]\gt ; PHILOSOPHICAL TRANSACTIONS .
A Memoir on the Theory of Form .
* A. B. KEMPE , M.A. , F.R. Received May 18 , \mdash ; Read June 18 , 1885 .
TABLE 0F CONTENTS .
CTIONS .
PAGII .
1- 2.\mdash ; Scope of the memoir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 3-13.\mdash ; Fundamental principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2-4 Units .
4 , 5 18- 24.\mdash ; Some definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5 , 6 25-36 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6 , 7 37-49.\mdash ; Heaps\mdash ; Graphical representation of units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7 , 8 50- \mdash ; Graphical representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9-14 73-83 .
Aspects . . . . . . . . . . . . . . . . . .
14-16 84-86.\mdash ; Letters , their sorts and positions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16 87- 88.\mdash ; Representation of aspects of collections by arrays of letters . . . . . . . . . . . . . . . . .
16 , 17 89- 99.\mdash ; Elementary properties of aspects . . . . . . . . . . . . . . . . . . . . . . . . .
17 , 18 100-111.\mdash ; Tabular representation of systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18-20 112-129.\mdash ; Correspondences of undistinguished collections . . . . . . . . . . . . . . . . . . . . . . . . . . .
20-23 \mdash ; Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
, 24 \mdash ; Aspects unique with respect to collections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
24 , 26 Associates . . . . . . . . . . . .
25 , 26 Unified aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
26 , 27 162-169.\mdash ; Correspondences of systems of like forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27-29 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
29 , 30 \mdash ; Independent and related systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30 , 31 \mdash ; Three modes of compounding systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31 , .\mdash ; General method of graphically representing a system . . . . . . . . . . . . . . . . . . . . . . . .
32 196-210.\mdash ; Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
32-34 \mdash ; Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
84 , 35 * In aocordance with criticisms and suggestions of ths referees , an alteration has been made in the title of the memoir , and certain sections towards the end have been omitted .
There is consequently a slight want of complete correspondence between the memoir and the abstract given in the vol. 38 , p. 393 \mdash ; A. B. K. PAGE .
CTIONS .
.\mdash ; Groups containing from one to twelve units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
37-43 \mdash ; Some general forms of groups . . . . . . . . . . . . . . . .
, . . . . . . . . . . . . . . . . . . . . . . .
43-47 256-269 .
family of groups 270-274 .
-adic groups . . . . . . . . . . . . . . . . . . . . . . . . . .
275-279 .
bstitutions . . . . . . . . . . . . . . . . . . . . . .
48-54 \mdash ; Algebras . . . .
54-57 \mdash ; Quadrates . . . . . . . . . . . . . . . . . . . . . . . .
\mdash ; Isolated Satisfied collections . . . . . .
57 , 58 332-343 .
\mdash ; Some isolated-triad systems.\mdash ; Family No. 1 . . . . . .
58-61 \mdash ; Some isolated-triad systems.\mdash ; Family No. 2 . . . . . . . . . . . .
\mdash ; Geometry in genel.al \mdash ; System of coplanar points and straight lines .
62 , 63 358-359.\mdash ; Coplanar points , lines , and conics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
63-70 Logic Scope of the 1 .
My object in this memoir is to separate the necessary matter of exact or mathematical thought from the accidental clothing\mdash ; geometrical , algebraical , logical , \amp ; c.\mdash ; in which it is usually presented for consideration ; and to indicate wherein consists the infinite variety which that necessary matter exhibits .
2 .
The memoir is confined to the exposition of fundamental principles , to their elementary , to their application to such a variety of cases as will vindicate their value , and to a description of some simple and uniform modes of putting the necessary matter in evidence .
I have been unable to ascertain that the principles here set forth have been previously formulated .
Fundamental Principles .
3 .
Whatever may be the true nature of things and of the conceptions which we have of them ( into which points we are not here concerned to inquire ) , in the operations of reasoning they are dealt with as a number of separate entities or units .
4 .
These units come under consideration in a variety of garbs\mdash ; as material objects , intervals or periods of time , processes of thought , points , lines , statements , relationships , arrangements , algebraical expressions , operators , operations , \amp ; c. , \amp ; c. , occupy various positions , and are otherwise variously circumstanced .
Thus , while some unlts are incapable of being distinguished from each other , others are by these peculiarities ldered distinguishable .
For example , the angular points of a square are distinguishable from the sides , but are not distinguishable from each other .
In some instances where distinctions exist they are ignored as not material .
Both cases are included in the ral statement that some units are distinguished from each other and some are not .
5 .
In like manner some of umts are distinguished from each other , while others are not .
Pairs may in some cases be distinguished even though the un.its MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
composing them are not .
Thus the angular points of a square are undistinguishable from each other , and a pair of such points lying at the extremities of a side are undistinguishable from the three other like pairs , but are distinguished from the two pairs formed by taking the angular points at the extremities of the diagonals , which pairs again are undistinguishable from each other .
Further , a pair may sometimes be distinguished from the pair though the units and are undistinguished .
