbolize any incon we shall have two equations = Oj and one inequation 0}.
{
={=
And
as a conclusion.
must give the contradictory of the This means that the contradictory of the
{
Oj
particular particular
must be
a x + b x
0,
all
to be thus obtainable.
will
be of the form a x
=
is
and
a b
x
0,
resultant of a x + b x
= =
whose contradictory
:
be a
0.
triad will
0,
x
0,
a b
=f
where a and
b are
positive or negative,
and x
is
any
positive term.
The
from
validity of
its
proposi
it
by changing
{=0}
to
197
And
may
be got by
how
Thus,
if
No
and
is y,
is y,
x y
z y
= = =
All notz
The
the triad
is
x z
it
4= 0.
x z
0,
or
is
"All
is
z".
(Incidentally
may
in
no one of
and
yz =
0,
no conclusion
is
same inconsistent
triad.
We can, then, frame a single canon for all strictly valid syllogistic reason
ing:
The premises and the contradictory of the conclusion, expressed = 0} or 4= 0}, must form a triad such that symbolic form,
{ {
in
(1)
(2)
and one particular The two universals have a term in common, which
{
0}
{4= 0).
is
once positive
The
particular puts
4=
com
mon term
in the
two
universals.
few experiments with traditional syllogisms will make this matter clear The validity of this canon depends solely upon the nature to the reader. three terms, three propositions and upon the law of
of the syllogism
elimination resultants,
"If
.r
+ b x
0,
then a
0".
Reasoning which involves conditional propositions hypothetical argu same process, if we first ments, dilemmas, etc. may be treated by the A reduce them to syllogistic form. For example, we may translate cases in which where x is the class of is J5, then C is by "All x is All cases in which .1 in which C is Di. e., A is B, and y the class of cases
"If
D"
y",
"
is
B
"
is
D".
And we may
translate
"But
is
B"
by
",
where
z is
"All
"
cases in which
all
is
is
D.
is
But
B.
is
Z)."
"Hence all
And
in some other arguments of this type are reducible to syllogisms to Thus the symbolic treatment of the syllogism extends similar fashion.
all
198
A
also.
them
more
easily
by another interpretation
The
its
an instrument of reasoning,
liberating us
Many who
object to
do not
realize
It
may
this fact.
We
shall offer
two
each in a
different
logic.
undreamt
of in formal
The
of these turns
whose
significance
was
first
It is characteristically
human
we
In fact, so
ingrained
some
will
correct
enough
If
the difference
is
preference.
Mars whose
coincided with our own, so that any conclusion which seemed valid to us would seem valid to them, and vice versa, but whose psychology otherwise
differed
from ours, these Martians might have an equally fundamental We can point out one such which
the
as
we regard
equality
their reasoning
might
diagrammed
+ ab c
is
as in figure
16,
since
accbc(a + c).
This relation
(Note that
(ac/b)
and
a be
precisely the information which we discard in drawing a syllogistic conclusion from two universal habitually If all a is b and all b is c, we have premises.
represents
a b
=
+ c) +
and
(a
6 c b c
=
0,
Hence a b
18
(c
+ a)
the Relation of the Logical Theory of Classes and the Geometrical Theory of Points/ , Proc. London Math. Soc., xxi, 14782. But the use we here make of this relation is due to Josiah Royce. For a further discussion of Kempe s triadic relation,
"On
199
0.
a b c + a b c
and
(1) is
(2)
a b
+ a
b c
"All
= =
a
a c
0.
is
(b
+ b)
a c
0,
c";
(2) is (ac/b).
all
Perhaps most
the information
that precisely
of
us would
feel
FIG. 16
what we overlook
curious
"illogical"
is
And
be capable of understanding and of getting for themselves any conclusion which a syllogism or a hypo
prejudice, they
still
would
thetical
argument can
stateable in terms of
and many others which are only very awkwardly a is our formal logic. Our relation, a c6, or
give,
"All
6",
would
(Qb/a)
is
equivalent to
a 6 +
a 6
=
"
a b
hence (Oc/a) This would, in fact, be only a special case of a more general principle which is one of those we may suppose the Martians would ordinarily rely upon
in
Barbara would be
(06/a)
and
(Oc/6),
".
for inference:
principle holds,
"If
(xb/a)
and
(xc/b),
then
(arc/a)".
That
this
general
is
proved as follows:
(xb/a)
(xc/b)
is
x a 6 + x a x
b c
b
c
is
+ x b
= =
=
or?
0,
+ x a 6 c
0.
200
[572] This equation
(1)
.r
A
is
a b c + x a b c + x
c c (b
= =
(2)
x ac
(b
+ b) + x a
+ b)
x a c +
c
a*
a
0.
x a
b c
+ x a b c + x a b
+ x a b c
b c (x
= =
(1) is (xc/a), of
a
+ x) + a b
c
c (x
+ x)
a
b c
+ a b
0.
which our
(2) is a
language to express.
Thus these Martians could deal with and understand our formal
by treating our dyads
as triads with one
logic
member
null.
In somewhat
gisms with a particular premise, dilemmas, etc., are all capable of state ment in terms of the relation (ac/b). As a fact, this relation is much more
powerful than any dyad for purposes of reasoning. Anyone who will trouble to study its properties will be convinced that the only sound reason for not using it, instead of our dyads, is the psychological difficulty of
common
but
Our differently placed, and a third member which is different in the two. But the operations of the algebra are inde attentionspan is too small.
pendent of such purely psychological limitations that is to say, a process too complicated for us in any other form becomes sufficiently simple to be
clear in the algebra.
logic
"
formal
"
cannot attain.
This illustration has indicated the possibility of entirely valid nonWe shall now exemplify the fact that by
are not so remote from familiar processes of reasoning,
modes which
any
number of nontraditional conclusions can be drawn. we make use of Poretsky s Law of Forms: 19
x
This law
is
For
this purpose,
is
equivalent to
x +
t
tx
"All
=
a
t0
is
6"
evident enough: if x = 0, then for any t, tx = t1 = t, and = t. Let us now take the syllogistic premises, + 0, while
and
"All
6 is
c",
and
see
what
from them by
this law.
All a
All b
19
is b,
is c,
ab =
bc =
0. 0.
201
Combining
these,
ab +
b c
And
[3441]
(ab + bc)
= 0. = (a
(a + b)(b +
c)
=
Let us make substitutions, in terms of
a+ b
a, 6,
a b + a
t
+b
c.
and
c,
for the
of this formula.
= (a + b) (a b + a c + b c) = abc+abc + bc = be
is
+ a b (ab + b c)
What
is
either a or 6
which
6
is
both
and
c".
c.
This
is
is
and
all
b is
Other such
conclusions
a
What
is
is
either a or c
which
is a, b,
and
c, all
three, or
b or c.
b + b c)
(a
b)
c
That which
or
is
is b
with what
is c
The number
"All
from
of
the premises,
is
6"
and
b is
c",
limited only
c,
by the
this
number
a, b,
and
of course, arbitrary.
By
method, the
is
number
of conclusions
entirely
unlimited.
In concluding this discussion of the application of the algebra to the more involved logic of classes, we may give a few examples in which problems
than those usually dealt with by formal logic are solved. The examples chosen are mostly taken from other sources, and some of them, like the
first,
Example
I.
20
(a)
The
financial
committee
shall
shall be
chosen from
among
(c)
(b)
No
one
be a
member both
on the
of the general
and
on
be
Xo member
committee
shall
financial committee.
2, p.
331.
202 Let /
g
/
A
= member = =
" "
of financial committee.
"
general
"
library
fcg,
(gl)cf,
or or
/ g
0.
0.
fgl =
// =
0.
We
is
redundancy here.
(c)
In
by
by
horizontal,
by
oblique.
(a)
and
(c)
I.
To
simplify the
and
(c)
in a single equation:
g)
[591]
= f g +
f g
+ fg
+ f g
=fg+(f+f)gl=fg
is
+ gl
0.
And
Thus the
(a
}
[572] this
and g
0.
The
committee
shall be
chosen from
among
the general
committee.
(6
)
No member
of the general
committee
shall be
mittee.
203
Example
a
is
2.
21
The members
s
in three
ways
as
or not, as 6
or not,
and as
c s
or not.
It is
c s
made up
is
precisely of the a s
and the
s.
How
Given:
a c + a
b (c
c.
To
6 c
solve for
b c.
c.
c.
22
+ c) = Hence, b c + b c
a c + a
c
c a b + a
b.
The
c s
comprise the a
6 s
and the
6 s
s.
Another solution of
+ ac to the form
[7
1]
{
this
a c
0}
a c + a
equivalent to
6 (a c +
a
c)
+ b
c
(a c
+ a
c)
And
[6 4]
(a c + a
c)
+ a
c.
c
Hence, a
b c
+ a
6 c
+ a 6 c + a 6
c
0.
We
a b + a
b,
FIG. 18
relation of a,
6,
and
c,
stated
c.
by the premise
is
we have
21
also a
b c + 6
of
Venn
s, first
printed in an article on
"Boole
System
of
Logic",
Mind,
22
(1876), p. 487.
intelligible
if
to.
204
A
3.
23
Example
If
x that
is
is
not a
is
the same as
6,
and a that
is
not x
is
the same as
c,
what
x in terms of
b
a, 6, c
and
c?
.r.
Given:
= ax and
To
solve for
.r.
[71] 6
=
6
a
.i
is
equivalent to
(a
.T)
+ a b x
(a
.r) 6
+ a 6
a:
=
And
c
a b + b x + a 6
.r
(1)
=
c
a x
is
equivalent to
(a x)
+ a c x
(a +
.T)
+ ac x
= a c
Combining
(1)
x + a c x
(2)
and
(2),
a b + a c + (a b + c) x + (b + a c) x
(3)
Hence
[5
72]
(a b + c) x + (b + a c) #
(4)
6 c
(5)
The
solution of (4)
(b
is
+ a c) c # c (a 6 +
c)
And by
(5),
c)
(a 6 +
=
x
(a 6 +
c)
+ 6
(a
b)
c + b c
=
Hence
[22]
a c + b
(c
+ c)
+ a c
b + a c.
FIG. 19
2S
i,
14.
205
This solution
is
verified
combines
all
the data.
(3),
which
b) c
This also
is
verified
by the diagram.
Example
4.
24
What
is
whom,
Let
asserts: s
s
=
p
p.
is
equivalent to
s(tp)+stp =
s
s t
p + s
(1)
(2)
cp
is
equivalent to
p =
t
And
(2)
is
equivalent to s t + s
=
=
(3)
and
(3)
p + s
(4)
(1)
represents
.4 s
assertion,
and
(4)
represents
s.
The
difference between
FIG. 20
the two
is
that between s tp
=
t
and *
s
t
=
t
0.
s
24
p + s
p
206
A
difference
is,
The
then, that
way
spread not having points nificantly assert this, for he has denied the existence of any space not having
".
stp = 0, while A does not. It is tpcs, "Threethis issue, stp = But B cannot sig as elements, is space
asserts
s
points as elements.
Both assert
p.
The
space, while
has
left
threeway spread has points as elements and is open the possibility that there should be threeway
Example
5.
Amongst the
enough to sink a
unless
it is
objects in a small
useful.
boy
which he regards as
either
But
all
And
heavy enough
to sink a fishline or
fishline,
not metal.
And
the
which he regards as
useful,
Specifically
what has he
in his pocket
which
he regards as useful?
Let x
= y = z = w =
bits of metal,
bent objects.
xy *
x z c w,
or
x z w
or
y w c (z + .i ), or zy ex ID,
x y z w = x y z + y z w =
z
and w,
w+x
w + x y
z
w+
x y z
w
=H
(w + w) + y
z
w
(x
+ x)
z
=
w =
x y z
w + x y
z
w + x y
z
w + x y
w + x y zw + x y
terms of the inequation appear also in this equation, with the exception of x y z w. Hence, by 8 17, x y z w 4= 0. The small boy has
All the
207
some bent
useful.
bits of
fishline,
which he considers
This appears in the diagram (figure 21) by the fact that while
FIG. 21
some subdivision
It
of x y
must be
is null.
appears also that anything else he may have which he considers useful may or may not be bent but is not metal.
Example
6.
25
The annelida
and
either
And
all
naked or enclosed
in a tube.
relation in
with respect to the possession of red blood, of an external covering, and of a vertebral column.
of the premises)
Let a
= = s n = = i = r =
/
annelida,
softbodied animals,
naked,
enclosed in a tube,
invertebrate,
etc.
t
and a cs (n + t), with the implied condition, n eliminate a and find an expression for s t.
Given: a
i
0.
To
25
208
a
A
=
i r is
equivalent to
(i r) a + a
i r
a i + a r + a
i r
(1)
cs (n+t)
n+
s
t)
is
0
s + n
t.
Hence,
as + ant =
(1)
(2)
Combining
and
(2)
and n
0,
a i + a r + a
i r
+ a s + a n t + n
(3)
Eliminating
(i
a,
by 7
4,
+ r + s +
7i 2
+ n
t) (i
+n
t)
= n
26
i r
s +
i r
n
t
The
is
i r
cs.
And
solution for
[53]
tisirnctcn.
s
t
Hence
irnc.stc.n, or
= irn+un, where u
is
un
determined.
The
FIG. 22
26
iv, "Symmetrical
of
an Equa
tion".
209
in a double system of circulating vessels and a body with an undetermined additional class (which may be covering, together This solution may be null) of other animals which have a body covering.
by the diagram of equation (3) the square formed by the two crossed
verified
(figure 22).
In this diagram,
is
rectangles.
The lower
half of this
qualification, n, in
n c s
t,
is
necessary.
i
senting a portion of
r (n)
which
is
not contained in
Example
7.
27
"All
Demonstrate that from the premises a is either b or c is no conclusion can be drawn which involves only two
"All
a",
c",
and
of the
classes, a, 6,
and
(b
c.
Given: a c
c)
and
a.
To prove
is
a c
(b c
c) is
is
equivalent to a 6 c
0.
And
= 0. Combining these, a b c + a c = 0. = 0. Eliminating a [74], (b c) c = 0, which is the identity, = 0. Eliminating c, (a b) a = 0, or c + a c) a c = a c = 0, which is the second Eliminating b, (a
ca
equivalent to a c
premise.
Example
8.
A
green,
or
more
and
blue,
And
all
oddnumbered blue
and
all
the even
numbered balls which are not both red and green, are on the table.
scribe the balls not
De
on the table.
Let
e r
g
t
Given:
27
(1)
r b g
= = = = = =
evennumbered, e
spotted with red,
= oddnumbered,
on the
table.
See
De Morgan, Formal
15
210
r
A
(2)
(3)
6
0.
[eb + e(rg)]ct,
x,
or
(e b + e r + e g) t
0.
To
find
an expression,
such that t c x, or t x
t.
Such an expression
t
with
b,
and
g,
and eliminate
t,
null terms.
(An alternative
is
method would be
t
to solve for
(r
briefer.)
t(e + e)
g + e
r b
r
g +
r b g + e r b g + e
b g
(4)
r b g + e
b g + e r b g + e r 6 g + e r b g)
From
(1), (2),
and
(3),
t (e b
r +
g +
b + r b g)
(5),
(5)
Eliminating from
t
(4)
terms involved in
g),
t (e r b g
+ e r b
or
t
(e r b
g + e r b g)
all
on the table are evennumbered and spotted with colors or oddnumbered and spotted with green only.
three
g...
211
is
by oblique,
9. 28
(3)
by
horizontal.
Example
Suppose that an analysis of the properties of a particular class of sub stances has led to the following general conclusions:
1st.
b are
property
2d.
or the property d,
is
present also; but they are not jointly present. b and c are combined, the properties
That wherever the properties a and b are both absent, the proper also; and vice versa, where the properties
and d are both absent, a and b are both absent also. Let it then be required from the above to determine what may be con cluded in any particular instance from the presence of the property a with
respect to the presence or absence of the properties b and regard to the property d.
c,
paying no
Given:
a b c
(c
d + c
d).
To
= =
0.
0.
But [64] (cd + cd) =cd + cd, and (a d + a d) = a d + a d. Hence we have, a b (c d + c d) =abcd+abcd = Q and b c (a d+ a d) =abcd+abcd = ()
(3) is
(4)
(5)
equivalent to
(c
a b
+ d) +
(a
b)
c d
=
Combining
function of
(4), d, (5),
a b
(6),
+ a b d + a c d + b c d
result the
(0)
and
form of a
(a b c + a 6 + a b c + a b
c)
d
c
+ (a b
28
+ a c +
b c
+ a b c + a b
c)
d
Franklin, loc. cit., pp. 5161, Venn, Symbolic Logic, Chap, Logik: Vol. I, Dreizehnte Vorlesung.
For furfher problems, see Mrs. Larldxm, and Schroder, Algebra dcr
212
A
Or, simplifying,
(_ a
fc
d + (a b
d,
+ a c +
b c
+a
b c)
d
Hence
[7
4]
eliminating
I)
(_ a
I)
c ) (a
b
+ ac +
b c
ab
c
c)
= ab
ab
cac
(b + c).
always present when c is present and b absent, and when ever a is present, either b is absent or c is absent. The diagram (figure 24) combines equations (4), (5), and (6).
The property a
is
FIG. 24
is
As Boole correctly claimed, the most powerful application of this algebra to problems of probability. But for this, additional laws which do not
Hence we omit
it.
Some
thing of what the algebra will do toward the solution of such problems will be evident if the reader imagine our Example 8 as giving numerically the
proportion of balls spotted with red, with blue, and with green, and the
quaesitum to be
"If
what
r
is
it will be spotted with all three colors? that it w ill be with green?" The algebra alone, without any additional laws, spotted answers the last question. As the reader will observe from the solution,
all
ta*ble
213
in
our postulates,
"prod
b,
we
have another interpretation of the algebra. Since a+b is the negative of is false that a and b are both false", or a 6, a + b will represent
"It
"At
least
"
6, is
true".
It
a+
b,
Either a or
or
"
Either a
"Either
is
true or b
or
true
".
But
this
is
some
in
what misleading,
algebra
is
since
...
..."
frequently
denotes,
is
For instance,
"
we should hardly
parallels
meet at
men
are
mortal".
We
"Either
or
But
Parallels
and
"All
men
are
is
a true proposition.
by +
and
between them.
in
"Either
a or
6",
we must bear
mind that no
6,
no relation of
"relevance"
or
is "logical import",
intended.
"a
The negative
false".
is
It
"contrary"
of a as well,
a be
"All
men
are mortal
",
then
"
Xo men
are mortal
"
But if the contrary as well as the contradictory be denoted then a will be an ambiguous function of a, whereas the algebra by a, 29 requires that a be unique.
should be a.
The
and
its
interpretation of
and
1 is
most
easily
made
clear
by considering
a,
The
prepositional sign,
may
equally
well be taken to represent the class of cases in which the proposition a is a b will then represent the class of cases in which a and b are both true,
true; a, the class of cases in
which a
is false,
and
so on.
The
"
universe", 1,
"actual"
Thus a =
"a
represents
"The
cases in which a
"a
true are
all
cases",
or
is
true in point of
cases,
fact",
or simply
"a
is
true".
Similarly
or
"a
is is
the class
of
no
It
and a
will
mean
is
true in no
case",
false".
May
not
a, b, c, etc.,
and sometimes
ace"?
false,
such as
is
Monday"
is
shows an
May
29
214
A
and these be not
1
true,
all
And
should not
be read
"a
is
always
"a
com
is true"?
The answer
But symbolic
is
made and
made
it
in his chapters
on
30
"Secondary
Propositions".
logicians
have come to
distinguish between assertions which are sometimes true and sometimes In the sense in which "Today is Monday" is false and propositions.
false, it is called
a propositional function
the BooleSchroder Algebra as a logic of propositional functions. the logic of propositional functions
is
this algebra,
and
in the
second place,
it is
much more complex than much more useful to restrict the a H= 0, then a = 1, and law
"If
=^ 1,
then a
0",
tions
false.
we
consequences of this law, which holds for proposi We need not pre
:
sume
this
law at present
is
throughout to propositions. But we shall either always true or never true if a proposi
:
true at
all,
it is
always
==
true.
Hence
"a
means
is
or
"a
is
always
either
true"
And
means
"a
is
or
"a
always
b,
false".
The
is false
true".
relation a c
since
it
is
equivalent to a b
",
=
"If
0,
may
is
be read
"
It
is
that
is
true and b
like
is false is
or loosely,
true, then b
But acb,
and
==
a+
b,
"relevance"
=
or a connection of
"Christmas
is
import".
Suppose a
that
"2
+2
"If
4"
b =
holiday".
We
Yet
4,
then Christmas
is
is
holiday".
it is false
+2
not a
holiday":
in this
example a b
called
is
true,
and
";
hold.
This relation, a c
b, is
"material
implication
we most frequently
"a
6"
interpret
"implies"
as a relation of intension.
