74 T H E FOUNDATIONS OF MATHEMATICS. [Nov.

he supplements this theorem by considering the case when

p = 2. The main results may be stated as follows : If the
Sylow subgroups of order 2m(m > 1) contained in any group G
are either cyclic or contain a cyclic subgroup of order 2m~l
which includes only two invariant operators under one of these
Sylow subgroups, then the number of operators of order 2 in
G is of the form 1 + 4k. When this condition is not satisfied
the number of these operators is always of the form 3 4- 4k.
When m = 1, G contains an invariant subgroup which is com-
posed of all its operators of odd order, and the number of the
subgroups of order 2 may have either of the two forms 1 + 4k,
3 + 4k. This is the only case where the form of the number
of the subgroups of order 2 is not determined by the form of
this number in a Sylow subgroup.
G. A. MILLER,
Secretary of the Section.

T H E FOUNDATIONS OF MATHEMATICS.*
The Principles of Mathematics. By BERTRAND BUSSELL.
Volume I. Cambridge. The University Press, 1903. xxix
+ 534 pp.
Essai sur les Fondements de la Géométrie. Par BERTRAND
RUSSELL. Traduction par A. CADENAT, revue et annotée
par l'auteur et par L. COUTURAT. Paris, Gauthier-Villars,
1901. x + 2 7 4 p p .
1. The Problem. — Pure mathematics has always been con-
ceived in the minds of its votaries and by the world at large to
be a science which makes up for whatever it lacks in human
interest, and in the stimulus of close contact with the infinite
variety of nature, by the sureness, the absolute accuracy, of
its methods and results. Yet what has been accepted as sure
and accurate in one generation has frequently required funda-
mental revision in the next. Euclid and his pupils could
doubtless have complained of the lack of rigor and logical pre-
cision in his predecessors just as forcibly as some modern pupils
of Weierstrass berate their scientific ancestors and companions.
* We may also refer our readers to the review by L. Couturat, Bulletin des
Sciences Mathématiques, vol. 28, pp. 129-147 (1904). So large is the work
of Russell that Coufcurat's review and our own supplement rather than
overlap one another.
1904.] THE FOUNDATIONS OF MATHEMATICS. 75

Euler, finding confusion in the theory of the infinite and infini-

tesimal, proceeded to explain away the difficulties, that others
might be free from the prevailing errors. We cannot accept
his reasoning to-day. At the beginning of the last century the
state of infinite series was lamentable and Cauchy's memoir on
the subject is said to have impressed itself on Laplace to such
an extent that he postponed publishing his Mécanique Céleste
until he became so hopeless of righting things that he gave up
trying to do it. The righting has been accomplished in the pres-
ent generation by Poincaré. Yet we very much doubt whether
Laplace, before hearing of Cauchy's treatment, would have
for a moment granted any possible inaccuracy in his own
methods. Somewhat later Dirichlet treated the problem of
determining a harmonic function from its boundary values and
so careful a mathematician as H . Weber extended the method to
the discussion of the equation A F + X F = 0 without any ap-
parent qualms as to error. Nevertheless, now-a-days, the theo-
retical importance and the practical use of the principles of
Dirichletand Thomson are completely obscured for many by the
too great emphasis laid upon the errors in the original demon-
stration of the principles.
We notice that the advance toward our present rigor has
been made step by step by great men who, however, were no
greater — one might almost say no more careful — than their
fellows working in apparent unconsciousness of the impending
trouble and perhaps even incredulous at first as to its reality.
When will this revision stop ? And whereunto will it finally
lead ? This is the problem of the ultimate foundation of math-
ematics. In attempting an answer one can learn only hesi-
tancy from the past. The delicacy of the question is such that
even the greatest mathematicians and philosophers of to-day
have made what seem to be substantial slips of judgment and
have shown on occasions an astounding ignorance of the essence
of the problem which they were discussing. At times this
has been due to the inevitable failings of individual intuition
in dealing with matters that are still unsettled ; but all too
frequently it has been the result of a wholly unpardonable dis-
regard of the work already accomplished by others. Even
when guarding as much as may be against this latter sin, those
who approach the depths of the subject upon which Russell has
so courageously entered may well expect to hear the warning :
Procul, o procul este profani !
76 THE FOUNDATIONS OF MATHEMATICS. [N~OV.,

2. The Solution. — Says Russell : Pure mathematics is the

class of all propositions of the form "p implies q," where p and
q are propositions containing one or more variables, the same in
the two propositions, and neither p nor q contains any con-
stants except logical constants. And logical constants are all
notions definable in terms of the following : Implication, the
relation of a term to a class of which it is a member, the notion
of such that, the notion of relation, and such further notions as
may be involved in the general notion of propositions of the
above form. In addition to these, mathematics uses a notion
which is not a constituent of the propositions which it considers,
namely the notion of truth.
This is probably the first attempt to give a complete defini-
tion of mathematics solely in terms of the laws of thought and
the other necessary paraphernalia of the thinking mind. Some
there are who, under the influence of arithmetic tendencies,
might be tempted to give a decidedly more superficial definition
in terms of integers. Some might regard a complete defini-
tion as impossible. The fact that a definition such as the above
may be given — and it is the purpose of Russell's Principles
of Mathematics to demonstrate that the definition is not illusory
nor too small nor too large — is attributable to two things :
first, the more careful discrimination of what pure mathematics
is ; second, the extraordinary development of logic since Boole
removed it from the trammels of medieval scholasticism.
He to whom the present highly developed state of the foun-
dations of mathematics is chiefly due is Peano — one whose
work unfortunately is very little known and still less appreci-
ated in this country. True, Leibniz had long since done
much and of recent years has been ably expounded by L. Coutu-
rat ;* true it is, too, that Boole had freed us from Aristotelian-
ism and that C. S. Peirce f and Schroeder had carried the
technique of logic much farther ; but they had never accom-
plished that intimate formal relation between logic and all
mathematics which was the necessary precursor to a yet more
intimate philosophic relation and which has been brought about
by Peano aided by a large school of pupils and fellow-workers.
The advance has been made largely by introducing into sym-