Thus , in fig. 1 the three black spots , are all undistinguished from each other , each has an Fig. 1 .
arrow proceeding to it and an arrow proceeding from it , but the pair is distinguished from the pair ba , for an arrow proceeds from to , but none from to 6 .
It will be convenient to speak of and ba as different aspects of the collection of two units .
Here the terms " " aspect\ldquo ; and " " collection\ldquo ; are each to be understood as referrin t , o two separate units , and not to those units regarded in the aggregate as a single unit .
7 .
Again , there are also distinguished and undistinguished , .
m-ads , . . . . .
; every m-ad being , of course , distinguished from every n-ad .
Just as we may have ab from ba , though is undistinguished from so we may have n-ad pqrst .
uv distinguished from qusvt . . .
, though the units \amp ; c. , are all Qhed from each other , and further , though their pairs are also undistinguished , as likewise their triads , \amp ; c. Here pqrst . .
and qusvt . .
will be termed , as in the case of pairs , different aspects of the collection , . .
: the term " " collection\ldquo ; being understood to refer to a number of separate mits without reference to the various ' aspects\ldquo ; of collection .
Different aspects of the same collection of units will be regarded as different n-ads .
8 .
The terms " " pair " " triad\ldquo ; " " n-ad , \ldquo ; " " collection " " aspect\ldquo ; will always be understood to refer to two , three , , \amp ; c. , units , and never to aggregations of units considered as a single unit .
Pairs , triads , n-ads , collections , aspects may , of course , be regarded as units , but when they are so regarded the fact will be distinctly pointed out .
9 .
Every collection of units has a definite form due ( 1 ) to the number of its Fig. 2 .
component units , and ( 2 ) to the way in which the distinguished and guished units , pairs , triads , \amp ; c. , are distributed through the collection .
Two collections of the same number of units but having differe1lt distributions will be of different forms .
Thus the two tetrads , and , of fig. 2 , contain the number ( four ) of unit spots , but they are of different forms ; for are all undistinguished from each other ; while , though undistinguished from eaoh other , are all distinguished from .
The distribution of the distinguished and undistinguished pairs and triads is also obviously different in the two cases .
The word " " form\ldquo ; will in this memoir be always employed in the sense here indicated .
10 .
Two collections of units which are undistinguished are of the same form , but two collections which are of the same form are not necessarily undistinguished ; there may be the same distribution of distinguished and undistinguished units , pairs , \amp ; c. , in each , while the units , pairs , \amp ; c. , of one are all distinguished from the units .
palrs , \amp ; c. , of the other .
11 .
Each of the forms which a system of any number of units can assume , owing to varieties of distribution , is one of a definite number of possible forms , and the peculiarities and properties of the collection depend , as far as the processes of reasoning are concerned , upon the particular form it assumes , and are independent of the dress , geometrical , algebraical , logical , \amp ; c. , in which it is presented ; so that two systems which are of the same form have precisely the same properties , although the garbs in which they are severally clothed may , by their dissimilarity , lead us to place the systems under very different categories , and even to regard them as belonging to " " different branches of science 1.2 .
It may seem in some cases that other considerations involved besides " " form but it will be found on investigation that the introduction of such considerations involves also the introduction of fresh units , and then we have merely to consider the form of the enlarged collection .
13 .
In order to put form in evidence some " " accidental\ldquo ; clothing is of course necessary ; if , however , we employ more than one species of clothing , each species being uniform and suited to forms of every kind , the likelihood of its accidental nature being overlooked will be reduced to a minimum .
Units .
14 .
The units } we have to consider exhibit endless variety ; thus we may have a material object dealt with as one unit , a quality it possesses as another , statement about it as a third , and a position it occupies in space as a fourth .
The task of specifying the units which are considered in an investigation may in some cases be one of considerable difficulty , and mistakes are likely to occur unless the operation is conducted with great care .
15 .
We have frequently to deal with things , \amp ; c. , pairs , \amp ; c. , of those things , the differences between which depend on the existence or non-existence of certain circumstances , or upon taking into account or ignoring certain circumstances .
Thus we may apparently have collections of units \amp ; c. , which are at one time , when .
A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
one state of things exists , of one form , and another time of another form .
This is , however , not so ; the form of a collection is absolutely invariable .
The apparent alteration of form arises from overlooking the fact that the units dealt with are not \amp ; c. , but ( at time ) , ( at time ) \amp ; c. , ( at time at time \amp ; c. The unit ( at time t ) may be quite different from at time Where there is only one alternative , the collection ( at time ) , ( at time ) , ( at time ) \amp ; c. , is of the same form as the collection \amp ; c. , and it matters not which is the collection dealt with .
16 .
Aggregations of those things which , in a symbolical representation of the units considered in any case , are already represented as units , must not be supposed to be sufficiently represented without additional symbols , must each be represented by a distinct symbolical unit .
An aggregation of things is , as far a@ the processes of reasoning are concerned , a mere unit , and must be so dealt with .
17 .
While it is important that a unit should be represented and dealt with as a unit , it is equally important that we should not be misled by our modes of thought and consequent use of language into regarding a number of distinct units as one only .