But
acb
common
true,
30
case in which a
with our usual meaning of is true but b is false does not occur.
b will
a c b holds, and a
is
then
Laws
not be
false,
though
it
may
be irrelevant.
Thus
"material
215
of ordinary
"implies"
logic: a
c6
implies
6"
holds;
it
also holds in
some cases
which
"a
implies
6"
The
so simple,
and so
resembles
will suffice.
sitions,
We
give
logic of conditional
propo
Example
If
1.
is
B,
is
C
B.
is
D.
(1) (2)
And A
Let x
= A
is B,
= C
is
D.
are
:
xcy,
x
or [49] x + y
or x
==
1.
1,
=
1
is
0.
[57] Since
x + y
"
=
C
and x =
0,
1.
1 is
the conclusion
".
Example
(1) (2)
2.
If
is
B,
is
is
D.
But C
not D.
Let
= A
(1)
is
B y
= C
is
D.
.re?/, or
x + y
=
1
1.
(2)
0.
[57] Since
x + y
=
"A
and y =
0,
r
or
1.
x
1 is
the conclusion
is 1? is
false",
"A
is
not
Example
(1)
If
3.
(3)
and
or
(2)
if
is
F,
is
H.
is Z).
Let
w = A isB;
(1)
(2) (3)
31
isF;
= G
is //.
"Material implication" is
more at length
in
Chap,
iv, Sect,
i,
and Chap.
v, Sect. v.
216
Since w + y
A
=
1,
?y,
yz =
1.
x + y
1.
#+3 =
is
the conclusion
if
Either
is
Z)
or
6?
is
//".
This dilemma
be diagrammed
0,
in
(2)
yz =
(3)
10
y
0.
In figure 25,
z
is
struck
FIG. 25
lines,
yz with
is
vertical,
is
H&gt;
# with oblique.
That
either x or z
evident.
Example
(1)
(2)
4.
.1 is 7?
Either Either
Either
or
is
not
F.
7).
is 7)
is Z?
(3)
or E or E
is is
not
2
/?.
Let x
= A
(1)
is /*;
= C
is 7);
= E
0.
is
^.
(2) (3)
= =
0.
By
+ #
(2),
(3
Hence by
+ 3) = x + x + yz = 1
z
And by
(3),
x y
0.
+ y 3 = 1. = x + y z (x + x) = x + x y Hence x + x y z = x = 1.
?/ z
+ x y
z.
give the categorical conclusion "A is 5", indi cating the fact that the traditional modes of conditional syllogism are by
no means exhaustive.
217
Example
5.
32
If
matter
is
is
2.
If
gravitation
is
necessarily absent,
is
not subject to
any presiding
3.
intelligence,
vacuum
is
necessary.
4.
5.
vacuum
is
matter
not a necessary being. a necessary being, the world is not subject to a presiding
is is
necessary, matter
intelligence.
Let x
= Matter is a necessary being. = Gravitation is necessarily present. y z = The world is not subject to a presiding w = Motion exists. = Gravitation is necessarily absent.
t
intelligence.
A vacuum
are
f),
:
is
necessary.
(1)
(2)
xc(y +
t
or
xyt =
0.
0.
(3)
(4)
(5)
And
and absent,
0.
:
xy
From
between
t
+ tz
w+
it
ij
v + v x + x z + y
first,
=
any
(7)
be required,
being",
to discover
collection
x, "Matter is
a necessary
it
and
y, "Gravitation is necessarily
present".
For
this purpose,
is
sufficient to discover
of the four, x y
0,
x y
0,
x y
0,
or
.T
y
0,
the relations which state any implication which holds between x, or .r, and ?/, or y. This can always be done by collecting the coefficients of
y,
x y, x
y,
and x
(7),
such as equation
32
comprehensive expression of the data, and finding which of them, if any, reduce to 1. But
y, in the
are supposed to be
metaphysics.
218
not necessary,
because the inspection of the equation representing the data readily reveals
such a relation.
From
Hence
If
if
(7), [572]
vx + vy =
0.
(v
[15] v
x y + v x y
+ v) x y
x y
is
0,
or x c y, y c x.
matter
is
gravitation necessarily present, matter is not a necessary being. Next, let any connection between x and w be required. Here no such
is
relation
easily to be discovered
by
inspection.
Remembering that
if
0,
then a b
and a b
(y t +
t
=
;
From
(7),
+ y v +
z +
+ z + y
t)
wx
w x
(8)
t)
Here the
coefficient of
wx
y+
reduces to
v,
1,
for [585],
t
i)
and
y
t
+ z
+ z
coefficient
)
is
But
[590]
(y t + y +
0, or
+ y + t + v + z + y t. + z + y t = l+v + z + yt
1.
Hence w x =
c x, x c iv.
7
IV
y
{
FIG. 26
219
None
being;
1.
is:
"If
motion
exists,
matter
not a necessary
matter
is
exist".
Further conclusions, relating other terms, might be derived from the same premises. All such conclusions are readily discoverable in the dia
gram
of equation (7).
is
more convenient
for such
problems than the transformation of equations in the algebra. Another method for discovering the implications involved in given data is to state the data entirely in terms of the relation c and, remembering
,
that
b
"If
then
ace",
as well as
"acb
is
equivalent to
ca",
any connection thus revealed between the propo Although by this method it is possible to
exists, the
danger
is
relatively small.
IV.
The
relatively unimportant,
Schroder Algebra, and requires more extensive treatment in order to be of We shall, consequently, confine our discussion simply to the service.
explanation of this interpretation of the algebra.
A relation,
etc.,
taken in extension,
is
the class of
all
which
of"
That
is,
the relation
is
"father
the class of
all
those couples,
(x\y)&gt;
such that x
father
the relation
R to
R is (.r; R y. The extension of a relation is the class of things y, We must distinguish between the class of couples relation.
the class of all couples
y) such that x has
and the
symmetrical
and
.r
R y commonly
from y
R x.
The
(x; y)
"
product
",
x S, or
such that
all
xRy and
couples
;
the class of
R + S, will be x S y are both true. The "sum .r R y and y) such that at least one of the two,
of R, R, will be the class of couples
x S
y, holds.
The negative
is false.
0;
y) for
which x
Ry
(t\
The
couples
If
the class of
u) for
which
Ru
is
true,
is
220
which
Sw
is
true
is
then
R
t
It is this
sents.
Thus
R =
and
u,
such
that
Ru
is
true
R to
anything.
Similarly,
the universalrelation,
course).
the class of
all
The
for
inclusion,
RcS,
which x R y is true is also such that x S y is true; or, to put it otherwise, that the class of couples (x; y) for which x R y is true is included in the
class of couples (u; v) for
which u S
v is true.
reading of
RcS
$".
"
is
The presence
of the relation
R implies
the presence of
the relation
signifies
R =
RcS
and S c R,
that
(x; y) for
which x
v is true.
Ry
is
true
is
which u S
It is
obvious that
all
all
the propositions, of
and S are
true,
relations (that
is, if
there
is
such that x
Ry is
and a
such that u S
K
is
true),
then
R xS is a relation (that is, z) w R z and w S z are both true). If R and S be such that there is no couple (w; z) for which w R z and w S z both hold, then R x $ is the nullrelation,
there
is
such that
i.
e.,
12
The
which x
Ry
and x
Ry
both hold
is
simply the class of couples for which x Ry holds. 1 3 The class of couples denoted by R x S is the same as that denoted
by
SxR
14
(x
y)
such that
xRy
y,
and
xSy
all
The
class of couples (x
y) for
which
xRy,
x S
and x T y
hold
i.
is
e.,
which x
Ry
is
for
For every relation, R, there is a relation R, the which x Ry is false, and R is such that:
161
If
the relation
is
RxS
Sy
is
is
null (that
is,
if
there
is
no couple such
is,
that x
Ry
true and x
is false), is
then
R xS = R
(that
the class of
Ry
true
;
Ry
and
221
162
If
R xS  R and RxS = R, then R = e., if the class of which x R y and x S y are both true is identically the class of which x R y is true, and if also the class of couples for which
i.
x x
R y is true and x S y is false is identically the class of couples for which R y is true, then the class of couples for which x R y is true is null. 171= 0 e., the universal class of couples is the negative of the
i.
18
R+S =
(RxS)i.
e.,
(x\y)
such that
Ry
and x S
true
is
which x
Ry
false.
19
for for
RxS =
R Ry
which x
e., if
which x
if
and
x
the presence of
R
is
(x; y) for
which x
Ry
which
Ry
33
true.
Chap,
iv, Sect. v.
CHAPTER
We
IV
calculus of propo
its
sitions"
or calculus of
"material
implication",
and with
extension to
We
shall
modes
by
of
two methods
side
side.
procedure takes the BooleSchroder Algebra as its foundation, the elements of this system as propositions, and adds to it a interprets
first
The
postulate which holds for propositions but not for logical classes.
result
is
The
Algebra",
because the
and
if
=t=
0,
1.
For any x, if x == 1, then x = 0, This Two Valued Algebra is one form of the
where x n
is
calculus of propositions.
The extension
&lt;px
propositions of the
{
form
,
a* 3 etc., gives the calculus of prepositional functions. composed of x II and 2 functions have a special significance in this system, and the relation
,
a&gt;,
of of
Hx
(&lt;px
c\j/x), is
particularly important.
In terms
in the last
can be established.
This
is
which
simpler and otherwise superior to the TwoValued Algebra, then pro ceeds from this to the calculus of propositional functions and formal impli
cation,
and upon
of
the
"calculus
It is especially
and the difference of method, two procedures, should be understood. Too often they appear to
a, b, p, q, etc., represent propositions, and a x b or a b represent the joint assertion of a and b, then the assumptions of the
If
the elements
summary
See Schroder, Algebra der Logik: n, especially Fiinfzehnte Vorlesung. is contained in Schroder s Abriss (ed. tiller), Teil n.
An
excellent
222
223
and p = 1, is true". Since and 1 are unique, it follows that any two propositions, p and q, such that and q = 0, p = or such that p = 1 and q = 1, are also such that = q. p = q, in the p
sent
"p
BooleSchroder Algebra will all be found to hold for propositions, as was 2 As was there made clear, p = explained in the last chapter. will repre
is
false",
"p
"truth
value",
not an equiva
its definition,
P+
p+q
"At
(P q)
q
is
false",
is
the denial of
"p
is
false
and
q,
or
it
is
the proposition
least
is
true",
p+q
may
be read loosely,
"Either
p
is
true or q
is
true".
The
p and
q should
be true
is
not excluded.
equivalent to p q
pc
= p and
to
p q =
0.
;;
material implication.
We
the section.
"It
is
false
For the present, we may note simply that p c that p is true and q false". It may be read
(materially) implies
q".
means exactly
p
is
"If
true, q
is
true",
or
"p
With the
all
Schroder Algebra are true for propositions. Hence all the theorems will also be true for propositions. But there is an additional law which holds
for propositions:
p =
"The
(P
=
p
i)
is
proposition, p,
this that
is
equivalent to
true
".
It follows
immediately
from
P = (P
"p
=
".
1)
(P
0)
is
equivalent to
is
false
p = (p =
1),
and hence
(P
1)
is
(P
0),
and
(p
0)
=
is
(p
1)
is
=
"p
1 is
false
equivalent to p
=
==
0",
and
"
p =
is
false
equivalent
to
1".
Thus the
is
calculus of propositions
a twovalued algebra:
every proposition
either
or
1,
We
may, then,
proceed as follows:
2
However, many
and
of the theorems, especially those concerning functions, eliminations, no importance in the calculus of propositions.
224
which were given in Chapter II may be regarded as already established in the Two Valued Algebra. We may, then, simply add another division of propositions the additional postulate of the TwoValued Algebra and the
additional theorems which result
from
it.
orems
in
Chapter
.
II
was numbered
8,
we
of
this section 9
The
901
additional postulate
is:
p,
(p
1).
:
And
902
(p
q) is
equivalent to p
of 901,
==
q&gt;
As a consequence
(p
we
shall
=
==
1)
1
and
and
=
0.
0).
p*
It follows immediately from 901 that the TwoValued Algebra cannot be viewed as a wholly abstract mathematical system. For whatever p and 1 may be, p = 1 is a proposition. Hence the postulate asserts that
any element,
interpretation
p, in the
system,
is
a proposition.
But even a
necessary
be abstracted from in one important sense no step in need be allowed to depend upon this interpretation. This is the proof
may
procedure we shall follow, though it is not the usual one. It will appear shortly that the validity of the interpretations can be demonstrated within
the system
itself.
we
shall indicate
previous propositions by which any step in proof is taken, by giving the number of the proposition in square brackets. Theorems of Chapter II may, of course, be used exactly as if they were repeated in this chapter.
91
p = (p =
[901]
0).
p = (p =
1).
And
[32]
p =
1 is
equivalent to p
0.
912
p = (p
[901]
+
=
1). 1).
p = (p =
(p
Hence
[32]
p = (p =
1)
(p
1).
913
(p*
1)
0).
914
(p
[9M2] + 0) = (P =
[913, 32]
1).
913 and 914 together express the fact that the algebra
is
twovalued.
Every proposition
is
225
written the
.",
Up
logical
.
. .
to this point
relations
. .
.",
that
.
is,
,
"If
we have
. .
.",
or
"Both
and
etc.,
in arithmetic or geometry or any other mathematical have had no right to do otherwise. That "... c ..." system. + is by inter and then is by interpretation or does not warrant us in identifying the "Either pretation
would be written
We
"If
".
."
",
.",
theorem
"If
a c
6,
then
be
a"
with
. . .
"
(a
,
c b) c
.
(/;
.
c a)
".
We
have
."in theorems with then had no more reason to identify c ..." than a geometrician would have to identify the period at "... end of a theorem with a geometrical point. The framework of logical the
"If
relations in terms of
to interchange the joint But we can now prove with pcq, etc. We can then of p and q wath p xq, assertion p, demonstrate that if p and q are members of the class A then p c q is a And we can is equivalent to p cq. member of K, and that p, then
r
is logic.
"If
&lt;/",
is
of the
system
in
.
.
itself.
We
. .
the theorems
,
or
.,"
...
and ..
.,"
"If
then
."in
terms of
+,
x,
c,
etc., is
a valid procedure.
in
which these things are proved are never needed here and their after, except in the sense of validating this interchange of symbols we need not give them any section number. Consequently interpretation.
The theorems
(1)
If
is
an element
in
A, p =
is
and
[901] If
If
is
is
an element in
= A, p = 1
/;
are elements in
is
an element
in
[9
A.
1]
/;
[10]
p is an element in A, p an element in K.
q,
an element
in
A, and hence
(2)
q,
or p
q.
pq =
is
is
equivalent
pair,
to
the
p and
in
q.
p and
[49]
if
q are
elements in A, then p c q
an element
A.
But
(4)
0,
1
and hence
1]
[91] to (pq).
(p q)
is
an element
in
A.
p
is
is
equivalent to
p
4=
is
false".
[912]
false",
p =
(p
1),
and
[801]
"p
4=
is
equivalent to
=
"p
and hence
[901] to
is
false".
16
226
A
(5)
pc
is
equivalent to
"If
p,
then
1,
g".
[564]
"If
p eg
g".
gives
"If
p =
then q
=
it
1",
and hence
[901]
p,
then
p,
And = g
[22]
"If
then
g"
gives
peg,
for [901]
gives
"If
p =
g
1,
then
1",
and
(a)
Suppose as a fact p
1.
Then, by hypothesis,
1,
and
peg.
Suppose that p
={=
(6)
1.
Then
[914]
p =
0,
and
[563]
in
peg.
(6)
If
p and
[71]
g are elements in
K, then
p =
+
is
an element
0,
K.
[91]
p =
is
equivalent to
pg pg =
and hence
to(pg + pg).
(7)
Hence
Q.E.D.
g".
p =
g is equivalent to
"p
equivalent to
"peg
[22]
p =
g is equivalent to
"p
and
cyj".
By
if g,
(5)
above,
p".
eg and
this
is
cp"
is
equivalent to
"p
"If
p,
then
g,
and
then
And
equivalent to
is
equivalent to
in
g".
(8)
If
p and
g are elements in
K, then p
g).
4= g is
an element
K.
[902] (p
g)
=(P =
Hence, by
(9)
(6)
==
is
equivalent to
is
not equivalent to
g".
By
(10) p + g
is
(4)
and
(2)
above, Q.E.D.
"At
equivalent to
least
g, is
true.
[18]
p+q =
and
(2)
(pg).
above, (pg)
is false)
g, is
".
By
(p
(4)
is
is
"It
is
false
that
false
and
And
this
equivalent to
"At
least
one
of the two,
p and
true".
".
for
"If
then
...","...
=
. .
..."
for
.",
".
is
equivalent
to
...","...
..."
for
"Either,
or
etc.,
for
we have
and the proved that not only all expressions formed from elements in relations x and + are elements in K, but also that expressions which in
volve
c, and =, and
=j=
TwoValued
. .
Algebra.
of
"Both
The equivalence
. .
then
.
..."
with
".
.",
no longer a matter = (p = 1). Also, we can of interpretation but a consequence of 901, p go back over the theorems of Chapter II and, considering them as propositions then of the TwoValued Algebra, we can replace etc.,
.
and
..."
with
".
.",
etc., is
"If
.",
227
is
Each theorem not wholly in symbols gives a wholly in symbols. But when we consider
the BooleSchroder Algebra, without the additional postulate, 901, this procedure is not valid. It is valid only where 9 01 is one of the postulates
i.
e.,
only in the system of the TwoValued Algebra. Henceforth we shall write all our theorems with pcq for
q for
"If
.
"
If p,
then
7",
p =
use
"p
is
equivalent to
r/",
etc.
But
".
in the proofs
.
we
shall frequently
then
..."
instead of
.
",
etc.,
because the
symbolism sometimes renders the proof obscure and makes hard reading. (That this is the case is due to the fact that the TwoValued Algebra does
not have what we shall hereafter explain as the true
"logistic"
form.)
915
+
(p4=
1.
=
.916
q)
0.
Hence [913]0
(p
4= 1.
=
If
=
q
q)
(p
=
1,
?).
(1)
And
if
p = = 1, p
If
and p
[32]
= = p
then q
4= 1
and
q.
[913] q
0.
0.
Hence p =
then [913] p
= 0, and [32] p = =h 1, (2) p 4= q = 1 = p. Hence if p q, then q 4= 0, and [9 14] q If p = q and q = 1, then p = 1, and [32] p = 0. (3)
and p
=j=
1.
Hence
[915]
If
4= q.
= 0. (4) p = q and q 4= 1, then p 4= 1, and [913] p Hence [32] p = 1, and p 4= qBy (1) and (2), if p 4= q, then p = q. And by (3) and (4), = q are equivalent. Hence p 4= q and qp = q, then p And [3 2] (p = q) = (p = q).
=)=
&gt;
if
This theorem illustrates the meaning of the relation, =, in the calculus or 7; = If p 4= (/. then either p = 1 and q = of material implication. = 0, and if p = 0, then p = 1. Hence and q = 1. But if p = 1, then p
the theorem.
Let p represent
home".
"Caesar died",
died"
and
q represent
"There
"There
is is
no place no place
is
{
like like
If
"Caesar
is
not equivalent to
die"
home",
then
"Caesar
did not
is
is
equivalent to
{
"There
no place
like
home".
The equivalence
0} or
1}
917
p =
(p
1)
(p
4= 0)
(p
0)
(p
4= 1).
[9OM31416]
918
p
(p
0)
(p
4=
1)
(P
1)
(P
0).
[9M31416]
228
92
A
(p
l)(p
[24]
0)
0. 0.
pp =
is
And
[901]
false.
p =
(p
1);
[91]
p = (p =
0).
No
proposition
(p
921
l)(p
0)
0.
922
(7;
[918J
p =
(p
1);
[917]
p = (p
0).
p + p =
is
1.
Hence [9011] Q. E. D.
Every proposition
923
(p*
l)
(p*
[48, 9011]
Theorems
of the
same
924
(p q)
925
= (p = l)(g = 1) = (p 4 0) (g = ==  (p = l)(g 4= 0) = (p + g) (p * 0)(g 1) = (P + g = 0) = [(p = 0) + (g = 0) = 0] = [(P * 1) + (g * 1) = 0], etc., etc. = (p + g = 1) = (p + g 4= 0) = (p = 1) + (g = 1) (p + g) = = (p g) = [(p = 0)(g = 0) = 0] (P * 0) + (g 0) = [(P * l)(g * 1) 4= 1], etc., etc.
=
(p
1)
(p g
0)
=t=
0)
=(=
These theorems
ways in which the same logical the TwoValued Algebra. This is one of the
redundancy
of forms.
defects of the
system
its
much
.
We
turn
now
We
some theorems which do not require the additional postulate, 901, for the sake of bringing together the propositions which illustrate the meaning of
"material implication".
93
(per/)
(p +
q)
(p q
0).