* La Logique de Leibniz, Paris, 1901.

t T h e fundamental importance of the logic of relations (see infra) was
emphasized by C. S. Peirce in 1880-1884 : but it is only beginning to have
its full effect.
1904.] THE FOUNDATIONS OF MATHEMATICS. 77

bolic logic such a simplification of notation as to relieve it of its

unwieldiness and to allow its development into a powerful
instrument without which one can hardly hope to get the best
results in the treacherous though treasure-laden fields of the
foundations of mathematics. Poincaré, to be sure, in his re-
view of Hubert's Foundations of Geometry * spurns this pasig-
raphy, characterizing it as disastrous in teaching, hurtful to
mental development, and deadening for investigators, nipping
their originality in the bud. However much we may agree in
the first statements (see § 7, page 91), we had best be cautious
in accepting such sweeping statements as the last, even from so
great an authority — especially in view of the fact that,
equipped with this pasigraphy, the Italian investigators, Peano
and his pupil Pieri,f with some rights of priority, had
given a more fundamental logical J treatment of the subject on
* Translated in BULLETIN, vol. 10, p. 5 (Oct., 1903).
f " I principii della geometria di posizione," Memorie della B. Academia
delle Scienze di Torino, vol. 48, pp. 1-62. And, " Delia geometria elementare
come sistema ipotetico-deduttivo ; Monografia del pun to e moto," ibid., vol.
49, pp. 173-222.
t While we appreciate and admire as much as anyone can the beauties of
Hubert's famous Grundlagen der Geometrie, we fail to see how the his-
torical facts can justify what Poincaré says (I. c , p. 23) : " H e has made the
philosophy of mathematics take a long step in advance, comparable to those
which were due to Lobachevsky, to Riemann, to Helmholtz, and to L i e . "
Poincaré makes the point that Hilbert regards his geometric elements as mere
things and on this seems to rest a large part of the praise (I. c. bottom p. 21
and top p. 22). If this be so, it ought to be mentioned as a matter of history
that Peano, in 1889, in his Principii di Geometria took precisely this stand, p.
24. In 1891-2, Vailati, Bivista, vol. 1, p. 127, vol. 2, p. 71, again formulated
the principle in words. By 1897 the Italian school had gone as far beyond this
point of view as to make it a postulate that points are classes — thus showing
a twofold advance, once in recognizing the presence of a postulate, again in
using the word class so as to bring the reasoning into form dependent upon
precise logical processes alone. It has also been said that the idea of the in-
dependence of the axioms was due to Hilbert. As a matter of fact in 1894,
Peano, " S u i fondamenti della geometria," Bivista, vol. 4, pp. 51 et seq.,
states the problem and, by actually setting up simple systems of elements,
proves the independence of certain axioms from certain others. So that by
1899 the idea and method were both five years old at least. Again, in 1889,
Peano laid it down as a principle that there should be as few undefined
symbols as possible, and he used but few. In 1897-9 Pieri used but two for
projective geometry and but two for metric geometry, whereas Hilbert was
using a considerable number, seven or eight. (The idea of compatibility
seems to have been first stated clearly by Hilbert.) There still remains in
the Grundlagen der Geometrie matter enough for the amplest praise. The
archimedean axiom, the theorems of Pascal and Desargues, the analysis of
segments and area*, and a host of things are treated either for the first time
or in a new way, and with consummate skill. We should say that it was in
the technique rather than in the philosophy of geometry that Hilbert created
an epoch.
78 THE FOUNDATIONS OF MATHEMATICS. [Nov.,

which Poincaré was writing than is to be found in the work

he was praising so highly. In the fields of arithmetic and
algebra, too, Burali-Forti and Padoa, adherents of Peano, had
reached a point far beyond the widest view of the chief of the
German school that deals with the same subjects.* Further-
more, on this one point Poincaré may not be regarded as an
authority ; for his own work f in the field should be charac-
terized as subjective rather than objective, speculative and
suggestive rather than purely logical. J Anyone who is
acquainted with the articles presented to the Philosophical
Congress at Paris in 1900 by Peano, Burali-Forti, Padoa,
and Pieri, cannot be convinced that* these authors had become
deadened, and the artificiality of their system is by no means so
certain as it might be. Since then, our author, Russell, has
simplified and improved the older work of C. S. Peirce on the
theory of relations, adapting it to the system of Peano, and has
produced a coherent treatment of the great problems underlying
mathematics. In view of accomplished facts one inclines more
readily to the praise given by Whitehead : " I believe the in-
vention of the Peano and Russell symbolism forms an epoch
in mathematical reasoning." §
3. The Reason. — I t is not hard to detect the reason why
mathematics has thus pushed its foundations back until they
* Compare the papers below referred to in the Bibliothèque du Congrès
International with No. 2 of Hubert's Mathematical Problems, BULLETIN, vol.
8, p. 447. We may refer also to Padoa, VEnseignement Mathématique, vol.
5, p. 85. See also ? 4 of the present review. For our readers, who may
be working on the problem No. 2, we may note—what we unfortunately
failed to note at the time of translating—namely, that a solution along the
lines proposed by Hilbert seems logically impossible. A solution has long
since been proposed in the article here referred to. There are those, however,
who hold that Padoa has gone so far as to overshoot the mark. Hilbert has
again taken up the matter much more searchingly than in 1900. It is to be
regretted that his paper which was presented at Heidelberg, August, 1904,
is not at hand for comparison.
t La science et l'hypothèse, Paris, 1903; and numerous scattered essays.
X That Poincaré seems frequently to have in mind the physical rather than
the mathematical, the psychological rather than the logical point of view can
be seen in several places in his review. On p. 8 he asserts that we know the
axioms are non-contradictory "since geometry exists," And on p. 22 he
seems to complain that the logical standpoint interests the author to the
utter disregard of the psychological. It should be remembered that the first
chief aim of the modern researches on the foundations of geometry is to be
entirely rid of the psychological element — and this for the very reason that
secondly we may decide just what that psychological element must be.
This latter problem belongs rather to the philosopher and psychologist than
to the mathematician.
§ American Journal of Mathematics, vol. 24, p. 367 (October, 1902).
1904.] THE FOUNDATIONS OF MATHEMATICS, 79

have come to rest solely in logic. In the first place mathe-

matical or other reasoning presupposes a mind capable of
rational, that is, non-selfcontradictory ratiocinative processes.
Now it always has been comfortably assumed that we can carry
out such processes if only we are careful enough, that there is
no need of formulating the laws of thought before beginning to
reason, or even that a formulation and analysis of those laws is
impossible.* Where then do the errors creep in? An exami-
nation of some typical cases shows that it is generally through
lack of a sufficiently careful definition of the terms. This
failure properly to define has led to interminable discussions
which from the start could only lead either to nothing or to
wrong results. I n mathematics it is the absence of precise
definition which brings in the erroneous statements concerning
differentiation, continuity, and infinity, with a host of others.
The perception of this difficulty was the origin of the principle
of arithmetization and of epsilon proofs. In the end, how-
ever, after one has really mastered the principles of modern
analysis he seldom needs the actual presence of epsilons to
establish a theorem. Nevertheless it is a satisfaction to have
this formal method to fall back on whenever challenged by
one's own hesitancy or by that of others. In like manner, who
has not at times during some long complicated or indirect
logical demonstration felt the least bit uncertain ; who would
not be glad to have at his hand some formal method such as
Peano's, based upon certain rudimentary propositions and
concepts ?
I n truth it is a matter of more consequence than is some-
times thought, to have clearly in mind those processes which
are definitely to be admitted as logical. The one process which
stands out most definite in our consciousness is the syllogism.
I f a piece of reasoning can be put in the form of major premise,
minor premise, conclusion, we are tolerably sure of its truth.
But numerous proofs cannot be so constructed and it is one of
the most frequent errors committed by the intuitive logicians
to say that reasoning consists in a sequence of syllogisms.
Perhaps the greatest advance made by Boole was the clear
recognition of the necessity of asyllogistic reasoning.
The question then becomes of fundamental importance :
What is at the bottom of our logic? When the matter is

* Duhamel, Des méthodes dans les sciences, Paris, 1875, vol. 1, p. 17,
80 T H E FOUNDATIONS OF MATHEMATICS. [Nov.,

looked into, it appears that we constantly use propositions,

passing from certain propositions as hypothesis to certain others
as conclusion. The laws of implication which govern the rela-
tion between hypothesis and conclusion constitute the logical
theory or calculus of propositions. Casting about for other
principles we come upon classes or sets of objects represented
in ordinary speech by common nouns. The development of
the interrelations of classes produces the logical calculus of
classes. This calculus has a remarkable analogy with the
calculus of propositions, but the relation is not quite dual. In
the third place we perceive that relations are of the utmost im-
portance. Every transformation, every function is a relation.
In common language the verb does but express a relation be-
tween the subject and object. Thus there appears the necessity
for a calculus of relations.* The complete logical calculus, as
now used, is a combination of these three types. The whole
number of laws of thought or logical premises which seem to
be required for establishing the calculus in all the generality
necessary for mathematics is small. In addition to these pre-
mises there are a certain number of elementary ideas or terms
such as implication, and the notions of proposition, class and
relation, which must be assumed as known. I t is the dis-
cussion of these questions which are of a philosophical rather
than mathematical nature, that fills the first Part of RusselPs
Principles.
We may grant, then, that logic is necessary to mathematics.
I t is affirmed to be sufficient This in reality is the remarkable
content of the definition given by the author. So immune are
we from logical error that the necessity of logic might never
force us to a critical examination of its principles ; but the affir-
mation of its sufficiency fully justifies and even renders im-
perative such an examination. RusselPs entire volume is
devoted to establishing this sufficiency. And although the
subject is very new and many difficulties philosophical and
mathematical are still outstanding, there can be little doubt
that to an unexpectedly large extent the author is successful in
his attempt and that in these Principles he has given a per-
* Peano and his immediate followers overlook the importance of this sub-
ject— so busy are they with other important questions. It is one of the
lasting services of Russell, following very closely on the work done twenty
years earlier by C. S. Peirce, to have recognized the necessity of this addition
to Peano's system and to have supplied the deficiency. See his articles in
the Revue de Mathématiques, vol. 7, nos. 2 and following (1901-1902).
1904.] THE FOUNDATIONS OF MATHEMATICS. 81