1).
materially implies
"It
g"
is
is
equivalent to
true
"Either
is
false or g is
true",
and to
is false
that p
and
false".
"
Since
pc
g has
It is false
may
be symbolized by (peg).
229
931
(p
cry)
[3 4]
(p + q
q)
0)
=
q.
(pq).
(p +
&lt;/)
= p
And
[9
3]
(p c
q)
(p +
q)
[9
"p
.02] (? +
(/;+&lt;/
is
0).
q"
equivalent to
"It
is
false
that either p
is
false or q is
and to
"p
is
true and q
false".
932
(p
0)c(pc&lt;7).
[503] Ocr/.
If
Hence Q.E.D.
ry,
is false,
the famous
proposition".
p materially implies
false
q.
This
is
"A
proposition
implies any
933
(q=
})c(pcq).
[501]
pel.
Hence Q.E.I).
"A
This
is
true proposition
is
implied by any
proposition".
934
(pcq)c(p =
since
if
1).
The theorem
(p
c&lt;y),
then [932] p
q,
4= 0,
If
there
is
any proposition,
is
true.
This
A similar consequence of 9
=
0.
33
is:
935
(pcq)c(q =
If
If
(pcq), then [933] q =(= 1, and [913] p does not materially imply q, then q is false.
(/;
936
q)
(7;
q)
(p c
cry),
0,
q)
c (p c
q)
1
(/;
q)
c (p c
q).
[93435]
[32] If
If
0;
then p
p =
1, y;
and
if
= =
and
0.
=
0,
then q
1.
[932] If
[9331 If
If
p = q =
0,
1,
;; cry.
then
y&gt;
of
p implies
of
p implies
the negative of
"The
ry.
If
"Today
is
Monday"
made of green cheese", then "Today is Monday" implies "The is not made of green cheese", and "Today is not Monday" not Monday" implies "The moon is made of green cheese", and "Today is implies "The moon is not made of green cheese". Some of the peculiar properties of material implication are due to the
moon moon
is
230
the system
of logical classes.
But 936
936
p =
(p
1).
For
b,
classes,
represents
contained
in":
but
if
is
not contained in
it
does not
follow that a
of
b.
is
contained in not6
may
36] If
"(p
q),
then p c
q, if
q,
and hence
[3
1]
p.
q,
then q
938
(pq) c[(pcq)(qcp)].
[924]
pq =
(p
l)(q
==
1).
Hence
[933]
Q.E.D.
If
p and
939
(pq)c[(pcq)(qcp)].
[924]
pq =
(p
l)(q
1)
(p
0)(g
0).
Hence
If
[932]
Q.E.D.
p and q are both false, then each materially implies the other. For any pair of propositions, p and q, there are four possibilities:
1)
2) 3)
4)
= p = p = p =
p
1,
q q
q q
0, 0,
1,
= = = =
1:
true, q true.
false, q false. false, q true.
0:
1
:
p
p
0:
true, q false.
1
Now
in
the algebra,
cO,
1,
and
1;
but
cO
is false.
Hence
in
the four cases, above, the material implications and equivalences are as
follows
:
1)
2) 3)
pc pc
q,
q,
qcp, p = qcp, p =
q.
q.
4)
q.
q.
of
q.
Thus p cq and p =
and
"equivalence"
speak
"truth
ing,
p and
values"
231
cation and material equivalence are relations of the extension of proposi whereas the "implication" and "equivalence" of ordinary logic are relations of intension or meaning. But, as has been mentioned, the material implication, pcq, has one most important property in common with
tions,
"q
p"
in ordinary logic;
if
is
the relation of material equivalence, p connects a true proposition with a false one.
And
pcq
never
q,
These theorems should make as clear as it can be made the exact meaning and character of material implication. This is
important, since
significance
this relation.
of
r)
(q
p c
r)
(q
r)]
[q
c (p c
q)
==
r)}.
[13]
pq =
qp.
Hence
[32]
(p
[(/; q)
r]
= (p q c r), and [(q p) + r] = And [3 41] [(p q) + r = + ?) + r] = [p + (ry + + r] = Iqc(pcr)]. Similarly, [(q p)
But
[9
3]
[(p q) +
r]
(q
r)]
[(/&gt;
per). = (p c
(q
r)}
Principle of Exportation,
[p
q)cr]c
c(qc r)]
r";
pq
implies
r,
and
his Principle of
Importation,
[y;
"
c
r,
(q
r)]
if
[(p q)
r]
If
p implies that
[(pq)c
r]
q implies
[(p r)
then
p and
c
q are
both true,
"
r is true.
95
q]
[(q r)
y,].
[9 3]
[(p q)
r]
= =
[(p q)
[(&lt;/
r]
=
y;]
[(p + q) +
r]
[(p +
r)
ry]
= =
[93]
[(p r ) +
ry]
[(pr)cg], and
[(qr)+p]
l(qr) cp].
p and q together imply r, then if p is true but r is false, q must be false, and if q is true but r is false, p must be false. This is a principle first stated by Aristotle, but especially important in Mrs. LaddFranklin s theory of
If
the syllogism.
We
sufficient
number
of
TwoValued Algebra
illustrate
232
A
p =
(p
p c q. Any further theorems of the system will be found to follow readily from the foregoing. A convention of notation which we shall make use of hereafter is the
postulate
1),
following:
sign
;
=, unless enclosed
,
over
any other sign a sign c unless enclosed in parentheses, takes precedence and the sign + unless enclosed in parentheses, takes over any + or x
;
This saves
many
II.
FUNCTIONS OF ONE
The
is
Valued Algebra to propositions which involve the values of variables. Fol 3 lowing 3Ir. jlussell, we may distinguish propositions from prepositional
functions as follows:
or false;
proposition
is
is
is
either true
an expression, containing one or more variables, which becomes a proposition when each of the variables is re
a propositional function
its
values.
"
There
is
one meaning of
1,
Today
2,
is Monday"
.
ambiguously Jan.
or Jan.
or
etc.
is
Today
Jan.
1 is
is
Monday
implies
Tomorrow
Tuesday
we mean that
if
is
Monday, then Jan. 2 is Tuesday; if Jan. 2 is Monday, then Jan. Tuesday; if July 4 is Monday, then July 5 is Tuesday, etc. Today
variables,
1,
Jan.
2,
Jan.
3,
that
is,
all
variable sense,
"Today is Monday" is
When
or more accurately,
it is
and
"Today is Monday" is
There
today
is
is
"Today is
which
In
this sense,
"Today
is
Monday" is
its
true
it
is
always true.
its
"
It is either
false:
meaning and
"
today Today is Monday is a proposition. When Today, meaning Jan. 16, 1916, is one value of the variable today this value is substituted for the variable, then the propositional function is
this
For
meaning
of
in question, his
view
is
the one
we
adopt.
Principia Mathematica
is
cited hereafter as
Principia.
233
We may
functions, in
use
^r,
t(x, y),
f (*, y z )
t
e tc.,
x, or
to represent prepositional
y, or x, y, and z, etc. These propositional functions must be carefully distinguished from the functions discussed in Chapter II. We there used /, F, and the Greek
x and
here
Also, for
any function
&lt;px,
of one variable,
x.
\l/y,
f(x), V(x, y), etc., Chapter II are confined to representing such expressions as can be formed from elements in the class A and the relations x and + If x and y in *(*, y) are logical classes, then *(.r, y) is some
.
in
logical class,
such as x + y or a x + b y.
Or
if
x in
/(.r)
b.
is
a proposition,
then
/(.r)
is
functions,
&lt;px
&lt;px,
y},
(x, y, z),
etc.,
a x or .r +
The
becomes a proposition when a is replaced by one of its values, but does not necessarily become any such x proposition as a x or .r + b. x is a citizen of y, Monday, these are y is between x and z
propositional functions.
is
typical
They
become
either true or false as soon as terms denoting individual things are sub stituted for the variables x, y, etc. All the functions in this chapter are
A
is
fundamental conception
"range
is
that of the
of significance
The range
class,
of significance of a function
values of
&lt;px,
variables.
,
All the
If
&lt;px
be
is
mortal
the aggregate of
all
which x
is
is
is
Thus the
"universe
of
significance"
is
to propositional functions
what the
of
discourse"
to class
terms.
Two
propositional functions,
&lt;px
and
&lt;px,
r/^,
is
may
class of values of
x in
&lt;px,
or the range of
values of y in
\J/y,
is
Or the two functions may have a man and .r is a poet will have the
true will differ.
.r
same
for
is
range,
.r
for
Any
.r
which x
man
is
is
a poet
But some
s
.r
for
which
is
is
a poet
.r
is
either true
x precedes
nonsense,
is
a poet
and
precedes x
It is
4 According to Mr. Russell s "theory of types" (see Principia, i, pp. 4142), the one fundamental restriction of the range of a propositional function is the principle that nothing
234
range of
function.
is
A
&lt;px
determined, not by
but by
&lt;p.
&lt;px
we have a prepositional function of two variables, say x is a citizen of y we must make two substitutions in order to turn it into a proposition which is either true or false. And we conceive of two aggregates or classes
If
,
first
variable, x,
and the
second
These two
classes
It
may,
they
may
be different.
Turkey" is
Jones
is
citizen of
Jones" is
a citizen of John
is
nonsense.
3".
"5
precedes
is either true or false, as But precedes The range of x and of y in $(x, y) depends upon
5"
also
^,
not
upon x and
y.
&lt;px
xz
is
is by x\, x 2 convenient method of representing the values of x in This is not to presume that the number of such values of x in etc.
&lt;px
Any
If
sort of tag
all
,
3,
&lt;pxi,
&lt;f&gt;x*,
&lt;p%3,
etc., will
x, x 2 x z be propositions; and
is
etc.,
&lt;px
be a proposition.
the values of x
,
&lt;pxz
is
&lt;px
a proposition about
if
a certain individual
in
\I/(x,
which
.r 3 ,
not specified. 6
Similarly,
y) be
,
Xi, x*,
etc.,
,
then
$(%*&gt;
2/s),
t(x z y n ), t(x m y n ),
propositions.
II
We
shall
now make
new
and
S, giving
them
a meaning similar to, but not identical with, the meaning which they had in Chapter II. To emphasize this difference in use, the operators are here
set in a different style of type.
We
x
.
shall let
2x
&lt;px
represent
&lt;pXi
&lt;px
&lt;f&gt;Xz
to as
many
&lt;pxi
&lt;px.
And U x
&lt;px
will represent
z
T
x
e
&lt;px
to as
many
values of x in
&lt;px.
(W
ax6
to a b or a b.
is
of expressions
that presupposes the function, or a function of the same range, can be a value of the func It seems to us that there are other restrictions, not derived from this, upon the tion. range of a function. But, fortunately, it is not necessary to decide this point here. 5 which is the only "Individuals" in the sense of being distinct values of x in
&lt;px
conception of
6
"individual"
which we require.
&lt;px
It
may be
is
urged that
n is
The
most difficult, and we cannot enter upon it. But this much may be said: Whenever, and in whatever sense, statements about an unspecified individual can be certain gentle If any object to this, we shall reply n is a proposition. asserted, man is confused". Peirce has discussed this question most acutely. (See above, pp.
question
"A
&lt;px
9394.)
235
chapter, always
be complex.
write
"products"
shall,
in
this
might be an infinite set of values of x in does not affect the theoretical adequacy of our definitions. For nothing here depends upon the order of m ^x n p and it is only required that the values of x which are distinct should be identifiable or The ob "tagable".
&lt;px &lt;px
The
&lt;px
is
more
serious,
may
z,
&lt;f&gt;x
is
proposition
the proposition,
"For
"Either
&lt;pxi
or
&lt;px
or #r 3 or ...
&lt;px
etc.".
Thus
2x 2x
of
all
&lt;px
represents
is
some value
Similarly,
3,
is
true".
And
"For
&lt;px
a proposition.
^.TJ
&lt;px
^.r 3
x ...
is
&lt;ATi
and
&lt;px
and
&lt;px
&lt;px
etc.
",
Thus H t
&lt;px
values of x,
true",
is
true
We may
"
translate
&lt;px
Z x ^.r loosely by
true".
"
&lt;px
is
sometimes
and
by inasmuch accuracy
&lt;px
nx
loosely
is
always
This trans
&lt;px
as the variations of x in
may
The conception
3,
of a prepositional function,
&lt;px,
and
(px n ,
2x
&lt;px,
and
U x(px.
new types
hold for propositions generally, all the theorems of that system will be true when propositions such as the above are substituted for ...
a, b,
p, q,
etc.
(We must,
of course,
a, b,
... p, q, etc., in
the
Twoyalued Algebra represent propositions, x in yx, tion but a variable whose values are individual
to follow,
etc., is not
things.
we
shall
which no
To avoid any
possible confusion,
we
shall represent
letter, P.)
We
proved any theorem which can be got by replacing any proposition of the TwoValued Algebra, by
6,
&lt;px
...
n,
p, q, etc., in
&lt;px,
&lt;px
3,
2x
or
Ti x
&lt;px.
Additional theorems, which can be proved for propositions involving values of variables, will be given below. These are to be proved by reference to
Chapter II and in Section I of this chapter. As before, theorem by which any step in proof is taken will be given in square brackets. Since the previous theorems are numbered up to 9, the additional theorems of this section will be numbered beginning with 10earlier theorems, in
the
number
of the
One
will
Two Valued
Algebra,
be needed.
The
in sufficiently
236
of
mathe
matical induction.
of this section,
in the proofs
we
values of x in
&lt;f&gt;x,
by the difficulty that the number of may and hence the number of terms in 2 x #r and H x
are confronted
&lt;px
not be
finite.
And any
then be invalid
in tttis
con
Short of abandoning the proposed procedure, two alternatives are open to us: we can assume that the number of values of any variable in a propositional function is always finite; or we can assume that any
law of the algebra which holds whatever finite number of elements be involved holds for any number of elements whatever. The first of these assumptions
would obviously be
false.
is
true,
and we
shall
make
it.
This also resolves our difficulty concerning the possibility that the number of values of x in might not be even denumerable, and hence
&lt;px
&lt;pxi
&lt;px
&lt;px
and
&lt;pxi
x
if
tpx 2
&lt;px
might be
of values of
"2
We
(f&gt;x
the
&lt;px
number
+
.
.
any function,
x
&lt;px
be not
or
finite,
&lt;fXi+
&lt;px
or
&lt;px,
and
&lt;?Xi
Hx
&lt;px,
shall
be proved will be demonstrated to hold for any finite number of values and this being proved, our assumption allows us to extend the of x in
&lt;px\
theorem to any case in which the values of the variable in the function are This principle will be satisfactorily covered by the infinite in number. x shall always + and convention that 3 x 2 x 3 + 2 +
. . . . . .
&lt;pxi
&lt;px
&lt;px
&lt;pxi
&lt;px
&lt;px
number
of terms,
and any
shall
7 any variable, x, which is involved. This postulate, and the convention which makes
it
needed
1001
2p.r
1002 1003
7
H(px
&lt;p.r
= 2x = Hx
=
=
&lt;px
&lt;pxi+
&lt;px
+
x
tftf 3
+ ....
x
.
Def.
=
&lt;px
&lt;pxi
(f&gt;x
&lt;pxz
..
Def.
{&lt;f&gt;x}.
Def.
is
far
from
ideal, as are
many
other details of
this general
criticisms together in the last section of this chapter. But it is a fact that in spite of the many defects of the method, the results which it gives are without exception valid.
method.
We
main
237
=
{&lt;px
n }.
Def.
Def.
Def.
1004
1005
U x
2 x
&lt;px
&lt;px
= {n x = {S,0*r}
&lt;px}.
The
one
merely serve to abbreviate the notation. Elementary theorems concerning propositions which involve values of
variable are as follows:
2&lt;px
last four
101
n^.r.
&lt;f&gt;xi+
[5951]
&lt;f&gt;x+
&lt;px
...
J.
1012
U&lt;px
2&lt;px.
[595]
^ix^r
"For
some values
&lt;px
of x,
is
&lt;px
is
true"
is
equivalent to the
"For
all
is
values of x,
true"
false".
all
"For some values These two represent the extension of De Morgan s Theorem to propositions which involve values of variables. They might
&lt;px
values of x,
&lt;px
is
of x,
is
false".
be otherwise stated:
false that
"It
is
true that
all
is
"
is
is
equivalent to
"
"It
is
some x
is
is
not
";
and
is
"It
is
is
equiva
lent to
"It
not
102
Il(px
&lt;px
[599]
&lt;pxi
x x
x
&lt;px~
x x x
&lt;**3
x x
c c c
&lt;px
and and
1021
&lt;px
&lt;pxi
(px%
&lt;^r
&lt;px
&lt;px%
&lt;px!
&lt;px
^.r 3 ,
etc., etc.
c2&lt;f&gt;x.
[5991]
^c^+^roH.^*...
&lt;pXz
and
c c
(pXi
&lt;pxi
(px^
&lt;px$
+ +
. . .
and
&lt;pXs
+ pxz +
&lt;px
etc., etc.
it is
By
of
102,
or
if
&lt;px
is
true for
all
values of
true of
.r,
then
true for
.r,
"What is
true of
all is
.r,
any given
It
one".
By
is
then
it is
or
"What
is
true of
is
some".
reversible.
&lt;px
because
&lt;px
example,
let
may =
&lt;^.r
be
&lt;p.r
2,
and 2
&lt;px
For
Then
2^.r will
mean
"Some
day
is Monday",
"Today
but
&lt;px
n will
mean
"Today
(Jan. 1)
"Some
is Monday",
or will
mean
(Feb. 23)
is Monday", etc.
day
is Monday"
does
238
not imply
"
.1
Jan.
1 is Monday",
"Feb.
23
is
Monday
n
"
Monday.
x n in
&lt;px
means
"
in a sense
which
&lt;px
is
"some
value of
respect.
.r".
Xo
translation of
1022
II&lt;&gt;.r
cS^.r.
[51, 10221]
Whatever
1023
is
true of
is
all is
true of some.
"Whatever
3
H&lt;px
equivalent to
value of x, in
(&lt;pxi
&lt;px,
xn x
.
may
.
be,
&lt;px
".
Ii&lt;px
=
&lt;pxi
&lt;px
x
&lt;px
&lt;px
x
&lt;^r
=
. .
&lt;px
&lt;px
=1)
to
[901]
set
And
[5971]
&lt;pxi
x
1,
==1
is
equivalent
the
=
&lt;pXi
1,
&lt;pX
I,
pXz
....
&lt;px
And
[901]
tt&lt;px
&lt;px
I is
equivalent to
Hence
is
&lt;px
lf
&lt;px
z,
&lt;p.T
3,
....
This proposition is not tautological. It states the equivalence of the x tpxz x with the system of separate propositions 3 x product
&lt;pxi
. .
&lt;f&gt;x
&lt;pxi,
&lt;f&gt;x
z,
&lt;f&gt;x
s,
etc.
It is
tt&lt;px
by virtue
as
"For
all
values of x, ^x
is
true"
is
legitimate.
In this proof
we make
p =
(p
1)
which
it is
we can pass directly from any theorem of the TwoBy Valued Algebra to a corresponding theorem of the calculus of propositional functions. If we have, for example, pc.p+q, we have also "Whatever
virtue of 10 23
value of x, in
1023,
&lt;px,
xn
may
be,
&lt;px
p.r n
P".
And hence we
importance of
have, by
this:
it
Ux
[&lt;px
&lt;px
+ P].
We
shall
later
see the
gives us,
for every
theorem concerning
"material
implication",
a cor
"formal implication".
Next, we give various forms of the principle by which any proposition may be imported into, or exported out of, the scope of a II or S operator.
103
Z#c
2
&lt;px
P = S^r + P). + P = +
(
&lt;.ri
&lt;px&lt;2
&lt;px
+...) +
+ P) +
P
...
P) +
(^
(^3 + P)+
[5981]
1031
P+Z^o  ? X (P+
Similar proof.
p.r ).
103
may
be read;
"
x,
&lt;px
is
true, or
is
true
is
equiva
239
1031
For some
"Either
x, either
is
&lt;?x
is
true or
is
true
".
And
is
may
be
read:
P
is
&lt;px
is
true
equivalent to
For
some
1032
x, either
true or
(&lt;^.r
is
true
".
U&lt;f&gt;x
P = nx II (pX + P =
+
+ P).
&lt;pXz
(&lt;pXiX
&lt;pX
P
P) x... [5941]
1033
P+tt&lt;px
= U X (P +
tpx).
Similar proof.
&lt;(
Either
P is true
is
or, for
&lt;px
every
true."
x,
&lt;px
is
true"
is
equivalent to
"For
every
x,
either
true or
is
1034
S X (^ + P) = S X (P
[43]
2&lt;px
+*&gt;*).
+P = P+
2&lt;px.
1035
ILOi
+ P)
= n x (P+
&lt;px).