manent set to the future philosophy of the questions which he

handles.
4. Some Notions. — Owing to the wide-spread diversity of
usage in the meaning of such fundamental notions as postulates,
axioms, undefined symbols, definitions, consistency, independence
(of postulates), irreducibility (of undefined symbols), complete-
ness (of systems of postulates and undefined symbols), we think
it best to enter upon some slight exposition * of these matters
instead of taking up the critical discussion of some of the more
abstruse problems which are treated by the author and which
could scarcely be appreciated before such exposition.
Axiom is a word which has so long been used in so many
vague ways that its use in pure mathematics had probably
best be abandoned. The familiar definition : An axiom is a
self-evident truth, means, if it means anything, that the propo-
sition which we call an axiom has been approved of by us in
the light of our experience and intuition. In this sense pure
mathematics has no axioms : for mathematics is a formal subject
over which formal and not material implication reigns.f The
proper word to use for those statements which we posit would
seem to be postulate. What self-evident truths can there be
concerning objects which are not dependent on any definite
interpretation but are merely marks to be operated upon in
accordance with the rules of formal logic? Postulates, how-
ever, may be laid down at will so long as they are not contra-
dictory. I t is the postulates which give the objects their
intellectual though not physical existence. Indeed before we
can apply to the physical world any of the systems of logical
geometry, for instance, we have the one great axiom : This
system fits nature sufficiently for our purposes. To postulate
such a statement would avail us naught. We must carefully
consider the totality of our experience and decide whether the
statement seems to represent a truth.
Definition is a term which has long been used by philoso-
phers to stand for a process of analysis and exemplification

*See also E. V. Huutington on " S e t s of independent postulates for the

algebra of logic," Transactions, vol. 5, p. 288 (July, 1904).
f i n regard to logic on which mathematics rests, we should incline to use
the word axiom (if indeed we do not prefer to hold to premise) and not pos-
tulate. For here we are dealing with the actual (mental) world and not
with a system of marks. The basis of rationality must go deeper than a
mere set of marks and postulates. I t is foundation of everything and must
be more real than anything else.
82 T H E FOUNDATIONS OF MATHEMATICS. [Nov.,

which brings before the mind a real consciousness of the object

defined. This sort of' definition has to be used in dictionaries.
In mathematics, however, no such vague process is permissible.
Mathematical definition is simply the attributing of a name to
some object whose existence has been established or momen-
tarily postulated. I t is the process of replacing a set of state-
ments by a single name and is resorted to solely for conven-
ience. In any science whose development has been perfected,
definitions may be entirely done away with by those who are
willing to sacrifice brevity. There can be little doubt that a
large number of definitions might better be thus put out of the
way.*
Although all definitions are thus merely nominal, there are
three distinct aspects f of definition which are worth consider-
ing in detail in connection with the theory of integers. These
may be characterized as (1) the particular definition, (2) the
definition by postulates, (3) the definition by abstraction. They
may be illustrated as follows : Suppose (1) that it is possible to
find a logical class K of which the elements are, let us say,
classes or propositions. Suppose further that by means of
logical processes alone we may define operations on the elements
like addition and multiplication of integers. Grant that there
exists in the class iTan element analogous to zero (in case iTis
a class of classes this would be the null-class ; in case K were
a class of propositions it would be the absurd). In fact sup-
pose that we could set up a class K and a set of operations in
K which have the properties of integers as we use them.
We then might say from a formal point of view that the class
K was the class of integers, that the elements of K were the
integers themselves, and that the operations we had set up were
the ordinary operations of arithmetic. This would be a satis-
factory though very particular definition of the integers and
would have the advantage that unless there were a contradiction
inherent in our logic there could be no contradiction in our
system of integers. Or (2) we may assume a certain set of
elementary terms, known as undefined symbols, such as num-

* Peano : Bibliothèque du Congrès International de Philosophie à Paris,

1900, vol. 3 (1901).
t Compare Burali-Forti ; Bibliothèque etc. l ' Sur les différentes définitions
du nombre réel." W e say aspect on account of a change in view which has
been established since 1900. Also u Le classe finite " Atti deUa Accadtmia
rcale di Scienze di Torino, vol. 32, p. 34 (1896) ; and more recently "Rulla
teoria generale delle grandezze e dei mimeri," ibid., vol. 39 (Jan., 1904).
1904.] THE FOUNDATIONS OF MATHEMATICS. 83

ber, zero, and successor. These we could connect, as Peano

and Padoa have done, by a system of postulates, and thus we
should have a definition of number through postulates. In
order to prove the non-contradictoriness of our postulates and
indefinables, that is, the existence of our elements, we should
have to set up some system which afforded one interpretation
of the indefinables and of the postulates. As this must be done
by going back to the laws of thought we finally get very near to
where we started in the other sort of definition. The definition
remains, however, slightly more general : for the integers thus
defined are not merely one set of elements but any set which
satisfies the postulates. Or (3) we may use the principle of
abstraction on which Russell places a great deal of emphasis.
We may say that two classes of objects, no matter what objects
they be, have the same number when there exists a one-to-one
relation between their elements.* Thus number becomes the
common property of all similar classes, and is their only com-
mon property. The class of numbers becomes the class of all
similar classes. Owing to RusselPs development of the theory
of relations this definition becomes also merely nominal and as
it seems to be the most fundamental and philosophic it may be
accepted as the best thus far given.
Although the use of postulates other than the premises of
logic, and the use of undefined symbols other than those of logic
seem needless and to be avoided in pure mathematics, the usage
is so common that we may go on to say a few words concerning
consistency, independence, irreducibility, and completeness —
especially as these ideas are somewhat usable in the founda-
tions of logic. To show the consistency of the system of pos-
tulates and undefined symbols it is evidently futile to attempt
to develop the consequences of the postulates until no contra-
diction is reached (this method of stating the thing is sufficient
to show wherein lies the futility) : for the most that can be ac-
complished in this way is to see that up to a certain point no
contradiction has been reached. The method of proof consists
merely in finding some system of entities known to exist and
affording a possible interpretation of the undefined symbols
and postulates. To make the proof really fundamental for the
system of integers it seems necessary to go quite out of the
* Russell shows, Principles, p. 113, that this idea is not dependent on the
general concept number, nor even on the concept unity. Two classes
which can be placed in one-to-one correspondence are called similar.
84 T H E FOUNDATIONS OF MATHEMATICS. [Nov.,

field of mathematics into the domain of logic* The method

of showing the independence is merely to set up for each postu-
late one existent system of elements in which there are possi-
ble interpretations of our undefined symbols and which satisfies
all the other postulates but not the particular one in question.
This shows the independence of that one. If one of the un-
defined symbols used in the statement of the postulates can
be given a nominal definition in terms of the others the sys-
tem of indefinables is redundant. I t was Padoa f who first
made effective use of this idea. To show the irreducibility
of the indefinables relative to the system of postulates it is
necessary to set up a system of elements which satisfies all the
postulates, which affords an interpretation of the undefined
symbols, and which continues to satisfy these conditions when
one of the undefined symbols is suitably altered : this must be
done for each. The problem is quite similar to that of the in-
dependence of the postulates and is not difficult to solve in case
the number of undefined symbols is small. All this difficulty
is avoided in dealing with the different branches of mathematics
when Russell's point of v i e w — n o new indefinables, no new
postulates — is taken.
Huntington $ seems to have been the first to bring to effec-
tive use the idea of completeness. The problem is to show that
if there are two sets M and Mf of objects each § of w7hich satis-
fies the postulates and affords interpretations of the indefinables,
then the two sets of objects may be brought into one-to-one
correspondence in such a way as to preserve the interpretation
of the symbols. With the statement of this last idea we have
arrived at the limit of present ideas concerning the interrela-
tions of the notions at the base of mathematics as defined by
postulates.
•*See references given in footnote under § 2, p. 77. The consistency is far
more important than the independence, irreducihility, or completeness : for
these are merely a matter of elegance, whereas that determines whether or
not all our reasoning upon the system in question is void.
t Bibliothèque e t c " Essai d'une théorie algébrique des nombres entiers,
précédé d'une introduction logique à une théorie deductive quelconque."
This remarkable essay should be read by every one. We may note that
Padoa uses * transformateurs ' but introduces no theory of relations. In this
respect Russell has introduced improvements.
% Transactions, vol. 3, pp. 264-282 (1902). See also Veblen, Transactions,
vol. 5, p. 346 (1904).
§ Serious mistakes, resulting in definitions of no essential content, have
been made by forgetting that the relations whicji connect the elements must
be in correspondence, in addition to any correspondence between the ele-
ments. See also footnote under § 6, p. 87.
1904.] T H E FOUNDATIONS OF MATHEMATICS. 85

5. Numbers. — The analysis of number, cardinal or ordinal,

finite or infinite, integral, rational, or real, with carefully drawn
distinctions between the many allied ideas such as counting,
quantity, magnitude, and distance, forms the content of Parts
I I - V . , pages 111-370, of the Principles. To do anything like
iustice to this masterpiece of analysis in a field so strewn with
difficulties would be impossible within the space at our com-
mand. The summaries given by the author at the close of each
Part afford a clear review of the ideas which have been dis-
cussed and the points which have been won. Leaving out of
account the advances which are made toward the precision of
the terms which lie at the bottom of logic we can at best merely
indicate some of the results which are of greatest interest to
mathematicians.
I t is shown that cardinal and ordinal integral numbers are
inherently different, that finite cardinals and ordinals may be
obtained in terms the one of the other but that this principle
cannot be applied to the infinite. With the guidance of the
principle of abstraction cardinal integer has been defined as a
class of similar classes. This definition has the immediate
advantage of giving finite and infinite cardinals at the same
time. The finite may then be distinguished from the infinite
by the fact that in the former the whole cannot be similar to its
part, whereas in the latter it can. Another point which Russell
establishes with the aid of Whitehead * is that by the use of
logical addition the numerical addition of a finite or infinite
number of finite or infinite cardinals may be and indeed (if we
invoke the principle of abstraction) should be defined in such a
manner that the order in which the numbers are added plays
no part. This is a great victory for common sense and must
appeal to everyone as a vindication of the school-child in his
inherent notion that he has the same number of marbles whether
he has five in one pocket and three in another or three in two
pockets and two in a third, no matter which of his pockets
these be. The principle of commutation and association of the
terms in addition is entirely done away with, except in so far
as mechanical difficulties prevent us from writing simultaneously
a number of terms and the signs of addition connecting them.
We may point to the fact that the work applies equally well to
finite and infinite sums as an indication of its extreme generality

American Journal of Mathematics, vol. 24 (1902).