[103233]
is
Exactly similar theorems hold where the relation of the two propositions x instead of + The proofs are so simple that only the first need be
.
given.
1036
2&lt;px
&lt;f&gt;x
&lt;px
P
xP) +
. . .
=
(&lt;^.ri
xP) +
(^2
xP) +
(cp.i3
[594]
"
&lt;px
is
is
true",
is
equivalent to
"For
some
.r,
&lt;px
and
10361
1037
10371
xS^.i
&lt;r&gt;.r).
j:
&lt;px
1038
10381
S z (^ xP) = S,(P x
^.r).
Hx
(&lt;px
xP) = n x (P
&lt;^.r).
should perhaps expect that a proposition, P, might be imported into and exported out of the scope of an operator when the relation of P But here the matter is not to the other member of the expression is c
.
We
quite so simple.
240
104
A
Pc?&lt;px
= S,(Pc
&lt;px).
[93]
PC
=
2(f&gt;x
P+2&lt;px
= P
[5981]
= (P c ^rO = S X (P c
The
But
it is
+ (P c
^.r 2 )
+ (P c
^.r s )
[9
3]
&lt;px)
relation
c, in the above,
is,
"p
materially implies
9",
g".
We
g".
shall,
then, translate
pcq
"P
104 reads:
"For
or
&lt;^.r
by
is
"If
p,
then
is
.r,
true"
equivalent to
some
.r,
implies that
:
is true".
^&gt;.r
IlzO.rcP).
U&lt;px
[93]
S^ccP = S#r + P =
2
P
. .
[1012]
= (^i x^.r X^.TS x .) + P = (^d + P) x + P) x (^3 + P) = (^.TiCP) X(^.T CP) X(^.T CP)... = U x (&lt;pxcP)
(&lt;pxz
[5941]
[93]
"
&lt;px
is
true for
P".
some x
implies
P"
is
equivalent to
"For
every
x,
&lt;p
implies
It is
If
the
first also:
&lt;px
easy to see that the second of these two expressions gives is sometimes true, P must always implies P, then if
&lt;px
be true.
It is
it
2&lt;px
can put
thus:
P".
is
ever true,
P gives then P is
c
tt x (&lt;px
c P).
But we
"
true"
must mean
&lt;px
always implies
1042
PcU&lt;px
=
c
Tl x (Pc&lt;px).
[93]
U&lt;px
= P + Il^.r = (P +
[5941]
==
P +
(&lt;pxi
&lt;pxi)
x (P + x (P c
&lt;px
x ^r 2 x 3 x x (P + #c 8 ) x z)
.
.)
...
 (P c = Ut (P c
"P
&lt;pxj
&lt;px
2)
x (P c ^c s ) ...
[93]
&lt;px)
implies that
&lt;px".
&lt;px
is
x"
is
equivalent to
"For
every x
implies
1043
S^.T +
+&lt;X3
P
. .
[ioi]
.
&lt;X
+
.1*
241
= (^1
+ P) +
(^
(&lt;^
+ P) +
(&lt;pOr
+ P) +
+
[5981]
C P) +
C P) +
(^3 C P)
[93]
"
&lt;px
is
implies that
is
true"
is
equivalent to
"For
some
x, (px
At
first
sight this
"fallacy
of
division"
of composition"
and baldness.
out".
Suppose
&lt;px
be
"If
x
is
is
example about the separate hairs a hair of Mr. Blank s, x has fallen
And
of
let
be
"Mr.
Blank
bald".
Then
H&lt;px
cP
will represent
is
bald".
"If
all
Mr. Blank
Mr. Blank
And 2 x (#rcP)
that
Ti(px
if
will represent
is
some
is
hair of
bald".
Mr. Blank
such
Mr. Blank
(&lt;px
In this example,
their equivalence
is
cP is
cP)
is
dubious, and
The explanation
".
of the equivalence
&lt;px
this:
"
we
is
here deal with material implication, and false that n is true but P is false)
(&lt;px
"
P means
means,
simply
It
U&lt;px
cP
in this
example,
is
It is false
bald";
that
all
Mr. Blank
Mr. Blank
s
not
is
some one
of
is
Mr. Blank
is
hairs
such that
Mr. Blank
not bald
false".
No
necessary connection
hair
and baldness
material implication
last four
If
we compare the
theorems,
we observe
that an operator in
the consequent of an implication is not changed by being extended in scope to include the whole relation, but an operator in the antecedent is changed
from
to
&lt;p
to 2,
f
from 2 to
II.
This
is
pcqis
equivalent
where the sign of the antecedent changes but the consequent remains the same; and to the law n() = 2(), 2() = n().
+ q
The above
tional function
principles, connecting
used in later proofs. In fact, all the proofs can be carried out simply by the various forms of this principle and theorems 1011023. Since P, in the above, may be any propo
and
its
operator, are
much
sition,
#r n 2#r, II#c, etc., can be substituted for P in these theorems. In order that + \l/x) and x^j) are, of course, functions of x. must be significant and $x must be significant, + \I/x) be significant, ( par Such or $x" have meaning. further requisite that "Either and it is
,
(&lt;px
(&lt;px
&lt;?x &lt;px
considerations determine the range of significance of complex functions A value of x in such a function must be at + \I/x) and like x^.r).
(&lt;px
(&lt;px
17
242
once a value of x
(&lt;px
A
in
&lt;px
and a value
x n in
&lt;px
and
in
fa n
in
105
&lt;px
+ 2 fa
= Sx
^ + #r)
is
Since addition
associative
and commutative,
"
.r,
^,
or for
some
x,
fa"
is
equivalent to
"For
some
x,
either
or
fa".
If it
&lt;px
different
ranges
e.,
not indicative
(&lt;p.r
of the range
+ fa) did
not.
will
But
in
in
such a case the proposition which states their equivalence We shall make the convention that x n in n
&lt;px
and x n
wherever
&lt;px
n +fa n ) and n xfa n ), but fa n are identical, not only in and fa are connected, as in S^.r + S#r. Where there is no
(&lt;px
(&lt;f&gt;x
such presumption, it is always possible to use different letters for the But even without this convention, the above variable, as S#e + 2^y.
theorem
will
it is
significant
i.
e., it is
never
false
of this section.
&lt;px
Ufa = n z (
is
&lt;px
x fa)
Since x
associative
proof.
it
We
"For "For
= 2 x (&lt;pxxfa)
x, x
to hold, but
does not.
to,
is beautiful", is
not equivalent
some
x,
is
ugly and x
is beautiful".
Instead of an equivalence, we
have an implication:
1052
Sx (
[521
&lt;px
x fa)
n
( (p.i1
x fa
,
(&lt;px
xfa n ) c
&lt;p.r
and
(&lt;px
xfa n )cfa n
2,(&lt;px
Hence Hence
Similarly,
x,
[531]
[534]
U&lt;px
xfa) c^fa
Hfa = U x
is
(&lt;f&gt;x
+ fa)
",
fails
is
.T
to hold.
"Either
for every
is
beautiful
".
either x
ful.
ugly or
.r
is
beautiful
Some
not equivalent to, "For every x, may be ugly and others beauti
But we have:
243
&lt;px
Ufa ctt x
&lt;px
(&lt;px
+ #r)
[521]
c(&lt;px
n +\I/x n ),
(&lt;px
Hence
Hence
[53]
[5
H&lt;px
cH x
+
cU x ((px +
lemma
in
if/x)
33] II
&lt;px
n^x c n x (
+ $x)
we
write a
for
&lt;px
instead
it
for
&lt;px
&lt;pxi,
for
&lt;pxz,
for
&lt;px
s,
etc.
For example,
1052 we write
x^.r n ) c
n,
instead of writing
(&lt;pxi
&lt;px3,
etc., etc.
The
proofs are
this explanation.
This method
of writing such
lemmas
will
be continued.
&lt;px
With two
prepositional functions,
and
\f/x,
impli
S x (#rc^) and IL^.r c i^.r). But 2 x (#rc#r) states is false or \j/x is true: is a value of x for which either only that there and this relation conveys so little information that it is hardly worth while
cation relations,
&lt;px
to study
its
properties.
is
Ux
The
(&lt;px
c^.c)
the relation of
is
&lt;f&gt;x
"jormal
implication"
is
"For
every
.r,
false
and $x
true
is
a true statement"
(
&lt;px
Ux
(&lt;f&gt;x
c^x)
is
S x (^.r
is
x^.r), so that
.r
Ux
is
i//.r)
may
l
also
be read
is
any
such that
"At
&lt;px
true and $x
false".
The material
false
implication,
is
peg,
states only
statement";
least
"It
is
and
q
true
is
a true
or,
n
is
is
true
least
and
false".
The material
is
&lt;p
implication,
&lt;px
c\l/x n
states only
,
"At
false of
&lt;px
x n and
is
true of x n
false".
is
a true
statement";
is false
that
is
true and
^.r n is
is
false that
is
&lt;px
true and $x n
is
false
in the
whole range of
false.
&lt;px
not a
case in which
&lt;px
is
true and $x
To put
it
another way,
(^
^.r)
means
"Whatever
^"
plies"
of "im This relation has more resemblance to the ordinary meaning it should than material implication has. But formal implication,
be remembered,
nx
(&gt;.r
i//.r)
is
material implications; simply a class or aggregate of c ^.ri, ^.r 2 c ^.r, ^.r 3 c tx~, simply the joint assertion of ^,r,
is
is
etc.,
s
a material implication.
ing of
244
A
The
upon this relation are based certain derivatives and in the calculus of relations.
106
n^arc^a) = U x
[9
(&lt;px
+ fa)
&lt;pX
= Hx
(&lt;px
xfa).
n
3]
(pX n
Cfan = Q.E.D.
n
+ fan
(
&lt;pX
X fa n )
Hence
1061
[1023]
tt x (&lt;pxcfa)
c(&lt;px
cfa n ).
materially implies fa n
[102]
If (px
&lt;px
10611
c $x) x
1061]
&lt;px
n]
c fa^
[94,
If
&lt;px
is
true of
ar n ,
then ^
is
true of x n
This
is
is
for every x, a; one form of the syllogism in Barbara: for example, man implies x is a mortal and Socrates is a man, then Socrates
"If
,
is
mortal".
10 62
nx nx
( ??ar
c #r) c S x (
&lt;^x
c fa)
[1022]
10 63
&lt;px
c fa)
c(Il&lt;pxc
Ufa).
then
&lt;px
[1061] If
U x (&lt;pxcfa),
cfa n
Hence
10631
[53]
Q.E.D.
[U x (&lt;pxcfa)
[94,
xn^cn^o*.
&lt;px
1062]
is
If (px
is
always true.
10 64
nx
&lt;^.r
c fa)
c(2&lt;px
c.
2 fa)
[1061, 531]
10641
[n x (
&lt;f&gt;x
cfa)*2
1064]
&lt;px]
c 2 fa.
[94,
If (px
&lt;px
is
is
sometimes
true.
10 65
c fa) x
If
n x (^.T
fa;)]
nx
^.T
far).
[1061]
iA.r n
far),
then
&lt;px
cfa n and
far n .
Hence
[51]
whatever value of
xn
may
be,
&lt;^.r
far B .
Hence
[1023]
n x (^arcfar)
245
It is
is
a transitive relation.
&lt;px
= x is a another form of the syllogism in Barbara. For example let 1065 will then read: Greek \f/x = x is a man and x = x is a mortal
,
l "If
for every x,
is
a Greek implies x
,
is
man
#
and
for every x,
x x
is
is
man
implies
".
.r
is
a mortal
x,
is
a Greek implies
mortal
1065
10651
may
Ux
(&lt;px
c #c) c [n x (#r c
1065]
x)
Ux nx
(&lt;px
fr)].
[94,
10 652
U r (ifrx
[94,
fa;)
c [U x (
&lt;px
c $x) c
&lt;px
far)].
1065]
10 66
nx
#r c
[31]
Hence
[22, 53]
Q.E.D.
Any
equivalence
is
The
properties of formal
1067
Ux
to
=
(&lt;px
jx)
[tt x (vx
&lt;px)].
be, [22]
n
.
&lt;px
is
equivalent
&lt;px
Hence
1068
[U x
[1023]
Q.E.D.
=
(&lt;px
jx)
xn x (^.r =
fr)]
cn
x (^.i
Tr).
&lt;px
be,
if
&lt;px
= ^x and
i:
==
far,
then
10681
Ux
=
(&lt;px
^.r) c[II,(^.r
far)
[1068, 94]
10682
n z (^.r 
f.c)
c[H x (^r =
^.r)
[1068, 94]
last three
theorems,
is
a transitive
1069
Ux
and
(&lt;p
n x (^c 
[22,
10616263]
246
10691
A
n(^c =
[32,
If
$x)
= Ux
=
(&lt;px
#r).
1023]
to investigate the propositions
(&lt;px
we wish
may
differ
(&lt;px
\l/y,
we
2 Jl^r; ty), H S x (^r; ^), etc. And these are special + \l/y) are special cases of a function of two variables. x\f/y) and
two operators
(&lt;px
(&lt;px
Hence we must
first
III.
PROPOSITIONAL FUNCTIONS OF
Two
OR
MORE VARIABLES
y), gives
prepositional function of
&lt;p(x
two
y n ),
variables,
&lt;p(x,
the derivative
y),
propositions
m y n ),
,
U*&lt;P(X,
S*2
tf
p(a:, y),
2 y n x p(a:,
etc.
The
range of significance of
&lt;p(x,
&lt;p(x,
y) will
comprise
all
the pairs
(x,
y) such that
y)
is
.T 3 ,
We
i/
#1, xz,
and a
etc.,
one of the x
of the
s,
&lt;p(x,
y)
is
As has already been pointed out, the function may be such that the class of values of x is the same as the class of values of y, or the values of x
may
y.
If,
for
example,
&lt;p(x,
y)
be
"x
is
brother of
the class of
.r
for
which
&lt;p(x,
y) is significant consists of
identically the
nificant.
all
9
class of y s for
which
&lt;p(x,
&lt;p(x,
y) is sig
y) is the class of
member
of
of
or with
any
other.
Thus
if
the
members
such a
G2
&gt;
(ai, ffi),
(fli,
(
2,
a 2 ),
a,),
(i,
as),
(a z
ai),
(2, a 3 ),
,
...
(a 3 ,
.
.
aO,
3,
a 2 ), (a 3 a 3 ), ...
Etc., etc.
is
But
y",
if
&lt;p(x,
y) represent
"x
a citizen of
y",
?/",
or
"a:
is
a proposition about
or
"x
is
member
is
of the class
the class of x
and the
class of y s
for
which
9
&lt;p(x,
y) is significant will
"A
be mutually exclusive.
is
Presuming that
Schroder treats
first
brother of
A"
significant
i.
e., false.
10
all relatives
as derived
three chapters.)
But
this
from such a class of ordered couples. (See is an unnecessary restriction of the logic of
247
does not neces
Although
&lt;p(x,
y) represents
some
relation of x
and
y, it
sarily represent
any
cy
or x
y;
and
it
y),
&lt;p(x*,
Hence H v v(xi,
and properties
y),
Uv
&lt;p(x,
y),
^v
lt
the meaning
of
And U y
&lt;p(x,y),
2 v v(x,y),
which follow from preceding definitions and theorems. and 2 t Il x &lt;p(x,y), y) are propositional
&lt;t&gt;(x,
We
1102
11
UxUy 2xUy
n*Z Sx S
&lt;f&gt;(x,
y)
==
ll t
{n y
&lt;t&gt;(x,
y)}.
Def
Def.
Def. Def.
==
&lt;p(x,y)
2 x lU y &lt;p(x,y)}.
03
tf &lt;p(x,
y)
==
nx
sy
&lt;p(x,
y)
} .
1104
tf
?(.r, y)
=
2,{S^(.r, y)}.
from these definitions that our explanation of the range of ^significance of functions of two variables was not strictly required; it The same con follows from the explanation for functions of one variable.
It will be seen
number
of values of variables
and interpretation
of
also
(Where the
will
first
comma between
.r,
the two
be omitted:
Since
&lt;p(x_y)
p(.r 2
y), etc.)
Uy
&lt;f&gt;(x,
y)
is
the defini
IUIX*,
And
y)
==
U^Il^x,
y)}
=
Tl y
&lt;p(x,y)
xll y
&lt;f&gt;(x
z y)
is
&lt;p(xiyi)
x
x
{
&lt;x
x ^Oi*/ 2 ) x x 2) x
&lt;&gt;x?
&lt;p(.ri?/s)
x
x
^(.r 2 z/ 3 )
&lt;p(x*yi)
&lt;p(x
3 y) x
And
similarly,
y)
by 1001,
SJlXz,
And
= S x {n^(;r,
y)}
=
U
is
x
x
x +
. .
Etc., etc.
248
Or, in general,
expanded
or
x)
letter
(3)
The operator
The
from
+ or
between
varies
Some
2xUv
&lt;p(x,
It is usually sufficient to
it
read ZJI y
x,
"For
some
such
x.
ami.
every
y",
but
&lt;p(x,
strictly
should be
"For
"For
some
every y
is
is
that".
Thus 2 x H y
true".
y} should be
&lt;p(x,
some
x,
every y every
such that
&lt;p(x,
y) is
And H v 2 x
is
true".
y) should be
"For
y,
some x
is
such that
v(x, y)
which
ma
y)
=
"Either
That
is,
2x
IIy&lt;p(x,
y)
means
&lt;p(x,
for x\
or for # 2
...
or for
&lt;p(x,
y) is true, y)
is
&lt;p(x,
true,
y)
is
true".
On
the
other hand,
Tl y
2x
&lt;p(x,
y)
= 2x
&lt;p(x,
i/i)
xS x ^(.r,
*/ 2 )
xS x
yi,
&lt;^(.T,
y
y) is true,
&lt;p(x,
That
y)
means
"For
some x and
for
&lt;p(x,
and
for
&lt;p{x,
y) is true,
and
is
some x and y 3
y) is true, y)
and
.";
or
"Given
any y, there
one x (at
&lt;p(x,
is true".
The
following illustration of the difference of these two is given in Principia Mathematical n Let y) be the propositipnal function y is a proper
&lt;p(x,
"If
fraction, then x of y,
is
y".
Then
we have 2 x
&lt;p(x,
that
"If
Ii y
2x
&lt;p(x,
y)
is
satisfied.
In fact,
U^
is
&lt;p(x,y)
is
always
?/".
But 2 x li y
is
&lt;p(x,
y) expresses the
propo
frac
"There is
greater than
any proper
which
is false.
if
In this example,
11
we should read S x n y
"For
#";
See
i,
p. 161.
249
these two
IlySz
"For
x",
required are
we
r
shall see.
are
II
since every rela propositions can be turned into a onedimensional array, tion throughout will be in the one case x in the other + and both of
,
,
and commutative.
It follows
the range of significance of a function of two variables that any such func the ordered tion, y), may be treated as a function of the single variable,
&lt;p(x,
1105
SsSy^Gc, y)
tt x ll y
&lt;p(x,
= S (x
V )&lt;p(x,
y)
y)
==
S Xf
&lt;f&gt;(x,
y).
y).
1106
y)
= n (x
V )&lt;p(x,
==
nx
&lt;?(x,
The second
1 1
is
x r and y a be any values of x and y, respectively, in == r y s ). such that a value of (x, y) y) n say, (.T, y) n
07
If
y)
there
&lt;f&gt;(x,
&lt;p(x
1105 and 1106 could be derived from 1107, but the process is tedious, and since our interest in such a derivation would be purely incidental,
we
prefer to set
If
down
all
three as assumptions.
y
&lt;p(x,
we wish
2x 2y
y) in
with a con
stituent of
is
y),
if
some convention
of the order of
terms
&lt;p(x,
Sz
&lt;p(x,
y)
required, because
y) be unaltered,
determined
be impossible unless the number of values of y is we which, by our convention, need not be the case. Hence
of
terms
in
S Xf
&lt;
&lt;p(x,
y),
m yn) precedes
&lt;p(x
y s)
if
m+
r
y
s,
and where
n =
+ *,
&lt;
s.
of
terms
in S*,
&lt;p(x,
y) will be
&lt;p(xiyj
2 yi)
This arrangement determines an order independent of the number of values r y e ) can always y) n in terms of of x, or of y, so that the equivalent of
&lt;p(x,
&lt;p(x
to govern the be specified. exactly similar convention is supposed with the terms y) and their identification arrangement of terms in n*, y
12
An
&lt;p(x,
of
U x ll y
12
&lt;p(x,
y).
These conventions
of order are
proof of theorems:
we note them
any
theoretical
familiar device for denumerating the rationals i. corner. agonals, beginning with the upper lefthand
the This arrangement turns the twodimensional array into a onedimensional by successive e., by proceeding along
250
A
The
y)
objection.
identification
2x2y
will
&lt;p(x,
y)
with
Sx
&lt;p(x,
y),
and
of
HJIy^Gr,
with
nx
is
&lt;p(x,
y),
of little
propositional functions
of relations
itself,
but
which
to be derived
two or
more
to
variables.
Having now somewhat tediously cleared the ground, we may proceed the proof of theorems. Since 2 X y y) may be re y) and H x y
,
&lt;p(x,
&lt;p(x,
(x, y),
many theorems
here follow
111
sx
&lt;p(x,
= ^X^ V
&lt;P(X,
y}
= H z
y &lt;p(x,
y)
= [U x U v
&lt;t&gt;(x,
y)}.
[110506, 1005]
1112
Xf y
&lt;p(x,
y)
= U x ll y
&lt;p(x,
y)
= S Xf
v &lt;p(x,
y)
= {S^pfo
y)}.
[110506, 1004]
112
Hx
&lt;p(x,
y)
&lt;p(x,
y) n
[102]
1121
&lt;p(x,
y) n
c2 Xry
y)
&lt;p(x,
y).
[1021]
1122
nx
nx
&lt;p(x,
c s*
&lt;p(x,
y)
[1022]
1123
(x, y) n
&lt;f&gt;(x,
y) is
equivalent to
".
"
Whatever value
of (x, y), in
&lt;p(x,
y),
may
be,
&lt;p(x,
y) n
[1023]
1124
U x Tl y
&lt;p(x,
y)
is
equivalent to
be,
&lt;p(x
"Whatever
values of x and
y,
in
v(x, y), x,
and
y,
may
r
r y^".
[1023]
IL y
&lt;p(x,y),
equivalent to
&lt;p(x
"Whatever
&lt;f&gt;(x
value of
.r,
in
yY\
y
s
And H y
y)
is
equivalent to
"Whatever
(f&gt;(x,y),
may
be, v(x
r y^)".
But
y).
[1101]
the values of x in
Hy
x in
&lt;p(x,
Hence
Q.E.D.
"
1125
is
"Whatever
value of (x,y), in
values of
.r
&lt;p(x,y),
(x,y) n
may
be,
&lt;p(x,
y) n
equivalent to
"
"Whatever
and
y, in (p(x, y),
x r and y,
may
be,
&lt;p(x
ys]
[11062324]
1126
n x n,
&lt;p(x,
y)
tf
&lt;f&gt;(x
n y)
[1101, 102]
251
1127
n x n^(.r,
S x S^Gr,
y)
is
= U v TL x
&lt;p(x,
?/).
Since x
associative
1128
y)
==
S S x ^(a;,
tf
Since +
11
is
associative
.
29
IIJI,
&lt;p(x,
y)
c n x ^(jr, y n )
[112(327]
11291
lUI^C*,
[22,
y)
c&lt;p(av#,).
1124]
y)
113
IIJI^O,
c SJI^Gc,
y).
[1101, 1021]
1131
SJI^(a;,
[1103]
y)
cU
2x
&lt;p(x,
y).
ny s x
&lt;p(;e,
y)
&lt;p(x
yi)
+ + +
&lt;K#i2/2)
^(^3^/2)
^(.TSZ/S)
&lt;p(xiyz)
^(.r 2 z/ 3 )
sum
to
z)
+ +
p(.ri2/i)
&lt;f&gt;(x
x x
&lt;f&gt;(xiy
&lt;p(.ri#
3)
x x
z yi) x
&lt;p(x
yz)
&lt;f&gt;(x
y3)
Etc., etc.
where
/I is
the
sum
is
of all crossproducts.
But
[1102] this
SJI^O,
==
y)
+ A.
/).
n^S^Gr,
y)
c I^Z^Gr,
y).
We
is
have already called attention to the fact that the implication of 1131 not reversible that S x n^(.r, y) and I^S^Cr, y) are not equivalent.
1132
nx
2y&lt;t&gt;(x,
y)
c Sx S
tf
^(.r, ?/).
[1103] n,S,
[1104]
S x S^(a,
S
c2
And
[5992]
We
have
where
252
functions
of
$(x,y)
by
f
or
UsUy
114
[&lt;f&gt;(x
y) Ci//(x, y)].
By
We
shall
these forms.
y) y)
y [&lt;p(x,
y)
c t(x,
y)]
= H = Ux
x&gt;
v [&lt;p(x, v [&lt;p(x,
+ j(x,
y)]
y)].
x^(z,
[106]
1141
nx
y [&lt;p(x,
y)
c t(x,
y)]
c[&lt;p(x,
y) n
c f(x,
y) n ].
[1061]
11411
( (
nx
y [&lt;p(x,
y)
c t( x
y)]
&lt;p(x,
y) n }
c j(x, y).
[10611]
1142
nx
y [&lt;p(x,
y)
c$(x,
y)]
c Sx
y [&lt;p(x,
y)
cf(x,
y)].
[1062]
1143
n*,
[&lt;f&gt;(x,
y)
^(.r, y)]
c [n,
^&gt;(a;,
?/)
Xf y
$(x, y)].
[1063]
11431
{n x
,[^(.r,
y)cf(x,
y)]
xH x
^(x,
?/)}
nx
y }(x,
y).
[10631]
1144
nr
y [&lt;p(x,
y)
^(.T, ?/)]
c [2 Xt
y&lt;p
(x, y)
c S,
y ^(.r,
?/)].
[1064]
11441
{n x
tf
[^(a;, y)
^(ic, y)]
x Sx
,^(.T, y)}
2,,
y $(x,
y).
[10641]
1145
{n x
y [&lt;p(x,
y)
cj(x,
y)]
xH
If v [^(a;,
y)
cf(a
[1065]
11451
n Xftf [^(ic,2/)c^(a:,y)]
c {n x
[10651]
,
[^fe
2/)
cr(.T, y)]
cn
x&gt;
J^GT,
y)
c{(x, y)}}.
11452
nx
[*(x,y)ct(x,y)] c {n^,
y [&lt;^(^
y) c^(.r,
?/)]
cn
tf
[^(a;, y)
c rfe
y)]}.
[10652]
1146
n,,
y [&lt;p(x,
y) c^(;r,
?/)]
= nx
,[^(.1%
i/)
c^(.r,
y)].
[1066]
Similarly,
of
253
1147
n*,
v [v(x, ?/)
}(x, y)]
=
(n x
v [&lt;p(x,
y)cf(x,
y)]
y)c&lt;p(x,
xn,, [*(*,
[1067]
y)]}.
1148
{n*.
y [*&gt;(ar,
y)
v [t(x,
y)
(x, y)}}
cU
[1068]
y [&lt;p(x,y)
(x, y)}.
11481
II X
[*&gt;(*,
?/)
iKz,
?/)]
{II X
,[^0, y)
f(z, y)]
f
c n,
[10681]
[*&gt;(*,
y)
f(x,y)].
11482
n,
y [^(o;,
y)
 rfe
2/)]
c fn x
tf
[^(ar, y)
t(x, y)]
v [&lt;p(x,y)
cU,,
[10682]
= rfe
1149
Hx
J^GT, y)
t(x, y)}
[^(a;, y) n
t(x, y) n ]
c[U c[2 x
[1069]
x&gt;l
,&lt;p(x,y)
&lt;p(x,
y)
= Ht = S Xf
y)
t(x,y)]
t
*(x
y)].
11491
n,
tf
[^(.r,
i/)
 ^fe
?/)]
= nx
y [^(ar,
j(x,
y)].
[10691]
The method by which such functions are treated readily extends to those of three or more variables. y, z) may be treated as a function
&lt;p(x,
(a:,
y, z)
of x and y, or of the ordered just as \l/(x, y) can be treated as a function with each extension of Strictly, new definitions are required pair (x, y). number of variables, but the method of such extension our theory to a larger
will
be entirely obvious.
For three
y, 2)
variables,
f
we should have
t
njiji^O,
SJIyll
,&gt;(&,
y, z)
= u x {u v u = S x {lI n
1/
tf&gt;(x
y, z)}
^(.T, y, z)\
Etc., etc.
It is interesting to note that the
will
be
n
and
Since
*)*&gt;(.r,
y, z) y, z)
==
S (I
y, z)
==
,)&lt;t&gt;(x,
njl^, S x S (y
,
g )&lt;p(x t
y, z)
f)
^(ar, y, z)
z y,
njl (v
,)&lt;p(x,
= n
(j/
,)&lt;p(xiy,
z)
xn
(y ,
g}
&lt;p(x
z)
xn
(l/
,)^(x 8 y, z)
254
x
A
.
.,
and
n (y
y
,
Z )&lt;p(x n y,
2)
= Uy
llz&lt;p(x
n y,
2), etc.,
we
n
And
( x,
Z)&lt;P(X,
y, 2)
= nji = n n
tf
(I ,,
2)
^(.r,
?/,
2)
(I&gt;
Z)&lt;P(X,
y, 2)
= n ^iwo, #, 2) = njiji^O, y, 2)
(X
&lt;p(x,
similarly for
S (I
2)
This
calls
as a function of three variables or as a function y, 2) can be treated not only of one, but also as a function of two, x and (y, 2) or (x, y) and 2 or (x, 2)
and
?/.
In general, the conventions of notation being extended to functions of any number of variables, in the obvious way, the analogues of preceding
will follow.
We
failed to treat of
such expressions as
U&lt;px
xUfy,
2&lt;px
U\f/y, etc.,
The reason
x^z/),
2 x Hy
(&lt;px
+ ty),
etc.,
special cases
of the
two
variables.
We may
&lt;p
+ U\l/x and n tpx + U\{/y. between two such expressions as II need not be identical; there and The ranges of the two functions,
difference
\f/,
may
be values of
.r
in
&lt;px
of y in $y.
But
in
any
&lt;px
with x n as a value of
.r
in ^.r.
For
this reason,
is
same range.
the more general case, in which the functions are not we have px and Theorems involving functions of this type restricted to the same range.
not always be significant for every choice of v and \J/. There may even be cases in which an implication is not significant though its hypothesis is
will
significant.
But
true;
for
significant,
they
they will never be false for any functions, however chosen. + \f/y) follows from the The meaning of an expression such as 2 x U y (
will
be
&lt;px
definition of
2x
Tly&lt;f&gt;(x,
y).
(
i
S x IIj,(
&lt;px
\j/y)
= II, =
+ +
&lt;px
i/^)
z
Il y (
&lt;px
.r
&lt;xi
+ $y) + U v ( + \/ z x
&lt;px
+
\/
^)
3
+
.
.
&lt;xi
{(VP.TS
. .
.
^I) x( ^3 +
^2/2)
x( ^3 +
^2/3)
Etc., etc. of
And
for
any such expression with two operators we have the same type
255
in general.
&lt;^.r
The
+
\fry
is
itself
12
(1)
[13]
II
&lt;&gt;.r
n^.r
(2)
II\l/y
= =
&lt;pX]_
&lt;/?.r
x
(&lt;p.r
(&lt;f&gt;X!
x n^i/) x
(&gt;.r 2
x n^z/) x
x Ityy) x
[598]
xn
(3)
tf
(^r 8 x^y)x...
[10371]
[iioi]
By
(2)
and
13,
tpxi)
x (n^y x
&lt;^.r
2)
(H^
3)
(p.r 3 )
x ...
n y (^
xll
.
&lt;^.r
2)
tf
(^x^.r
^.r)
x ...
[10 37]
[1101]
(4)
"
Similarly, ll^y
.r
xH^r = U y U x (\f^y
is
=
is
11^(^.1 x^y).
"For
&lt;px
is
a;
and
\f/y
z/"
equivalent to
every
122
and every
y,
&lt;px
and
^z/
are both
etc.
S^.r + S^z/
= 2^z/+2^.r = ^^^(^.r
S^i +
(1) (2)
[43]
2&lt;p.i
S^
 S^+S^r.
i
S^
= (^
^2 + ^3 +...)
&gt;/
[5981]
[1031]
SxZ^ +
ifc/).
[1104]
(3)
By
(2)
and 4 3,
[103]
S x S,(^+^r).
tf
[1104]
(4)
Similarly,
S^+2^.r = S
256
for
A
some
x,
&lt;px,
"Either
some
etc.
y,
tyy"
is
equivalent to
"For
some
x and some
123
y, either
&lt;px
or
S^r
=
[13]
S5
(1) (2)
O?2
(
X S^y) + (^T 3 X
[594]
+ S
tf
x $y) + S y ( +
x ^y)
[10361]
...
S^C^cx^).
S
x
i
[1104]
(3)
By
(2)
and
x
13,
Si =
S^
y
x ^cj +
&lt;^x
+ Sy
=
(^x^ )+
8
tf
..
[1036]
SA(^x^).
#,
^?/"
[H04]
#r)
(4)
"For
Similarly,
&lt;^.r,
 S S x (^c x^).
"For
some
y,
&lt;px
x,
and
for
equivalent to
some x and
some
124
and
+
^y",
etc.
v?.r
n^
H&lt;px
n^ + n^
IL&lt;f&gt;x
= nji^z
=
^) = n*n y (^ +
(&lt;px
&lt;?x)
UyUx
3
== Il v
U x (frj+
&lt;px).
(1) (2)
[43]
+ TIty
= U^y +
x
&lt;^x
U^px.
+ n^?/
=
=
(^aji
(&lt;pxi
&lt;px
+ ttty) x
(^2
...
[5941]
+
=
M
x
x...
[10371]
n s n (^ + ^).
tf
[iioi]
(3)
By
(2)
and 4 3,
+ n^?/
= (n^y +
^o x (n^ +
^2)
n
.T 3 )
x...
[1037]
(4)
"Either
Similarly,
x,
&lt;px,
Uty
H&lt;f&gt;x
= U v U x (^y +
\f/y"
&lt;px)
= U y tt x
(&lt;px
for
every
or for every y,
is
equivalent to
"For
every x
257
and every
invalid.
y, either
&lt;px
or
\f/y",
etc.
At
&lt;px
first
may seem
it is
One may
If
say:
"Suppose
it is
be
.
is
a number,
odd
and $y be
is
a number,
even
Then
even
,
U&lt;px
U^y
will
be
\f/y)
Either
will
every number is odd or every number Every number is either odd or even
illustration lies in misreading TL x
is
".
but
U x ll v
(&lt;px
be
The mistake
\^y).
of this supposed
Uv
(&lt;px
It
is
is
legitimate to choose,
as in this case,
&lt;px
and
\f/y
identical:
but
it is
not
legitimate to read
UxUy Ux
(&lt;px
with
a corresponding value of y.
lly(&lt;px
To put
each given value of x were connected nzllx (#c + #c), it another way:
as a special case of
either
of x
&lt;px
or
\l/x",
but would be
&lt;px
+ ty), would not be "For every value of x, "For any two values of x, or for any value
$x"
.
and
itself,
either
or
Thus HJIj/^T +
is
1/^)
in the
supposed
in fact
is
or for
itself,
either one
pair of
numbers,
so
is even"
that
lent.
U&lt;px
Il x lly(&lt;px
false,
A
125
somewhat
two theorems.
2&lt;f&gt;x
The analogues
of these, in
&lt;p(x,
y),
do not hold.
^x
Wy
By
= n^y + 2
&lt;px
= 2 X II V
&lt;f&gt;x
+ ty)
==
(1)
(2)
[43]
proof similar to
(2) in 122,
2&lt;px
&lt;px
ty)
&lt;px)
H^y + S + S^r
==
Uy ^x
"For
for
some
or,
^"
equivalent to
some x
and every
126
y, either
or
^?/",
etc.
s&lt;^.r
[13]
(2)
(2) in 123,
(3) in 123,
S^c xltyy
2&lt;px
==
==
S*n (^c x
tf
xH^z/
S,H y (^ x
(2) in 121,
(3) in 121,
is
Ityy
Ityy
xS^c xS^r
==
n H
tf
2 2
some
y,
&lt;px
x,
&lt;px,
and
for every y,
etc.
fy"
equivalent to
"For
some x and
every
18
and
^?/",
258
A
We may
by saying that
for functions
of the type
(&lt;px
\f/y)
and
(px
\j/y)
members
and
n
l 2J
the operators
may
between the propositions. It will be unnecessary to give here the numerous theorems which follow from 105126 by the principles pqcp, pcp + q, and Il^xc Spx, etc.
gives at once
(i)
nx
(&lt;px
(2) (3)
(4) (5) (6)
Etc., etc.
And
122,
2&lt;px
2\j/y
= 2x 2v
(1) (2) (3) (4) (5) (6)
+
(&lt;f&gt;x
\fry),
etc.,
gives
S,
2&lt;px
2^z
from
Etc., etc.
H&lt;px
or
S#ccS#r,
with
Ii\f/y
c S^y, giving
by 5 3,
(3) (4)
&lt;px
+
x
Tl\{/y
TL\f/y
&lt;px
Etc., etc.
Each
of these has a
whole
set of derivatives in
which
259
etc.
by
H. x lly(&lt;px
1/
?/),
etc.,
H&lt;px
xZi/ Z/ by
(2),
n x S (^x^),
J/
We
give, in
summary form,
the derivatives of
by way
any one
of illustration:
Any one
of
of
xU\l/y
x
x
&lt;px)
&lt;px)
S n x (^ x
tf
U y Ux
9x x ^)
S vnz (
&lt;px
x
&lt;?x
n ipx
zty
xn
etc., etc.
etc., etc.
^),
This table summarizes one hundred
fiftysix
etc., etc.
portion of those to be got by such procedures. and ((px + Functions of the type of
(&lt;f&gt;xx$y)
kinds
(3)
of
implication
,
relation:
(4)
(1)
tf
njl^rc^);
$y).
"implications",
n x S v (^c^);
of the
of
all
2 x n y (&lt;pxcty)
and
S x S (^c c
field
first,
them
mathematics.
Xor
by means
of
formal
implications.
Perhaps
UJl y
127
(&lt;px
c $y)
tf
is
n x n (#cc
[1101]
y^
i
c Ufy) x
(^
c
...
[1042]
And
Uifry,
&lt;pxz
c H^y,
&lt;px
c nty,
etc.
1271
{IIJI
J/
Ozci/
?/)
x^r B
[94, 127]
260
If for
A
every x and every
is
y,
&lt;px
\l/y,
and
for
some given
x,
&lt;px
is
true
then $y
1272
Il t Tl v (
II
 #c
&lt;p
(1)
If
Hence
[5991]
And
if
tt y (2
&lt;px
c ty)
(2)
[93]
is
S^c
"If
HxHy
(&lt;px
c $y)
is
equivalent to
y".
there
is
is
true,
then $y
1273
{nji w (#rc^)
[1272]
If
xn y n,(^cf2)} cn x n,(^c^).
S*ccII3, and
[12 72]
:
This implication relation is here demonstrated to be transitive. In fact, it is, so to speak, more than transitive, as the next theorem shows.
1274
{(S^ccS^) xn n,(^cfz)} cIIJI.C^c cfz). = s^z/cn^. [1272] n v n (^crz) And [51] if S^c cZfy and S^y c Hfz, then S^.T
tf
1275
(n x
[1272]
ns
if 2&lt;pxcttty
And
[51]
IV.
The
logic of classes
and the
logic of relations
begun with a calculus of propositions, the TwoValued Algebra, which includes all the theorems of the BooleSchroder Algebra, giving these
theorems the propositional interpretation.
sidered as belonging to the calculus of
We
then
."
261
...","...
is
equivalent to
..."
by
".
.",
etc.
The
it
Schroder Algebra;
simply includes it. the calculus of propositions the TwoValued Suppose, then, Algebra our fundamental branch of symbolic logic. We derive from it
we make
We may
logical classes,
two and a
calculus of relations,
in this section
The present section will not develop the logic of classes, but will present the method of this development, and prove the possibility and adequacy At the same time, certain differences will be pointed out between of it. the calculus of classes as derived from that of propositional functions and
the BooleSchroder Algebra considered as a logic of classes.
distinguish classsymbols from the variables, x,
tions,
y, z, in
In order to
propositional func
we
by
a, 0, y, etc.
conceived as the aggregate of individuals for functions, a given class is a man", n represent which some propositional function is true. If
"a:
&lt;px
s for
which
&lt;px
is
men.
If,
then,
propositional
represent the aggregate of individuals for which the will be "the class determined by is true, function
z(&lt;?z)
&lt;pz
z(&lt;pz)
the function
acter
etc.
13
&lt;&gt;".
&lt;pz",
or
"the
class
We
can use
a, 0, 7, as
"a
an abbreviation
for
z(&lt;pz),
2(^2), z(&),
&lt;pz",
z(tpz) will
the function
&lt;pz
are the
same
function.)
The
relation of
an individual member
"
member
of
a "or
briefly
a".
here borrow the notation of Principia. The corresponding notation of Peirce and Schroder involves the use of S, which is most confusing, because this S has a meaning But in of a propositional function. entirely different from the S which is an operator such does not represent an aggregate of individuals; it represents Principia, And z(tpz) is not a primitive idea but a notation supported by an elaborate that Our procedure above is inelegant and theoretically objectionable: we adopt it theory. because our purpose here is expository only, and the working out of an elaborate technique would impede the exposition and very likely confuse the reader. As a fact, a more satis
13
We
"
z(&lt;pz)
&lt;pz".
makes no important
difference.