86 THE FOUNDATIONS OF MATHEMATICS. [Nov,,

and as evidence that at last we have a principle of addition dis-

tinctly above the plane of counting on one's fingers. In like
manner the definition of multiplication is such as to be free
from the laws of commutation and association of the factors and
to apply equally to finite or infinite products of finite or infinite
cardinals. Again a vindication of the school-child who rightly
cannot see why it should make any difference whether he puts
down four rows of three marks or three rows of four.
The discussion of the meaning of quantity and magnitude in
Part I I I . and its connection with number we will not pause to
consider, but we pass directly to the theory of order as de-
veloped in Part I V . The treatment of this subject is greatly
simplified by the theory of relations. Order is shown to be an
asymmetric transitive relation, an essential property of serial
relations. I t is clearly pointed out and it is important to notice
that when a set of objects is given the relation is not necessarily
included ; whereas when the relation is given the field in which
it operates must also be given. If recourse is had to the prin-
ciple of abstraction the ordinal integer appears as " the common
property of classes of serial relations which generate ordinally
similar series." As the cardinals are classes of similar classes,
so the ordinals are classes of like relations. The principle of
induction is intimately associated with the system of ordinals
rather than with the system of cardinals although for finite
numbers the distinction is not so great as for infinites. We
may say that the finite ordinal is that which can be reached by
induction from 1. I t appears that those who generate their
system of numbers by a relation of succession or by counting
— that is, by successive acts of attention — must in reality be
coming at something which resembles ordinals much more
nearly than cardinals. The difference between the theory of
infinite cardinals and infinite ordinals brings to light the im-
portant fact that in mathematics we have two kinds of infinite :
the cardinal, which has the property of being similar to a part
of itself, and the ordinal, which cannot be reached by induction
from 1.* The discussion naturally brings up the old question
of extensive and intensive definition. The definition of an
*This would seem to render invalid the contention of Poincaré in his La
science et l'hypothèse to the effect that the principle of induction is the
essence of the infinite. We have seen that it is the essence of the finite. The
difficulty seems to be that Poincaré has in mind the definition of the infinite
as a growing variable. If this be so, the apparent contradiction resolves
itself into a mere difference of definitions.
1904.] T H E FOUNDATIONS OF MATHEMATICS. 87

object is said to be extensive when the object is given by the

enumeration of its parts ; it is said to be intensive when the
object is characterized by its properties. In the treatment of
these questions and of transfinite cardinals and ordinals there is
much which is instructive for the mathematician and the phi-
losopher. The author points out with his customary frank
desire to state no more than the truth that there still remain
difficulties to solve. Thanks to his lucid and modest presenta-
tion there is no reason why he should not find adherents who
will take up the work and attempt the solution in a spirit of
hearty cooperation.
There is a school of creationists who, when they find that
certain infinite processes lead to no rational limit nor yet to a
number which becomes infinite, postulate the existence of a limit
and thus obtain the irrational numbers. The author does not
consider an ipse dixit like this to be a sufficiently good theorem
of existence. He therefore considers infinite sets of rationals
and by means of them he forms a set of things which he calls
real numbers. A real number is neither a rational nor an
irrational ; it is a certain infinite set of rationals. The real
numbers thus defined are shown to satisfy the notion of a con-
tinuum. According to the method followed, the continuum
appears as an idea wholly ordinal in nature. With the aids
thus prepared the author is able to give a very satisfactory
account of the philosophy of the infinite and of the continuous.
His treatment of the paradoxes of Zeno shows that the argu-
ments of the ancient philosopher are by no means so far from
right as might be imagined and that the contradictions are
more apparent than real.
6. Geometry and Mechanics. — A short study of the proper-
ties of multiple series leads to a point from which it may be
seen that : Geometry is the study of series of two or more di-
mensions.* In this manner the necessity of new postulates and
new indefinables is avoided. The procedure is evidently reason-
able. Mathematical geometry has long since been divested of
all spatial relations between its elements. The above definition
* As the serial relation is emphasized rather than its domain (see discus-
sion of order given above) the author avoids a definition which is null and
which makes dimensious impossible. Compare discussion of completeness
and footnote, \ 4. For a fuller discussion of this important point we may
refer to " T h e so-called foundations of geometry," by the present reviewer,
in the Archiv der Mathematik und Phyxik, vol. 6, pp. 104-122 (1903). Toward
the end of the discussion a change, which may cause some confusion, is made
to the point of view of physical geometry. The first part, however, deals
solely with purely mathematical geometry.
88 THE FOUNDATIONS OF MATHEMATICS. [Nov.,

is but the culmination of the ideas of manifolds introduced by

Grassmann and Riemann. As those who define geometry by
postulates are forced to show the existence of their elements
by having recourse to systems of numbers the question is quite
pertinent : Why not begin with a purely nominal definition
like the above and avoid the trouble of proofs of existence, of
independence, and of irreducibility ?
At this juncture it is interesting to compare the attitude
taken in the Principles with that taken in the older Founda-
tions of Geometry. I t should be remembered that the author
originally started with the study of the philosophy of dynamics
and hence necessarily of geometry. To render the examination
really searching the foundations of geometry had to be inves-
tigated. But, once started, the end was not so easily to be
reached. Probing into the mysteries of infinity and continuity
led to arithmetic in its wider sense. Trying to render precise
the meaning of important words such as element, set, operation,
conclusion, proof, etc., could but conduct to the study of logic,
and the desire to be rid so far as possible from the contamina-
tions of the personal element brought up at last at formal logic.
And then the entire field had to be traversed in the forward
direction with the necessity of constant acquisition of original
results at every step ! Surely the present work is a monument
to patience, perseverance, and thoroughness.
I n the essay on the foundations of geometry the author had
not yet reached the logical stage — scarcely the arithmetic
stage. He was content, as some still are, to analyse the ideas
in the rough, to use a large number of indefinables, to state
broad indefinite axioms instead of brief incisive postulates —
in a word, to forego all the modern technique. The result was
an extremely suggestive essay — one which still can be read
with profit and by rights ought to be read, if only for the sake
of contrast, in connection with the newer work. To-day we
have mathematical geometry, then we had physical. If one
wishes to read an excellent account of space from the physico-
metaphysical point of view he has but to turn to the Russell
of a few years since ; if he would know the extreme point of
modern mathematico-logical geometry he has merely to take up
the Russell of to-day.
In Part V I I . the analysis proceeds to mechanics. Here
space is merely a certain three- (or n-) dimensional series ; time,
a simple series. There is a relation which connects part of space
1904.] T H E FOUNDATIONS OF MATHEMATICS. 89