262
1301
"x
A
xn
a
z(&lt;pz)
=
&lt;px
n is
member
of the class
determined by
&lt;pz"
is
equivalent to
"
&lt;px
n is
true".
we continue
is
and
theorem a number.)
The
class
"a
relation
"a
is
contained in
of
ft"
ft
is
member
ft"
a
ft.
is
by a c
"is
a member of (3 also. We shall symbolize The sign c between a and ft, or between
in";
z(vz)
2(^2), will be
",
contained
e
c between propositions
e
will
be
"implies
as before.
xn
is,
of course, a proposition; x
a, a prepo
sitional function.
1302
acft =
is
U x (xeacxeft)
"For
Def.
c
a c
ft
equivalent to
e o:
every
x,
is
an a implies x
is
ft
".
nx
(&gt;
cx
e (3)
is
a formal implication.
is
It will appear, as we proceed the logic of the formal implications and formal
mine the
1303
(
classes.
ft)
= U x (x e a =
"For
e (3)
{
Def.
a
to
=
x
ft
is
equivalent to
every
is
member
a c
of
/3
".
a =
e.,
ft
ft
i.
consist of identical
members.
propositions.
a,
ft,
and a
But
the logical product of two classes, and the logical sum, are not assertable
relations.
They
axft
= x[(xea) x(xe
of
ft)}
ft,
Def.
is
The product
which
this
two
classes,
"x
a and
the class of x
is
propositional function
is
is
an a and x
ft".
true constitute
a*
ft,
both as and
/3 s.
The relation x between a and ft is, of course, a different relation from x between propositions or between propositional functions. A similar remark applies to the use of + which will represent the logical sum of two classes,
,
two propositions or propositional functions. This double use of symbols will cause no confusion if it be remembered that a and ft, and 2(^2), etc., are classes, while x e a is a propositional function,
as well as of
(&lt;pz)
and x n
a,
a c
ft,
and a =
ft
are propositions.
263
a+
of
/3
= x{(xea)
classes,
+ (x6(3)}
Def.
is
The sum
two
x
is
a and
/?,
the class of z
,
of the two,
an
is
and x
an a or x
is is
a
/3.
is
The negative
1306
of a class
a = x(xta)
The negative of a is the class of or s for which x is an a The "universe of discourse", 1, may be defined by the
device of selecting
some prepositional function which is true for all values of the variable. Such a function is (f.r cfz), whatever prepositional function x may be.
1307
1 is
= x({xcx)
xs
x
s.
Def.
the class of
all
for
which x implies
"nullclass",
x.
Since this
is
always true,
1.
1 is
the class of
The
Def.
0, will
be the negative of
1308
=
is,
1
That
by
1306,,
= x(txcx), and
s will
/3
since
(fac^)
is
This will be tt x #r). c$x), and a = assertable relations of establishes at once the connection between the
2(^2)
and
Then, by 1301, x n
==
(&lt;px
==
&lt;px
n.
Ux
(&lt;f&gt;x
classes
To
illustrate the
way
in
which
from that
of
propositional functions,
It will
prove a number of typical theorems. be convenient to assume for the whole set of theorems:
we
shall
a
131
= 2O),
]8
2(*z),
&(&)
= x({xc{x). = 1. Hence [1306] = x (x * 1). = f.r c c. Hence = x (s*.r [13 01 07] x e 1
f .r)
132
n x (xel).
[130106] x n el
Hence
U x (x

1)
== (x n cxn). = U x (xc{x).
But
[2
2]
fr n c
rr n .
Hence
a
n,(^ c
x)
is
member
.
"universe
of
discourse".
the range of discourse", 14 This defines, not the universe of discourse, but "universe With 1 so defined, propositions which invol of significance of the chosen function, f whenever t, etc., a the classes (**), (*), etc., and 1, will be significant
&lt;p,
if
significant.
264
133
II x
A
(.eO).
[13 01 06 07]
a n
eO =
(f.r n cf.r n ).
Hence
[32] ~(x
f.r.
n e
0)
(r n
c far)
But $x n c
For every
class.
Hence
that x
[1023]
e
n x (fx
fr),
and
is
H x (a;
0).
x, it is false
no individual
member
of the null
134
cl.
[130106] X n tl
Since a
frr n Cf.r n ).
e
%(&lt;pz),
[1301] x n
ipx n 
c [*c n c (fz n c &)]. [933] r n ). Hence since f.r n c f.r n n c ($x n c c ({x c{x)], and [132] a cl. Hence [1023] tt x
cf.r n )
,
&lt;p.T
On
[&lt;px
Any
class,
a, is
It will
be noted
make
c $x n
use of 933,
is
true proposition
it is
is
implied
n.
by any
proposition".
$x n
true.
x,
Hence
implied by
&lt;px
And
whatever value of
xn
may
be, therefore,
But
(px is
fa:
which determines
Hence
&lt;px
n is
xn
a,
and
xn
cx
n is
xn
e 1.
Conse
And by
135
Oca.
[91]
O n e 0)
is
e
equivalent to (x n
0)
0.
Hence
[133] (x n
0)
0,
and
a,
[932]
(x n e 0)
&lt;px
n.
Hence
Hence
[1301] x n
[1302]
is
eOcx n e
a.
and [1023]
U x (xeQcxe
a).
The
"A
nullclass
contained in every
class, a.
In this proof,
(fa: n
we
use 932,
hence implies
and
&lt;px,
and any proposition". But (fxcfx) is the function which determines 0; n the function which determines a. Hence c a.
cfa: n )
is false,
The
proofs of the five theorems just given are fairly typical of those
which involve
and
1.
of propositions
of classes
make more
and formal
implications or equivalences.
following:
may
be illustrated by the
265
= n x ( (px c \j/x) [1302] z((pz) cz(^z) = U x [xe z((pz) c.ce = (px n and x n e z(\f/z) = $ [1301] x n e z((pz} ex e (^z)] = U x Hence [21] U t [x e
(pz)
z(^z)
z(&lt;pz)
(&lt;px
c
\f/z"
"The
class
determined by
"For
(pz is
is
equivalent to
every
x, (px implies
[1303]
[1301] x n
e
=
[z(&lt;pz)
z(^z)]
&lt;px
= Hx
[xez(&lt;pz)
aez(^z)].
.
z((pz)
e
==
n,
e
and x n
z(^z)]
z(^z)
(&lt;px
Hence
"The
H x [x
=
z(&lt;pz)
= Hx
= \l/x n = fa)^z"
class
determined by
"For
&lt;pz
is
is
equivalent to
every
x,
&lt;px
equivalent to
$x".
[1066]
Il x [x e
z(&lt;pz)
Hence
Hence
139
II x (a: e
ex
e
e /3)
ex e z(tz)] = n x [(.i e
==
E x ([.r
c(x
e
e
z(^z)] c[.r
z(^z)]).
/3)
a)].
.T e
[130106]
(a;
a)
a; e
0)
/3.
[1302]
(c/3)  (/3c).
[(aC8) x(/3c T )]
[136] (a c
j8)
c(aC7).
= Ux
c
(&lt;px
c ^ r),
t
(8
7)
= H x (tx c
^.i),
and (a c
7)
And
The
[1065] [U x
"is
(&lt;f&gt;x
\f/x)
x
is
n x (^.r
.r)]
n x (^a;
is
c ^).
the
first
relation
contained
in"
transitive.
is:
139
form
of the
syllogism in Barbara.
1391
= n x (^c^.r).
&lt;px
[1301] (x n
a)
==
n,
and
(x n
0)
^
.
n.
And
If
is
^.r n ] c\j/x n
the class a
a
&
and
.r n
is
member
of a, then
.r n
member
[(a
of
1392
]8)
x(/3
\&lt;2
==
7)]
c (a
=:
7)~
[lo/J
Py
iix^cp.c
/&gt;
\fj
And
The
[1068] [H x (^,r
^.r)
xH x (^.r =
^.r)]
cH x (^ =
^r)
last three
theorems
266
between formal implications and the relations of classes. 136 and 137 = 0. Similar alternative defini are alternative definitions of a c /3 and a
tions of the other relations
would be
l5
:
[z(&lt;pz*)]
z(&lt;pz)
z(&lt;pz)
We may
way
by
in
which
TwoValued Algebra,
since
it
gives,
1023, a
formal implication or equivalence, gives a corresponding proposition con cerning classes. We choose for this example the Law of Absorption.
1392
[a+(aX]8)]
a.
[130405] [a+(aX8)]
= x{(xea)
x/3)]}
+ [(xea) x(.re/3)]}.
Hence
[1301]
{x n e
[a + (a
=
But
But
[1303]
{(x n
ea) +
[(x n
ea)x(x n
0)]}
==
e(3)}}.
(1)
{[+(
=
x/3)]
=
e
a}
a) + [(x
e
n,[{(.T
a) X (x
,
(x
)J.
(2)
[1301]
(x n e
a)
=
&lt;px
n,
(x n e 0)
= $x n and by
+
(&lt;px
(2),
{[a+(a*{3)]
=
+
a}
But
[54]
[&lt;pX
+(&lt;pX n X\l/X n )]
= Ux =
xifrx)]
{[&lt;px
xjx)]
=
&lt;px}
V%n&gt;
Hence
[1023]
TL x
=
&lt;px}.
{[&lt;px
(&lt;?x
two
= a and
n a {[#c+(#ex#r)] =
Once
this connection
is
of the
Two Valued
/3)]
namely
54, (p + .p q)
p,
for
g,
and then
generalize,
by 1023,
to
An
cedure will give, for most theorems of the TwoValued Algebra, a corre sponding theorem of the calculus of classes. The exceptions are such
propositions as
p =
its
(p
or an equivalence.
which unite an element p with an implication In other words, every theorem concerning classes can
1),
be derived from
We may
15
As a fact, these definitions would be much more convenient for us, but we have chosen to give them in a form exactly analogous to the corresponding definitions of Principia (see
i,
p. 217).
267
set of postulates
will
from the
by deriving the
II.
prove that, beginning with the Two Valued Algebra, as a calculus of propositions, the This procedure may have the appear calculus of classes may be derived. ance of circularity, since in Section I of this chapter we presumed the
propositions of the BooleSchroder Algebra without repeating them.
Chapter
This
But
the circularity
is
is
a distinct
system.
The
The postulates of the existence, in the system, of a when a exists, of a x when a and exist, and of the class 0, must be supposed satisfied by the fact that w e have exhibited, in their definitions, the logical functions which
r
determine a x
142
(a Xa)
(3,
a, and
a.
e
O.
16
[1301] x n
=
&lt;fX
n.
Hence
Hence
[1304] x n
[1303]
e(a*a) =
a]
[(x n
a) x (x n
a)]
[(aXa) =
x
= H x {[.re
(a x a)]
= = x = Ht
(&lt;px
^.r n ).
a]
[(&lt;px
==
&lt;px)
&lt;&].
But
[12]
(&lt;f&gt;x
=
&lt;pxn)
&lt;px
n.
Hence
143
[1023]
(/3
Ut
a).
[(&lt;px
==
&lt;px)
&lt;px].
(ax0) =
[1303]
[(ax0)
[x e (ft
x a)]}
a)]}.
[130104]
Hence
= (0 x a)] = nj(#rx#r) = [1301] [(ax/3) = (#r x ^r n ). But [13] n xtx n = (^ x ^.r)]. Hence [1023] U x x^.r)
(&lt;px
(*x x
&lt;px)}.
[(&lt;f&gt;x
144
(aX0) X T
[1303]
==
ax(0x 7 ). [(X0)X 7 =
==
X(0X T )]
x
(.r
{a?e[ax(0x 7 )]}J
e
7)
==
J {
(.r e
a) x
[(.r
0)
(.r e
7 )] }J.
[130104]
[(x0)
x 7
 ax(0x 7 )]
is possible when the satisfactory derivation of these existence postulates functions is treated in greater detail. See Principia, I, pp. 21718. theory of propositional
16
more
268
A
But
[104]
(&lt;px
xtx n )
{[(&lt;px
xx
=
&lt;px
n x(\f/x n
Hence
145
[1023]
0.
Hx
x^x) x &]
=
[&lt;px
axO =
[13101] .TncO
= (x n eO)} = (fzcMOj. n.u^xozc^)] But [22, 901] (r.rcfar B ) = 1, and [32] ({x n c{x n ) = 0. = = (fr cc). Hence [15] n x(x n c x n )] Hence [1023] n s {[^r xfrr c^)] = (fa; cfr)}.
[130304]
eQ)]
=
[&lt;px
[1301]
0, in
is
the
of the
TwoValued
method
of
now
what abbreviated.
1461
[(x8)
0]
c[( a x/3)
a].
is
The theorem
equivalent to
But
[133]
Il x
IL x [(x e 0)
0].
equivalent to
(xeQ)]c[(&lt;pxxifrx)
[Or x i/a*) =
is
But
[133]
H x (xeQ), and
Ux
{[(&lt;px
hence [91]
TL x [(x e 0)
= =
&lt;px]}
0].
equivalent to
xfr)
=
0]
0]
But
[161]
[(&lt;f&gt;X
n X\j/Xn)
C[(&lt;px
&lt;px]}
&lt;pX
n }.
Hence
1462
[1023]
Q.E.D.
{[(Xj8)
a]
x[(x/3) =
is
a]}
c (a
The theorem
But
equivalent to
==
n,[{[(#rxr)
[133, 91]
^]x[(^x^) =
0)
&lt;px]}
c[&lt;px
(x
0)]]
n x [(.r
is ==
0].
equivalent
to,
n,[{[(#cx^.T)
^i ]x[(^x^.r)
==
&lt;px
^r]}
==
c(&lt;px
0)]
But
[162]
{[(?.rxr n )
Q.E.D.
n]
x[(^x n x^x n }
&lt;px
n }}
c(^.r n
0).
Hence
[1023]
The
definition, 1
of
in
this section.
definitions of
148
The theorem
is
equivalent to
Il x [(&lt;px
\f/x)
= ($x
x\f/x)].
269
But
[18]
(&lt;px
+tXn) =
Q.E.D.
(&lt;?x
x\l/x n ).
Hence
149
(aCjS)
[1023]
[(ax/3)
is
a].
V/.r)
The theorem
But
[19]
(&lt;f&gt;x
equivalent to IIjOz c
is
= Uz
tt )
[(&lt;px
x$x)
==
&lt;px].
c^x n )
equivalent to
[(^cx^ic
==
v*c n ].
important
the BooleSchroder Algebra, as a logic of classes, are two: (1) The BooleSchroder Algebra lacks the erelation, and is thus defective in application,
since
it it
which
cannot distinguish the relation of an individual to the class of of which is con is a member from the relation of two classes one
(2)
The theorems
of the
BooleSchroder Algebra
of the cannot validly be given the completely symbolic form, while those functions can calculus of classes derived from the calculus of prepositional
17 be given this form.
V.
THE
is
LOGIC OF RELATIONS
The
tions of
logic of relations
two or more
variables, just
derived from the theory of propositional func as the logic of classes may be based
of one variable. upon the theory of propositional functions in extension when we logically exhibit the A relation, R, is determined
If to y. y) class of all the couples (x, y) such that x has the relation then x $[&lt;p(x, y)} is the relation "parent of of is
&lt;p(x,
"
represent
"x
parent
y",
is This defines the relation in extension: just as the extension of so the the class of all those things which have the property of being red,
"red"
"parent
of"
is
the class of
all
a property that is the extension of the relation is, couples (or triads, etc.) of a certain class; The calculus of relations, like the calculus thus, the class of couples itself.
relation
is
of propositions,
"Oftentimes,
and
of classes,
is
a calculus of extensions.
i,
the propositional of classes have been represented in the symbols of the BookSchroder Algebra. calculus has been treated otherwise than as an interpretation of to the system, tl in such a case, if these symbols are regarded as belonging But
is
invalid.
270
A We
R meaning
etc.
the class
R to
y.
R =
This notation
is
x y(x Ry),
S = w z(w S
z),
simpler
S = w
z[\[/(iv, z)],
but
it
R = x A triadic
y[&lt;p(x,
y)],
relation,
T, will be
such that
T = x
or
y z[T(x y
}
z)}
is
(x, y, z)
for
T(x, y, z)
true.
But
all
dyad to an individual
y, z),
that
is
to say,
can be treated as a
.
two variables, the couple (x, y) and z, or x and the couple (y, z) This follows from the considerations presented in concluding discussion of the theorems numbered 11, in section III. 18 Similarly, a tetradic relation
function of
15 01
(x, y) n e z
w[R(z,
ic)]
R(x, y) n
Def.
lence of
&lt;p(x,
y) n
and
&lt;p(x
y a ) becomes important.
For
this allows us to
as a function of one or of
two
variables, at will;
(xm
y n )ezw(zRw)
xm
R yn R yn
Def.
or extension, of the relation deter
true.
field,
mined by
1502
R w)
is
"
means that x m
is
RcS = U
x, v
[(xRy) c(xSy)].
Def.
This definition
(ac/3)
because, by 1501, (x
= Hx
y)
e
(xe
acxe
(x
/3)
R y)
,
is (x,
and
y) is (x, y)
S.
similar
definitions.
y)].
(R
S)
= nx
y [(x
Ry) = (xS
Def.
and
for every
x and every
y, (x
R y)
ff.
271
1504
R xS =
x y
[(x
Ry)* (x
y)].
Def.
product of two relations, R and S, is the class of couples (x, y. such that x has the relation R to y and x has the relation S to y. If R is
The
logical
"friend
of",
and S
is "colleague
of",
R x S will be
Def.
S,
is
"friend
and colleague
of")
1505
R + S = xy[(xRy)
sum
of
y
+ (xSy)].
The
will
logical
tw o relations,
R and
(x, y)
such
S to
y.
R+S
be
"
Either
fl of or
of".
1506
R = xy(xRy).
the relation of x to y
Def.
R
is
when x does not have the relation R to y. note that R x S, R + S and # are relations: x(R x *$)#,
and the
"nullrelation"
The
"universalrelation"
1507
x y
[?(x,
y)c{(x,
y)].
Def.
is
a function,
f,
such that
c(x,
y),
i.
e.,
in case x
relation.
1508
1.
Def.
for relations different
from their
But these
meanings ot
every theorem
involving functions of one variable, there is functions of two variables, due to the fact that a function
y)
may
be
for
regarded as a function of the single variable (x, y). Consequently, each theorem of the calculus of classes, there is an exactly corresponding
theorem
We
may, then,
cite as illustrations of
to hold
for
and no proofs
will here
be necessary.
theorems of Section
are given
III,
numbered
numbered 10
151
= x$[t(x,y)c(x,
nullrelation
i.
The
is
the relation of x to y
when
it is
e.,
when
x has no relation to #. 19
Of course, there
no such
couple which can significantly be called a couple. As in the case of the 1 and of the class calculus, the 1 and
of relations, defined as
272
152
A
nx
,
y [(x,
y)
e 1].
Every couple
is
member
(dyadic) relation.
153
nx
[(x, 2/)eO].
No
154
155
jRcl. Ocfl.
relation, R, is implied
Every
by the
nullrelation
relation; or, whatever couple (.r, y) has the nullrelation has also the relation 72, and whatever couple has any relation, R, has also the universalrelation.
156
(RcS) =
S".
tt x
For
relations,
RcS
By
is
"
implies
S"
than
"
is
con
tained in
if
156,
R to
y,
means "For every x and every y, implies then x has the relation S to Or R implies
S"
"
y".
S"
means
"Every (x,
y) couple related
y [(x
by
by
S".
157
(R =
S)
= Ux
Ry) =
(x
y)].
Two
relations,
R
S,
and
S, are equivalent
when the
couples related by
JR.
are
also related
by
and
is
always a reciprocal
relation
c).
158
(RcS) = (ScR).
If the relation
when S
is
absent
also will
be absent.
159
[(RcS) *(ScT)]c(RcT).
implication of one relation
The
by another
is
a transitive relation.
1591
If
[(RcS)x(x m
Ry
)]c(x m
Sy n ).
this couple are
implies
related also
1592
[(R
x(S
==
T)]c(R =
is
T).
The equivalence
If it
of relations
transitive.
be supposed that the postulates concerning the existence of rela tions are satisfied by exhibiting the functions which determine them, then,
we have
nificant,
defined them, are such that propositions involving them are true whenever sig and significant whenever the prepositional functions determining the functions in question have the same range.
273
we can
sitional functions.
from the theorems of the calculus of prepo The demonstrations would be simply the analogues of
may
be omitted.
162
163
(RxR) = (RxS) =
R.
(S xR).
164
165
1661
R].
x[(RxS) =
R}}
c(R =
0).
These theorems
may
In
fact,
"calculus
of
just that
So
far,
But there
analogies,
many
which
classes present no
h. Fact, and these peculiar properties are mosl important, of mathematics, worked out by Peirce, Schroder, the logistic development Peano and his collaborators, and Whitehead and Russell, has de
Frege,
While pended very largely upon a further study of the logic of relations. manner we can do no more, within reasonable limits, than to suggest the of this development, it seems best that the most important of these proper
ties of relations
But even
this outline
cannot
Every
1701
If
*R =
y x (x Rij).
Def.
R, to
x.
It follows at
that
xm
because
(x m
R yn = R
yn
*R x m
e
"R
Ry
n)
(x m y n )
R =
(y n x m )
is
Vn
"R
xm
&gt;
The converse
of the converse of
R.
w(wfl)
= R
19
274
since
A
=
(*&gt;R)
R x) =
R.
(This
is
not
proof:
n,
[(*, y) e
~(~R)
(x, y) e
R]
But
it
is
may
be given.
In general,
we shall not pause for proofs here, but merely indicate the method of proof.) The properties of symmetrical relations follow from the theorems con
cerning converses.
versal relation,
(x 1 y)
1,
The uni For any symmetrical relation T, T = and the nullrelation, 0, are both symmetrical:
*&gt;T.
[rO, y) cf(s,
y)]
==
==
[f(y, x)
c{(y,
.r)]
(y I x)
(The
"1"
in the
middle of this
proof
is
propositions.
(
Similarly for
in the next.)
y)
[r(.r, y)
if
$(*, y)]
[f (y, x)
{(y, x)]
(y
a)
It is
obvious that
two
equivalent:
(#
(R
$)
in
Not
quite so obvious
is
the equivalent of
(RcS),
terms of
Instead
*R and
^S.
We
c^).
we have
(RcS) = (*R c
for
(fl
S)
y [(y
"R
,S)
= H Xf
y [(.i
E y)
(x
y)]
= Ux
is
x)
(y
^S
x)]
= (vRc^S)
"
Parent of implies
of".
ancestor
of"
equivalent to
"Child
of
implies
descendant
The converses
of
compound
relations
is
as follows:
(/JxS)
for
= vRx^S
(x
(R x S)y
= y(R
x S)x
(y
R x) x(yS x) =
y ) x (x ^S y)
=
If
x(*Rx"S)y
is
employer and
by".
.r
is
"employee
of
and exploited
Similarly
If
is
either
employer or benefactor of
y,
the relation of y to x
is "either
of or benefitted
of relations
concern
"relative
sums"
and
275
These must be distinguished from the nonrelative sum and product of relations, symbolized by + and x The nonrelative of "friend and "colleague is "friend and product colleague
products".
.
of"
of"
of":
is
"friend
of a colleague
of":
of".
Their nonrelative
sum
by
is "either
friend of or colleague
of".
their relative
sum
is "friend
of
every noncolleague
We
shall
and S
S, their relative
sum by
R t S.
Def.
relation
z
1702
R\S
to y
for
(x, z) when for some y, x has the and y has the relation S to z. x is friend of a colleague of some y, x is friend of y and y is colleague of z.
is
when,
1703
+ (ySz)}}.
Def.
R t S is the R to y or y
It is
when, for every y, either x has the relation z. x is friend of all noncolleagues of z
is
is
friend of y or y
colleague of
z.
noteworthy that neither relative products nor relative sums are commutative. "Friend of a colleague is not "colleague of a friend
of"
of".
Nor
is "friend
of".
of all noncolleagues
of"
the same as
"colleague
of all non
friends
But both
R\(S\T) = (R\S)\T
for
2 t {(wR
x)
x X (S T)z]
[
2 X (w R x) x z v [( x Sy)x(yTz)]} = 2 y 2 x {(wRx)x[(xSy)x(yTz)]}
=
{
of a (colleague of a neighbor
of".
of)"
is "(friend
of a colleague) of a
neighbor
Similarly,
"Friend
R
all
t (S t T)
= (R
t S) t
T
of)"
is
"(friend
of all
noncolleagues) of
nonneighbors
of".
De Morgan
ucts.
Theorem holds
(R\S)
for
= fltS
=
}
{
2 t [(x
Ry)*(yS z)]
tf
[(*
R y)
x (y S
z)}
276
A
of
"friend
The negative
Similarly,
of a colleague
of"
of all colleagues
(nonnoncolleagues)
of".
(R t S) = R S
of
"friend
The negative
colleague
of".
of all noncolleagues
of"
is "nonfriend
of a
non
for
x *(R S)z

= z(R
S)x
If
is
fitted
z,
z to
"
is
bene
Similarly,
If
^(R t
all
/S)
= ^S
^R
a;
is
hater of
nonhelpers of
by".
z,
the relation of z to
is "helped
by
all
relative product
is
for
x[R (S + T)]z
\
X ^^2/
f 1
(T v**
7? tl/
ii\ c//
(11 L\c/
F
Si
?^
/
+ +
T?y
\"
y 2^1 /J
1
)
= 2 y {[(x Ry)*(yS
Similarly,
"Either
z)]
[(x
Ry)x(y T z)}}
(R + S) T = (R T) + (S T)
\ \
\
of"
is
the same as
"either
friend
of".
is
the following:
It holds since
x[R (S x T)]z
\
T)z]
Z y {(xRy)x[(ySz)x(yTz)}}
and
since a x (b xc)
(a
x6) x (a xc),
=
z y {[(x
2,[(.T
Ry)*(yS z)] x [(x Ry)x(yT z)] Ry)*(yS z)] x ? y [(x Ry)*(yT z)]
And
is
277
is
z,
then x
is
student of a colleague of
cause
"student
The converse
colleague"
of a friend
and
implication does not hold, be requires that the friend and the
friend and student of a col colleague be identical, while "student of a where S y is repeated, last step in the proof does not. (Note the league"
,
(flxS) T c (R\ T) x (S T)
\
instead of
are
more complicated
is
no particular importance, but the relative product In terms of this idea, "powers" of a relation are a very useful concept.
The
relative
sum
is
of
definable
# = R
2
JR,
# = R
3
2

R,
etc.
transitive relation, S,
S n cS.
The
predecessors
S 2 c 8, and hence distinguished by the fact that of predecessors ... of x are predecessors of predecessors of the powers of a relation plays a This of x.
is
conception
prominent
ties of
and
of the
fundamental proper
the
of
number
20
series.
By
method
"mathematical
induction"
pletely deductive.
were not given were letters by which relations were symbolized separate treatment. The Any also interpreted as relative terms by a sort of systematic ambiguity. class of entities which have that relation also stood for the
In the work of
De Morgan and
Peirce,
"relative terms"
relation
symbol
to something.
But
in the logistic
development
of
Mathematical
relative
The "domain" of a given separate treatment which they really require. relation R to some that is, the class of entities which have the relation, R be defined as follows: fl, which can thing or othermay be symbolized by
1704
DR =
of
&[2y(xRy)].
Def.
"For
The domain
x has the
some
y,
be the class
y".
of employers.
The
20
21
"converse
i,
domain"
of fl
that
is,
See Principia,
See
i,
Bk. n, Sect. E.
*33.
The notation we
cipia.
278
R may
be symbolized by
CL
R =
x,
y[2,(xRy)].
Def.
is
The
will
converse domain of
the class of y
"For
some
a
to
y".
R,
R. + (yRx)]}.
all
1706
C R = xi2 y [(xRy)
The
field of
will
If
be the class of be
"employer
the relation.
of",
CR
is
the class of
all
those
who
are
The elementary
properties of such
"relative terms"
are
all
obvious:
xn
(yRx n )}
However,
are of
"
development of mathematics, these properties 22 the highest importance. We quote from Prirwipia Mathematical
for the logistic
is
say the relation of less to greater among integers. Then that are less than some other integer = all integers, Q
that are greater than some other integer
case,
DR =
R =
all
integers integers
all
all
integers except
0.
In this
C R = all integers that are either greater or less than some other = all integers .... Thus when R generates a series, C R becomes integer
important.
..."
have now surveyed the most fundamental and important characters of the logic of relations, and we could not well proceed further without
elaboration of a kind which
is
We
here inadmissible.
is
warned
that
About
a pity!
it".
make
beginning
development
in the
of certain
theory
of
number.
development
Peano
Formulaire
p. 261.
279
and
of Principia Mathematica.
is
ment
previously given.
We
more extended and complete than any can here adapt the prophetic words which Leibniz
"I
begin to get a very different opinion I had regarded it as a scholar s of logic from that which I formerly had. in the way you understand it, it is a kind of diversion, but I now see that,
of Philalethes:
mouth
universal mathematics."
VI.
THE LOGIC OF
Principia Mathematica
of the BooleSchroder Algebra
We
the TwoValued Algebra, prepositional functions and the propositions derived from them, and the application to these of the laws of the Two
Valued Algebra, giving the calculus of propositional functions. Beyond with the Twothis, we have shown in outline how it is possible, beginning Valued Algebra as a calculus of propositions, to derive the logic of classes in a form somewhat more satisfactory than the BooleSchroder Algebra, In so doing, we have presented of relations and relative terms. and the
logic
as
much
work
of that
of
the
development which begins with Boole and passes through Peirce to Schroder as is likely to be permanently significant. we have not here expository rather than historical,
Instead,
we have
in the light of
work
of
from the
development
of
will
and VIII
Chapter
II
I.
be able to do so by reference to Sections VII But it seems best here to point out briefly what
we have made.
In the
first
place,
we have
sums or products of Peirce and Schroder avoided etc. But while theoretical difficulties.
explicitly as
still as an actual product, S#c as an actual sum, they did not treat II which result the laws which they give for propositions of this type are those method by from such a treatment. There is no slightest doubt that the and formulated these laws is substantially the one
symbol for a a sum, makes demonstration possible where product, S#r as the symbol of made and, for further otherwise a large number of assumptions must be resorted to. a much more difficult and less obvious style of proof
And
principles,
280
In this part of their work, Peirce and Schroder can hardly be said to have formulated the assumptions or given the proofs.
In the second place, the BooleSchroder Algebra
of
which
is
work
of classes
With this system before in the paper of 1880 that Peirce felt its defects). them, they neglected the possibility of a better procedure, by beginning
with the calculus of propositions and deriving the logic of classes from the
laws which govern propositional functions. And although the principles which they formulate for propositional functions are as applicable to func
tions of one as of
their interest
relatives
two
variables,
for
was almost
entirely in functions of
which
may
The
logic of classes
laid the
foundation
logic with the later and better treatment to be discussed in this section;
and second, to present the logic of propositional functions and their deriva tives in a form somewhat simpler and more easily intelligible than it might
otherwise be.
The
theoretically sounder
logic of Prin
cipia Mathematica
is
and
their derivatives
concerned
seems to us to obscure, by
notation,
much
With regard to this second purpose, we that the development we have given is theoretically
less
obvious.
adequate
its
tory to the
oretical shortcomings.
how much we
logic
is
connection
Formulaire de Mathematiques and of Principia Mathematica. This is easily overlooked by the student, with the result that the sub
the BooleSchroder Algebra and
its
applications
is
seem quite unrelated to the topic which later interests him the development of mathematics. Both the connections of these two
281
and
We
shall
And
we
stick to a single
illustration,
our comparison
will
Book
I,
Principia
Mathematical
is
interpreting an algebra intended primarily to deal with the relations of And it has several defects which reflect this origin. In the first classes.
place, the
same
two
q
different
ways.
per",
We
peg
. .
and
.
,
r,
then
..."
where
and
"
are propositions.
relation
"
But
"if
then
q,
is
supposed to be the
Also,
same
in
which
is
r
expressed by c in p c
is
r,
and per.
system
"either
"and
pc
q and q c
"
is
other
wise expressed
by
and
use
relations.
The
. .
.",
involves
.
the
. .
.",
then
to
.
",
"...
and
or
.
.
"...
is
equivalent
.
.",
and
"...
is
not
equivalent to
.
",
just as
,
any mathematical system may; yet these are + = and =H whose properties are supposed to
,
,
Thus
the
Xor
is
this
interchange
and
..."
and x,
etc.
For the very demonstration of this interchangeability takes for granted the logic of propositions; and furthermore, in the system as developed,
it is
form impossible in most cases to give a law the completely symbolic until it has first been proved in the form which involves the nonsymbolic
So that there
demonstration of the laws of
expression of relations.
in the
no way
in
system.
Another defect
of the
"p
is
The
23
proposition p or
is
symbolized by
method
of
by p =
1,
by p
is
4=
Peano
Formulaire
a sort of
The general intermediary between the PeirceSchroder mode of procedure and Principia. method of analysis and much of the notation follows that of the Formulaire. But the Formulaire is somewhat less concerned with the extreme of logical rigor, and somewhat more concerned with the detail of the various branches of mathematics. Perhaps for this fundamentals which is the dis reason, it lacks that detailed examination and analysis of
tinguishing characteristic of Principia. of the relation D (in our notation, c
):
For example, the Formulaire retains the ambiguity p?q may be either the class p is contained in
q".
the class
laire
282
the negation of
A
p
or
etc.,
"p
by
p,
p =
0,
p =
1,
=}=
1,
etc.
These various forms may, it is true, be reduced in number; p and p may be made to do service for all their various equivalents. But these equivalents
cannot be banished, for in the proofs
that
it is
necessary to
0)
make
is
p =
(p
1)
(p 4= 0),
p =
(p
(p
1), etc., in
order to
Hence
this
redundance
not
altogether
Both these defects are removed by the procedure adopted for the Mathematical Here p = 1, p = 0,
are not used;
instead
its
negative, symbolized
not presumed.
(2)
The
primitive
(1)
elementary propositions,
elementary
the
and
finally,
idea of
but
is
which does not belong in the system a notation to indicate that one symbol or complex of symbols merely
"equivalent
by
definition",
be replaced by another. An elementary proposition is one which does not involve any variables, and an elementary propositional function is such as "notp" where p is an undetermined elementary proposition.
may
The
what would be supposed a proposition may be asserted or merely considered. The sign h prefaces all propositions
idea of assertion
is
just
An
q.
is
such as
"A
is
"
is
undetermined.
pv
The disjunction of p and q is symbolized by At least one of the two propositions, q means
"
The
#101
"p
poq. =
(
~pvq.
q"
Df.
is
is is
(materially) implies
"At
least
one of
the two,
is
false
and
is
a true
is
proposition".
(The explana
ours.)
its
p^q
we
have symbolized by p c
q,
not
converse.
Principia.
(The propositions quoted will be given the number which they have in The asterisk which precedes the number will distinguish
in earlier chapters or earlier sections of this
The
24
logical
i,
product of p and q
is
symbolized by p
q,
or p
q&gt;
See Bk.
Sect. A.
283
*301
"p
p.q. =
true
~(~p
v~ry).
is
I)f.
"It
is
and
is
true"
is
false that at
least
This is false". one of the two, p and Morgan s Theorem in our notation, (p q) =
is,
of course, a
form of
De
.
(/&gt;
(/).
The
#401
"p
(material) equivalence of
p and
q
.
is
symbolized by p
q or
r/.
= ,p?q q^p.
.
q"
Df
is
is
(materially) equivalent to
"p
(ma
terially)
implies q
q)
and
g (materially) implies
would be (p =
= (pcq)(qcp).
Note that
...
...
and
Df
The dots
and brackets.
takes precedence over one, as a bracket over a In #401 we have only one dot after =, parenthesis, three over two, etc. because the dot between p D q and q D p indicates a product a dot, or two
dots,
:,
:
Two
is
by the
same number
of dots
The
*!!
*111
is
true.
Pp.
("Pp."
"Primitive proposition".)
can be asserted, where x is a real variable, and &lt;pxo\f/x can be asserted, where x is a real variable, then #r can be asserted, where x
When
is
a real variable.
Pp.
A
#12
"real
variable" is
such as p in p.
Pp.
p.
\:pvp.?.p.
In our notation, (p + p} c
*l3
\lq*D.pvq.
Pp.
q).
In our notation, q c (p +
*l4
\mpvqmOmqvp.
q)
Pp.
In our notation, (p +
c(q +
p}.
*l5
\:p
v (p vr).
r)]
Pp.
r)].
In our notation, [p +
(q
+
.
[q
+ (p +
r.
#16
h!
^
&lt;7
pv
c
pv
+
Pp.
r)].
In our notation,
(&lt;/
r)
[Q;
(/)
c Q; +
above
is
followed by a
284
sufficient
A
number
is
asserted.
$Kl7
If
is
an elementary proposition, ~p
is
an elementary proposition.
PP.
*171
If
p and
Pp.
&lt;pp
q are
elementary propositions,
pvq
is
an elementary
proposition.
#172
If
and
\I/p
which take
v
&lt;pp
\f/p
is
an elementary proposi
Pp.
list of assumptions. The last three have to do with the method by which the system is developed. By *l7, directly any proposition which is assumed or proved for p may also be asserted to
is
any proposition
or q or
r,
of
~p may be substituted for p or q or r, etc., in the system. By #171, p v q may be substituted for p
to say,
72,
if
etc.
And by *1
"
of the foregoing
symbols
in the
statements"
way
all
By
q,
the use of
three
q
q,
q,
p3
q,
i&gt;
p,
~p v . p v q, ~p v ~ry, etc., etc., may be substituted for p or q or r in any assumed proposition or any theorem. Such substitution, for which no
postulates would ordinarily be stated,
is
is
developed.
of substitution of
which
is
fundamental
is
the substitution
exists.
".
any complex
is
symbols
where such
This operation
covered by the meaning assigned to "... = ... Df Only one other operation is used in the development of this calculus
elementary propositions the operation for which *!! and ^111 are assumed. If by such substitutions as have just been explained there
of
results a
complex
of
and
if
that part of
which the main, or asserted, relation is o the expression which precedes this sign is identical with a
symbols
in
,
may
. . .
be asserted as a
lemma
or
new theorem.
In other
words, a main, or asserted, sign D has, by *!! and *111, the significant then This property is explicitly assumed property of
"If
,
.".
in the postulates.
to be noted
about
this operation of
inference
strictly
much a piece of reasoning as a mechanical, or mathematical, operation for which a rule has been given. No
is
that
it is
285
involved except that required to recognize a previous proposition followed by the main implication sign, and to set off what
operation
is
then,
assertion. The use of this operation does not, that the processes and principles of ordinary logic are tacitly presupposed as warrant for the operations which give proof.
new
mean
1,
#1
1 1,
Precisely
this:
these
is
postulates explicitly
assume
much
not assumed.
To
will
proof or two.
#201
\ipD~p.o.~p
Dem.
Taut M: ~p v ~p
01)]
.
~p
(1)
Vip?~p.?.~p
#12 above.
}
"Taut"
is
~p
is
giving (1).
This operation is valid by #17. Then by the definition #101, above, p D ~/j is substituted for its defined equivalent, ~p v ~p, and the proof is
complete.
*205
\mmqDrm3ipoqm3mp3r
Dem.
Sum [(1)
.
M:
=&gt;
~p v
~p v
.
(1)
O101)]
\: .
Here ~p
is
"Sum"
refers to
*l6, above.
what
*!(&gt;
becomes when
Then, by #101, p^q and pir are substituted for their defined equivalents, ~pvq and ~pvr, in (1), and the resulting
substituted for p.
expression
is
*1
and #111.
*2 06
\l ."3
zr
."3
Dr
Dem.
p,
Dr.D:pDg.D.pDr:.3:p3g
[#2 05] hs
[(1)
.
.3Sg
DrD.p3r
por
(1) (2)
3r
D:p3

s. par
(2)
*111] h:
/&gt;3r/.D:
q?r .0.
286
"
A
is
"Comm
When,
it
in
*204, previously proved, which is p . D . q D r : D : q D . p D this theorem, q D r is substituted for p, p D q for q, and p D r for
r.
r,
becomes the long expression (1). Such substitutions are valid by *l7, #171, and the definition ^101: if p is a proposition, ~p is a proposition; if ~p and q are propositions, ~p v q is a proposition; and p D is the defined If we replace equivalent of ~p v q. Thus p D q can be substituted for p.
ry
etc., (1)
becomes
r)]
}
(q
r)
[(p
D
&lt;/)
D
(/&gt;
(p
ry)
[(ry
r)
D (p D
r)]
But, as (2) states, what here precedes the main implication sign is identical with a previous theorem, *205. Hence, by #111, what follows this
the theorem to be proved can be asserted. sign Further proofs would, naturally, be more complicated, but they involve no principle not exemplified in the above. These three operations sub
stitutions according to *l7,
main implication
substitution of defined
are the only
equivalents;
and
"inference"
The
result
is
of propositions for
that this development avoids the paradox of taking the logic granted in order to prove it. Nothing of the sort is
explicitly stated postulates
And
it
results
from
this
mode
of
completely symbolic, except for a few postulates, #1 1, #17, etc., involving no further use of then "either ... or ...","... and
"if
. .