(the material points) with all time, that is, a, b, c = R(aQ, b09 c0, t),
where a, 6, c are the coordinates of the material points. This
relation R is so chosen as to allow for the impossibility of gen-
erating or destroying matter. The relation is also chosen so
that if the relation between matter and time is known at two
instants it is known at every instant. In this way is stated
the causality in the universe. This seems very far off from
the real world. I t must delight the hearts of philosophers who
believe in a pure idealism. I t is found that arithmetic may
be handled adequately with no help save from logic. This
does not surprise us. Then geometry is put in the same cate-
gory. Modern mathematicians have so accustomed us to look
on merely the logical side of the subject that we are not troubled.
Finally comes dynamics. Why not thermodynamics, electro-
dynamics, biodynamics, anything we please? There is no
reason why not. There is in reality no place to stop, save
when we have become tired of pure logic, if once we include
geometry. As a matter of fact all our concepts whether of
space, or matter, or electricity, or life, are but idealizations more
or less well-defined, and, if we insist on subjecting the world
to purely logical explanation, they all belong in the same class.
Upon this matter we may best quote Russell who, amid all
his refinements, keeps a clear idea of their proper place in the
system of all knowledge. He says : The laws of motion, like
the axiom of parallels in regard to space, may be viewed either
as parts of a definition of a class of possible material universes,
or as empirically verified assertions concerning the actual mate-
rial universe. But in no way can they be taken as à priori
truths necessarily applicable to any possible material wTorld.
The à priori truths involved in dynamics are only those of
logic ; as a system of deductive reasoning, dynamics requires
nothing further, while as a science of what exists, it requires
experiment and observation. Those who have admitted a sim-
ilar conclusion in geometry are not likely to question it here ;
but it is important to establish separately every instance of the
principle that knowledge as to what exists is never derivable
from general philosophical considerations, but is always and
wholly empirical. *
* I t would be interesting to discuss in how far this attitude is really in
accord or out of accord with the apparently very different view of Poincaré
(La science et l'hypothèse) that the question whether the parallel axiom is
true or not true is devoid of sense owing to the fact that it is merely a con-
venient method of correlating experience and a convenlion can have neither
truth nor falsity.
90 THE FOUNDATIONS OF MATHEMATICS. [Nov.,

7. Some Conclusions.—There is one conclusion in logic which

suggests itself almost inevitably at this point. For there are a
considerable number of systems of logic current at present.
Different authors have treated the subject differently — each
choosing the system of indefinables and laws of thought which
seemed best to him at the time. Now it is by no means true
that these various systems of logic have been proved coexten-
sive or even not mutually contradictory. If it should appear
that they cannot be brought into harmonious relation one with
another there will be some instructive, if bewildering, conclu-
sions to draw. And as we have such complex entities as infinity
and continua with which to deal it might not be regarded as
surprising if some points were found to stand out permanently,
so that logicians will permanently disagree. In fact at present
there seems to be a grave logical difficulty in our logical system
as developed by Russell. This trouble had been felt by
Frege and a solution had been proposed by him ; but it does not
seem entirely satisfactory. * In view of the outstanding diffi-
culties and the possible divergence of systems of logic held by
equally good authorities, we come to the conclusion that it is
dangerous to accept the naïve point of view of those who claim
that a certain piece of reasoning depends on the operation of
logic alone but who fail to state what those operations are and to
use all the means possible to avoid the intrusion of extraneous
ideas. They may not fall into error, but they are merely fol-
lowing in the footsteps of those who " knew " what infinity and
continuity were.
From the pedagogic point of view we may also draw some
conclusions. I t is hardly necessary to trouble the student
with the commutative and associative laws in multiplica-
tion of integers or with elaborate deductions of a number
system before he is readily able to appreciate the needlessness
of the former and the relation which the latter bears to the
theory of finite and infinite cardinals and ordinals, the ideas
of compactness and continuity, and the two kinds of infinity.

* I n a long appendix, Eussell gives a detailed exposition of the important

work of Frege, which culminated in the Grundgesetze der Arithmetik, and
he discusses this troublesome contradiction again from a different standpoint.
I t is this contradiction which Hubert had in mind in his Heidelberg address
referred to under \ 2. He, therefore, attempts to recast the principles of logic
and of arithmetic in such a manner as to render them sufficient for mathe-
matical reasoning. We certainly hope that he has succeeded in doing so to
the satisfaction of both mathematicians and philosophers.
1904.] T H E FOUNDATIONS OF MATHEMATICS. 91

A clear-cut physical conception that numbers possess order

and may be associated with the points on a line is a workable
idea which in practice is both necessary and sufficient for
ordinary rigorous analysis. An inadequate vague idea re-
garded as a useful working hypothesis seems, on the whole,
productive of more good and less harm than an inadequate defi-
nite idea regarded as final. In geometry and mechanics the
physical attitude may be taken. Axioms, things deemed
worthy of credence on the basis of experience, should take the
place of postulates. This does not prevent, in fact it en-
courages, the statement of a large number of axioms without
troubling too much as to their independence. At the same
time these statements should include the essential idea of order
and the useful idea of continuity and other ideas which are
usually passed over.* In short we should use and train intui-
tion to the utmost in connection with some logic ; for pure logic
alone is, as Poincaré states (§ 2), harmful to the earlier develop-
ment of the mind.
From the mathematical standpoint we have learned that
many of the objects which have been thought of as individual
must be regarded as classes. We cannot define euclidean
space, but wTe can define the class of all euclidean spaces.f
The principle of abstraction, here involved, seems to arise from
the necessity of taking the terms in a logical equation to repre-
sent the common attribute of all the objects which may in some
way satisfy the equation. As during the progress of the dis-
cussion, we have introduced no new indefinables, no new pos-
tulates, no processes other than those of logic, there is no possi-
bility of our arriving at contradictions except through the failure
of our logical system to be logical ; and behind this we cannot
go. I t remains merely to show the existence of the classes
with which we have dealt; otherwise our work would be null.
To quote freely from our author : The existence of zero is
derived from the fact that the null-class is a member of zero ; the

* Compare, for example, the series of articles by the reviewer on Spherical

Geometry, American Mathematical Monthly, commencing January, 1904.
t This apparently considerably lowers the importance of the idea of com-
pfetenes* discussed in §4. For it appears as if the one-to-one correspondence
between the different euclidean spaces were really of minor significance.
This is but another instance of the fact that the elements themselves are un-
important— that it is the abstraction from them which is most funda-
mental. However, we believe that the idea of completeness is a new step, a
step onward and toward a fuller description of the systems dealt with.
92 T H E FOUNDATIONS OF MATHEMATICS. [Nov.,

existence of 1, from the fact that zero is a unit-class (the null-

class being its only member). By an evident induction we get
all finite numbers. From the class of finite cardinals follows
first the existence of the smallest of the infinite cardinals, and
second, by considering them in the order of magnitude, the ex-
istence of ordinals and the smallest of the infinite ordinals.
We may go on to obtain the rationals, compact enumerable
series, continuous series. From the last we may see the ex-
istence of complex numbers, of the class of euclidean spaces, of
projective spaces, of hyperbolic spaces, and of spaces with va-
rious metrical properties. Finally we may prove the existence
of the class of dynamical worlds. Throughout this process no
entities are employed but such as are definable in terms of the
fundamental logical constants. Thus the chain of definitions
and existence-theorems is complete, and the purely logical
nature of mathematics is established throughout.
This is as far as we are conducted. But we are promised a
second volume — may it be soon forthcoming — written with
the collaboration of Whitehead. Herein will be contained
actual chains of deduction leading from the premises of logic
through arithmetic to geometry. Herein will also be found
various original developments in which the notations of Peano
and Russell have been found useful. For those who wish
sooner to get at the Peano-Russell point of view in the matter,
we append a bibliography, which while very incomplete may
still be found useful in tracing the development of the ideas :
(1) Arithmetices principia nova methodo exposita, Turin,
Bocca Frères, 1889.
(2) I principii di geometria logicamente esposti, Turin, 1889.
These two works by Peano are the starting point of the whole
movement. They were written in the days when a careful ex-
planation and translation of the symbolic method was in vogue
and form a good starting point for study. The Formulaire de
Mathématiques, edited by Peano, is rather hard to begin on.
The Bivista di Matematica, now the Revue de Mathématiques,
also edited by Peano, furnishes much easy and instructive read-
ing matter. Logica matematica by Burali-Forti in the series of
Manuali Hoepli may serve as a textbook. Omitting important
memoirs by Burali-Forti on arithmetic and by Pieri on geom-
etry which we have quoted in footnotes, we cite again
(3) Bibliothèque du congrès international de philosophie,
volume 3 (1901).
1904.] NOTES. 93

The articles by Peano, Burali-Forti, Padoa, and Pieri show

the point at which the Italian school had arrived in 1900. I t
is since that time that most of Russell's technical work has ap-
peared. For the present state of the science, we would note
a memoir by Whitehead :
(4) " O n cardinal numbers," American Journal of Mathemat-
ics, volume 24 (1902), pages 367-394; and a still more recent
paper by Burali-Forti, " Sulla teoria generale delle grandezze e
dei numeri," Atti délia B. Accademia delle Scienze di Torino,
volume 39, (January, 1904).
EDWIN BIDWELL WILSON.
YALE UNIVERSITY,
July, 1904.

NOTES.
T H E closing (October) number of volume 26 of the Ameri-
can Journal of Mathematics contains the following papers : " In-
variants of a system of linear partial differential equations, and
the theory of congruences of rays," by E. J. WILCZYNSKI ;
" On elements connected each to each by one or the other of
two reciprocal relations/' by C. DE POLIGNAC.
T H E opening (October) number of volume 6 of the Annals
of Mathematics contains the following papers : " On the sub-
groups of an abelian group," by G. A. M I L L E R ; " Note on the
continued product of the operators of any group of finite order,"
by W. B. F I T E ; " Reduction of an elliptic integral to Legen-
dre's normal form," by N. R. WILSON ; " The necessary and
sufficient condition under which two linear homogeneous differ-
ential equations have integrals in common," by A. B. P I E R C E ;
" A general method of evaluating determinants," by G. M A O
LOSKIE ; " Application of groups to a complex problem in
arrangements," by L. E. DICKSON ; " On functions defined by
an infinite series of analytic functions of a complex variable,"
by M. B. PORTER.

A T the Cambridge meeting of the British association for the

advancement of science (cf. October B U L L E T I N , page 28), Pro-
fessor A. R. FORSYTH presided over the subsection of pure
mathematics, whose programme included the following papers :
" A fragment of elementary mathematics," " Geometry of
the complex variable," by Professor F . M O R L E Y ; " Peano's