.",
"
etc M
\^
We have now seen that the calculus of propositions in Principia Mathematica avoids both the defects of the TwoValued Algebra. The further comparison of the two systems can be made in a sentence Except for the
:
==
0, etc., etc.,
redundance of forms, p, p = 1, and the absence of the entities and 1, the two systems
of this part of Principia
are identical.
into a valid
Any theorem
of the
can be translated
TwoValued Algebra, and any theorem of the TwoValued Algebra not involving and 1 otherwise than as {=0} or
can be translated into a valid theorem of Principia. In fact, the qualification is not particularly significant, because any use of and 1 in the TwoValued Algebra reduces to their use as = and = 1 For
{
theorem
as a term of a sum,
and
presence of
as a factor
wise expressed.
other
1 is p.
287
they contain
is
concerned. 25
of our treatment of propositional functions with the
The comparison
same topic
In the
in Principia is
first place,
there
in Principia, the
"theory
of
types,"
which
But we
of
shall
omit con
notation.
Where
we
or
write
U&lt;px,
or
H x (px,
by
Principia has
.
&lt;px.
2x
&lt;px,
and where we write S#r, (or) further and more important difference
.
&lt;px\
may
be
made
{(.?)
clear
citing the
.
assumptions of Principia.
~#r.
~&lt;px.
*901
.?&}. =
(3.r)
.
Df. Df.
Df.
*902
*903
25
~{(3a:)
(x)
.
&lt;px
.?&}. =
fa)
(x)
.vp
&lt;px
vp.
This
may be proved by
are contained
noting that, properly translated, the postulates of each system of the other. Of the postulates in Principia,
^101
is
(peg)
(p +
q),
is is
which
is
^1 2
sfcl
is
is,
(p + p) cp, which
3
pc
p+
(p + q),
which
^1 4
^15
by
^el6
is
is,
by
22.
+
(&lt;/
r)
cq
+ (p +
?),
which
is
22.
is (q cr) c [(p + q) c(p + r)], which is a consequence of our theorem The remaining (nonsymbolic) postulates are tacitly assumed in our system.
531,
by
22.
in
Chap,
n and
a consequence of
in Principia.
2
3
is
is
^43 in Principia.
4
5
of
is
^23
in Principia.
"If
is
equivalent to
0,
then a x
=
0",
hence to x c (a
x),
which
is
a consequence
*327
in Principia,
by *216.
= x), is a consequence of #471 and ^461 in Principia, 161, in the form (x a) c (x a by ^401 and *326. = y)]cy, is a consequence of #471, *516, and 162, in the form [(y a = y)(y a ^221 in Principia. to [(a; = 1)Q/ = 0)] c (x = ?/), hence to (x y) c (x = y), which is an 17 is
equivalent
immediate consequence
18
19
is is
of ^51 in Principia.
^K457 in Principia.
^471
is
in Principia.
(q
901
equivalent to
1)
[p
(p
q)},
hence to q c
[p
(p
q)],
which
is
an
immediate consequence
26
of ^5501 in Principia.
See Principia,
i,
1521.
288
In this
A
last,
two
sides.
note the difference in the scope of the "quantifier" (x) on the If the dots be replaced by parentheses, *9 03 will be
{[(x)
.
^r]
vp] =
{(x)
[tpxyp]}
A similar difference
*904
p
.
(x) or (3.r)
(x)
&lt;px
&lt;px
=
:
(x)
.
pv
.
&lt;px.
Df.
p.
*9 05
#9 OG
*907
(3.r)
v
.
= =
.
(3.r)
&lt;px
Df
p ***
(.r)
.
("3.x)
&lt;px
(3z)
:
pv
(#)
(.T)
:
&lt;px.
Df.
.
&lt;px
mv
.
(3#)
.
(.I)
^
^r
= =
(3#)
^ v ^.
$y v
^.r
.
Df.
*9 OS
(3y)
^
.
(3y)
Df
Besides these definitions, there are four postulates (in addition to those
*9 *9
1 1 1
h: hs
&lt;?x
(3
.
z)
&lt;pz.
Pp.
.
&lt;p
&lt;f&gt;x
^?/
(3
2)
Pp.
is
*9 12
What
is
true.
Pp.
#913 In any assertion containing a real variable, this real variable may be turned into an apparent variable for which all possible values are asserted
to satisfy the function in question.
Pp.
is
By
#901
for
&lt;px.
proved proposition.
is
H&lt;px
our method, every one of these assumptions, except ^912, In our notation,
= S
=
&lt;px,
which
is
&lt;px
substituted
*902
is
2^.i&lt;px.
n^.r, which
is
&lt;?x
substi
tuted for
*903
is
tt&lt;px
P = n x (^c + P),
II X
which which
is
is
*904
*905 *906
is
is
is
is is
(P+
&lt;px),
is is
2 X (P+
&lt;px),
which
$y),
^.r),
1031.
is is
*907
*908
which which
contained in 125,
also contained in 125.
The
&lt;py,
&lt;pz,
etc.,
&lt;px.
we have written
"Ibid.,
i,
letter.
This
is
a valid con
13538.
289
it
&lt;px
determined by not by x, and x is conventions aside indifferent. z in where we should write n is called a "real variable", x in (x) . and (3or) . an "apparent variable". With this it is clear
&lt;pz,
&lt;px
&lt;px
&lt;px,
explanation,
that:
*9l
#911
is
&lt;px
c2&lt;px,
which
c2&lt;px,
is
1021.
is
is
&lt;px
m+
&lt;px
which
an immediate consequence
of 1021,
by 533.
*913
is
"If
whatever value of
is
x, in
&lt;px,
xn
may
be,
&lt;px
n,
then
n^.r,"
and
this implication
without assuming
etc.), (3a)
.
&lt;px
(x)
.
&lt;*r
to be the product of
of
&lt;px
2,
etc. (or
&lt;py,
&lt;pz,
to be the
sum
.
lf
&lt;f&gt;x
2,
"
etc.
&lt;px
as
new
&lt;?x
meaning
a:".
(3or)
.
&lt;px
meaning
&lt;px
for
some values
of
all
questions
in which troubled us and secures the universality of theorems involving prepositional functions without any discussion or convention covering the cases in which the values of the vari
The
method.
all
They
+
are, in general,
if
we had
based
necting
2&lt;f&gt;x
with 2 x
(&lt;px
+ P),
as a sum.
The theory
*11 *11
of functions of
two
y)
=
.
(x)
:
(y)
:
.
&lt;p(
y).
&lt;p(
Df.
,
03
(3z, y)
&lt;p(x,
y)
(3.x)
(30)
y).
Df.
nx
&lt;p(x,
y)
2 XlV
&lt;p(x,y)
The
difference
of propositional functions
which
the treatment in Principia is not necessarily correlated with the difference between our treatment of propositions and theirs. The
of propositional functions
290
might exactly as well have been based upon the calculus of elementary A few minor propositions in Principia as upon the TwoValued Algebra.
alterations
would be
The
different procedure
a difference to be adjudged
Twothere
to the
differences
Neglecting these, we may say that the two methods give the same results, with the following exceptions: 1. There are certain complexities in Principia due to the theory of
types.
2.
3.
In Principia the conditions of significance are explicitly investigated. Principia contains a theory of "descriptions", account of which is
here omitted.
But none
They
are due to
the more elementary character of our presentation of the subject. may, then, say loosely that the two methods give identical results.
We
The
calculus of classes
and
of relations
in the
preceding sections bear a similar relation to the logic of classes and of relations in Principia; that is to say, there is much more detail and com
plexity of theory in Principia, but so far as our exposition goes, the
two are
And
here there
is
It should now be clear how the logic of Principia is related to the logic we have presented, following in the main the methods of Peirce and Schroder. There is much difference of method, and, especially in the case of the cal
is
in favor of Principia.
And
in
Principia there
is
much more
of theoretical rigor
and the
logic of relatives.
But
which we have
expounded goes, the two methods give roughly identical results When we remember the date of the work of Peirce and Schroder, it becomes clear
what
is
our debt to them for the better developments which have since
been made.
CHAPTER
The systems
implication,
is
V
1
all
*
a false
is
statement, p is true and q false/ p cq meaning exactly We have already called attention to the fact that statement".
"implies".
this
Its
"implies"
false
proposition",
and
"A
true proposition
is
implied
based upon an entirely different meaning of implies one more in accord with the customary uses of that relation in inference
is
"
which
and
proof.
We
And we
shall
meaning
it
either the
TwoValued Algebra or
It will appear that the logical import of these two systems is identical. Strict Implication is neither a calculus of extensions, like Material Impli
and the BooleSchroder Algebra, nor a calculus of intensions, like It includes relations the unsuccessful systems of Lambert and Castillon. Strict of both types, but distinguishes them and shows their connections.
cation
Implication contains Material Implication, as it appears in Principia Mathematica, as a partialsystem, and it contains also a supplementary
partialsystem the relations of which are those of intension. The numerous questions concerning the exact significance of implication,
"proper"
meaning
of
"implies",
will
be discussed in
be indicated
how
Strict Implication,
by an extension to proposi
tional functions, gives a calculus of classes and classconcepts which exhibits In this, it provides the their relations both in extension and in intension.
Various studies toward this system have appeared in Mind and the Journal of Phi losophy (see Bibliography). But the complete system has not previously been printed. We here correct, also, certain errors of these earlier papers, most notably with reference to
1
triadic
2
"strict"
relations.
illustrations, see
For further
Logic,"
Chap, n, Sect,
etc.
i,
and Lewis,
"Interesting
Theorems in
Symbolic
(1913), p. 239.
291
292
possible at
all.
I.
QUENCES
The fundamental
Symbolic Logic and
1.
MacColl
its
Applications.
q, r,
They
are as follows:
Propositions: p,
etc.
"p
2.
3.
is
false".
meaning
"p
is
impossible",
or
"It
is
impossible
that p be
4.
The
q
logical product:
is
true".
p xg or p
q,
meaning
"p
and
both",
or
"p
is
true
and
5.
Equivalence:
p =
q,
Systems previously developed, except MacColl s, have only two truthThe addition of the idea of impossibility and values,
"true"
"false".
all of
logical ideas:
p,
"pis
true".
p,
"pis
false".
(3)
(4)
~p,
"p
is impossible". is is
p,
"It
false that
is impossible"
i.
e.,
i.
"p
is possible".
"p
(5)
true".
p,
"It
impossible that p be
false"
e.,
is
necessarily
as simple ideas:
understand p and ~p, and to remember that each such It should prefix affects the letter as already modified by those nearer it. is equivalent be noted that there are also more complex truthvalues. p as will be shown, but  ~ p, to p, etc., are irreducible. p,
sufficient to
p,
We
shall
"
p,
It is
4
false".
,
p be
true"
i.
e.,
"p
is
possibly
Each one
3
of these
complex truthvalues
is
in logic or in
mathematics.
here use a symbol, ~, which appears in Prindpia Mathematica with a different meaning. The excuse for this is its typographical convenience. 4 is possibly is possibly true" and MacColl uses a single symbol for ~p, p,
"p "p
We
false".
The System of
Strict Implication
293
The dyadic
truth values
and the
p q?
Def.
true"
101
~(pq),
are
Consistency.
"It
poq = ~(pq}.
"It
is
would be
"p
and
inconsistent".
possible that
p and
q both be
true",
represents
"p
102
Strict Implication.
p*q
~(pq)(pq).
Def.
103 104
Material Implication.
Strict Logical
pcq =
p+
q
Def.
Def. Def.
p)
.
Sum.
p *q = ~(pr/).
105
= =
(p q)(p
*
106
(p
q)
q) (q
*
Def.
We
itself,
and
strict implication.
makes
it
p =
5 something which could not otherwise be done. q mains a primitive idea as the idea that one set of symbols
But p =
q re
may
Def.
be replaced
by another.
1 07
Material Equivalence,
(p
q)
(p c
q) (q
c p).
These eight relations the seven defined above and the primitive rela = q are the relations divide into two sets, p q, p c q, p + q, and p tion, p q
which
figure in
any calculus
of Material Implication.
We
shall refer to
them
as the
"material
relations",
p o
q,
i
q,
p*q, and p =
".
q involve
the idea of impossibility, and do not belong to systems of Material Impli We may anticipate a These may be called the strict relations cation. of these two sets, which results from the exhibit the
"
little
and
analogy
theorem
~(pq)
shortly to be proved.
Strict relations
:
= (pti
Material relations
pcq =
P+&lt;1
(p q)
o&lt;7)
q)
= (poq) x(qop)
(p
=
"
q)
l~P"&gt;
hesitates to The "circularity" here belongs inevitably to logic. No mathematician If theorem A, then theorem B, of two propositions by showing that prove the equivalence know that a reciprc and if theorem B, then theorem A". But to do this he must already And the italicized relation is equivalent to an equivalence. then . .
"if
. .
.
."
to"
294
The reader will, very likely, have some difficulty in distinguishing in meaning p q from p c q, p A q from p + q. The above comparison may be of assist
J
it
p oq
and p
g,"
q.
We
q,
shall be in
no danger of confusing p o
true".
is
consistent with
with p
"p
and
q are both q
Both p A
and p +
would be read
"Either
p or
q".
But p A
q denotes
"
a necessary connection;
day
is
Monday",
and
+
is
.
4:".
false.
In point of
two
is
propositions,
"Today is
Monday"
and
"2
4",
true;
.
.
but there
."
no necessary connection
in this respect.
between them.
the
"Either
or
is
ambiguous
Ask
members
or 2
of
Monday
"Either
+
or
any company whether the proposition "Either today is = is true, and they will disagree. Some will confine
4"
...
..."
to the
p*q
meaning, others
will
make
p =
it
include
make
between p
and p
q is that
q denotes
an equivalence of logical import or meaning, while p = q denotes simply an equivalence of truthvalue. As was shown in Chapter II, p = q may be and q are both true or both false". Here again, accurately rendered
"p
q,
g,
The
11
If
postulates of the
qi q
p
q are both true, then q
p and
q
true.
12
I
p
true, then
If q
is
true.
1 3
If
p
then p
is
true,
true and
is
true.
14
If
p(qr)4q(pr) p
is
is
true and
p and
are both
true.
15
If
pi(p)
is
true, then
it is
false that
is false.
16
If
(ptq)(qlr)l(plr)
strictly implies q
and
q strictly implies
r,
r.
The System of
1 7
Strict
Implication
295
~pip
If it is
is false.
18
p
"p
J
~q
J
~p
"
strictly implies
is
q"
is
equivalent to
is
impossible
strictly
im
plies
impossible
".
The
as
first six of
J
They
But,
do not, so
we
principles of trans
formation;
(piq)
obvious enough. i J
(
~p
~q)
p implies
is
then
is
possible
implies
possible
is
(
(~p
J
~q)
J
(p
J
q)
If
impossible
is
implies
q
is
is
impossible
.
then
false
implies
false
These two
propositions are
more
"selfevident"
and that
J
and
[(p q)
is
+ (p
[pc.(q
*
&lt;/)]
(p
q)
r)]
[(p
c q)(p c
shall
r)]
a possibility of confusion,
we
put in the
The
1.
operations
are three:
Substitution.
If
Any
proposition
may
be substituted for p or q or
If
r,
etc.
is
p and q are
propositions,
is
a proposition.
Also, of
any
is
by =
2.
either
may
Inference.
is
asserted
and p
is
J
asserted, then q
may
be
asserted.
cation,
3.
pc
Production
It
p and
may
be asserted.
of in proof.
296
In order to
A
make
and particu
we
shall
wish to derive from the postulates their This can be done by the use of
postulate
8 and
its
But
we must prove
number
of simple but
fundamental theorems.
The
first
theorem
be proved in
full
The
21
pqtp
16 [pq/p; qp/q; p/r}:
11 xl2i (p
16,
q*p)
q
is
This proof
may
be read:
for
r,
"Proposition
when p
substituted for p,
qp
(P q
is
for
*
q,
"
and p
P)
The number
Next, in braces,
indicates that
indication of
any
substitutions to be made,
q/p"
"p+ q/r"
to be substituted,
p+
was to
be substituted for
r,
etc.
Suppose we take
6,
which
is
q/p;
qp/p; p/r}.
We
then
(pqlqp)(qplp)
This
since
is
(pqlp).
But
is
11,
and
J
is
12,
we
write 11 x
2 instead of
qiqp)(qplp). This calls attention to the fact that what precedes the main implication sign is the product of two previous propositions.
(p
may
be asserted;
and
may
be asserted, what
it
implies
the theorem to be
is its
be asserted.
The advantage of this way of writing the proofs Yet anyone who wishes to reconstruct the demon
211
(p
= q)4(p4q)
21 {plq/p; qip/q}: (p *
106: (p
q) (q *
p) 4 (p
* q)
q)
(p
The System of
Strict Implication
297
212
(p
= (/HfeHp)
Similar proof, 12 instead of 21.
22
(pi q)
J
(~g
J
~p)
106: 18
[(p4q)4(~q4~p)}[(~q*~P}*(P*&lt;l)}
l)
we
and placed after a lemma which has been established, thereafter in the same proof we write (1), or (2), etc., instead of that lemma. line of proof instead Also, we shall frequently write "Q.E.D." in the last In the first line of this proof, the of repeating the theorem to be proved. substitutions which it is necessary to make in order to get
or (2), etc.,
is
1 8
[(p
q)
* (~q
*
~p)][(~&lt;7
H ~p)
J
(P
&lt;
&lt;/)]
And
in the
second
line, state
ment
is
Such
abbreviations will be used frequently in later proofs. Theorem 22 is one of the implications contained in postulate 1S. of a By the definition, 16, any strict equivalence may be replaced by pair
strict implications.
By
implications
may
2
be taken separately.
221
(~q*~p)4(p4q)
1
:
[(1) in
proof of 2
2]
J
Q.E.D.
This
is
23
(P *
* (~q * P)
11
2 .2
{q/p
p/q}:
qp*pq
q/q}
:
{g p/p;
p
(1)
J
*
[~(/&gt;
&lt;?)
~H
"rfl
102: (2)
= Q.E.D.
pplp
24
pip
12 {p/q}:
1G: l3x(l){ Q.E.D.
25
(p)*p
2i
{p/p}:
p*p
(1)
I
23 {p/q}:
Q.E.D.
251
(p}
=p
25x15 = Q.E.D.
106:
298 26
A
(p
i
g)
J
(g
J
p)
J
g)
i
[(q) H p]
(1)
= Q.E.D.
p)
261
(p
j
g)
i
26 {pip}:
251:
(1)
(g
&lt;
P)
(1)
262
(p
q)
I
(g
p)
i
261 {q/q}: [p
251:
(1)
(g)]
J
(g H p)
(1)
= Q.E.D.
26J
(png) = Htp)
106:
262x26 = Q.E.D.
261x23 = Q.E.D.
and 262 are the four forms
is
264
(p*q)
= (q*p)
23, 26, 261,
of the familiar
106:
Theorems
terms.
27
(~p4~q}4(p4q)
221 [plq; q/p}: (~p J ~g) j (q * p ) 262 {p/q; q/p}: (q J p) H (p t q) 16: (1) x (2) H Q.E.D.
(1)
(2)
271
(p
J
q)
i
(~p
i
J
q)
2.6: (p
?) H (g
p)
i
(1)
(g S p)
(~jM
~g)
(2)
(l)x(2)^Q.E.D.
(ptq)
106: 271
(~
p
J
g)
J
[(p)
J
(g)]
(1)
= Q.E.D.
:
273
(p4q)i(~pi~ q )
271 {p/p q/q} [(p) 251: (1) = Q.E.D.
,
i
(g)] 4 (
p s 
g)
(1)
27J1
(pig)
274
(p^g)H(^p^^)
262 {g/p; ~p/g}
16:
(~g ^ p) H (
^p
i
^g)
(1)
22x(l)^ Q.E.D.
The System of
Strict Implication
299
275
(
~p
i
~g)
i
(p H
ry)
(
~p
J
~q)
J
(~g
I
~p)
(1)
276
(pJg)
= (T
=
(
!&lt;/)
106:
274x275  Q.E.D.
i
277
(p
i
(g
{
p)
(~g
I
~p)
"
"p
implies
g"
is
"
equivalent to
p
"
is
possible
q
is
implies
q
"
is is
possible
is
equivalent to
"q
is
necessary
implies
is
necessary
"
equivalent to
implies
is
false
implies
".
is
false
equivalent to
is
impossible
is
impossible
26277 are various principles for transforming a strict implication. These are all summed up in 2 77. The importance of this theorem will be
illustrated shortly.
28
pq = qp
11
{g/p; pfq}: q
I
(1)
 Q.E.D.
(1)
281
p =
pp
29
r)
i
p(q
r)
(1)
= Q.E.D.
q)
q)
291
p(qr)
(p q)r
28: p(qr)
29: p(rq)
28: r(pq)
= = =
p(r
r(p
(p q)r
most further
proofs.
II.
We