A collection of units must in a symbolical representation be represented as a single unit where it is so regarded ; but where the word ' ' collection " " is ussd , as it is here ( sec. 8 ) , merely to " " denote\ldquo ; or mark off a number of things each of which is considered as a distinct unit , we must be careful to represent each of those things by a distinct symbol .
Some Definitions .
18 .
Any collection of units which consists entirely of units selected from another collection will be termed a component of the latter .
Any aspect of a component of a collection also be spoken of as a component of the collection .
19 .
Two collections of units will be said to be detached if they have no component in common .
20 .
An n-ad which has one or more units in common with each of a number of collections will be said to be an n-ad mecting those collections ; , the pairs which a single unit makes with the various units of a collection will be termed the pairs connecting the unit and the collection .
An n-ad connecting detached collections has one unit and one only in common with each .
If , . .
be collections of which , . .
are component units respectively ; when an n-ad A B . .
is spoken of , it must be understood that an n-ad such as .
is meant , and not one such as a . .
, which will be spoken of as connecting BD 21 .
If the component units of collection are all undistinguished from each other the collection will be said to be single .
22 .
Units which are undistinguished from the same unit are undistinguished from each other ; thus if a collection of units is not single it consists of two or more detached single collections , and will be termed a double , treble , ection , accordin to the number of component single collections which it contains .
23 .
Just as collections of units break up into single collections of undistinguished units , so collections of pairs break up into single collections of undistinguished pairs , and , generally , collections of -ads into single collections of undistinguished n-ads .
24 .
ferences between the valious units , pairs , \amp ; c. , of a collection may be termed the internal differences of that collection .
Differe ces between the various units , pairs , \amp ; c. , of a collection A and the units , pairs , \amp ; c. , of a detached collection may be said to be external to A. Systems .
25 .
If every component unit of a collection is distinguished from every unit which is detached from the collection , the collection will be termed a system .
26 .
Every collection of units is a component of some stem .
27 .
The whole collection of units which come under consideration in any inquiry is a syste1n ; for the units are distinguished from all others by being the only ones considered .
28 .
The -ads of one single system of units are distinguished from all those of any other single system of units , and from all the -ads connecting any two systems of units .
The connecting -ads of any systems of units are distinguished from those of any other systems of units , and themselves break up into single systems of n-ads .
29 .
The units of a single system of units must be dealt with as a whole ; for , as they are undistinguished from each other , no definition can be given of , or remark made about , one which is not equally applicable to each of the others ' .
Each can only be spoken of as " " one of\ldquo ; the units of the system .
A similar observation applies in the case of single systems of pairs , triads , \amp ; c. 30 .
Many systems of units are defined by stating that certain of their components constitute a system .
It must be borne in mind that such a statement does not mean that the components are undistinguished from each other , but merely that they are distinguished from all others ; the system of components may be a multiple one .
It is by no means unnecessary to emphasize this , as we are prone to assume that the mode of classifying things is that of putting like things into the same class , rather than that of putting unlike things into different classes .
31 .
The distribution of the various distinguished and undistinguished components of a system is regulated by definite laws ; so that a knowledge of the mode of distribution of some only of the distinguished and undistinguished components may determine the form of the system .
There are in general several ways in which the form of a given system may be thus determined , and accordingly different definitions of the same system may be adopted .
32 .
For the statement of some of the properties of a system it may be necessary to have the form fully defined ; in the case of others this may not be necessary , it being sufficient to state that components are distinguished , without asserting anything as to others , i.e. , to state that certain components constitute a system , MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
without asserting whether the system be single or multiple .
The result of this may of course be that we deal with different units at different of the investigation of ( secs .
126 ff ) .
33 .
Systems are frequently more readily dealt with when regarded as components of more extensive systems .
Much ingenuity has been expended o11 the discovery of systems , the addition of which to the particular system under consideration may assist in its investigation .
It is to this that the existence of such units as substitutions , quarternions , quadrates , \amp ; c. ( as to some of which I shall have to speak presently ) , is due .
We shall see that we can always , by the addition of a proper system to any given system , completely define the latter by merely in ( dicating the mode of distribution of certain pairs ( secs .
81 , 82 ) .
34 .
Notwithstanding the great assistance derived from the use of added systems , much reluctance is exhibited in them unless they can be shown to have their representatives in nature , i. e. , unless " " accidental\ldquo ; clothing can be found fit them .
The objections raised to symbolical methods which cannot be " " interpreted\ldquo ; are strong evidence of the fact that the accidental nature of much that comes under our consideration is not really appreciated . .
It is by no means unnecessary to state that the form of a system is independent of the particular method of definmg it which we adopt ; and that because it is easier to define a system by adding fresh units , it does not owe its form to the existence of those units .
36 .
The system which is the actual subject of investigation in an inquiry may be termed the base system ; those systems which are added for the purpose of assisting in the investigation being termed auxiliary systems .
\mdash ; Graphical of Units .
87 .
There are two forms of systems of units the consideration of which properly precedes that of all others .
The one is that of a system which consists of units , each of which is distinguished each of the others , so that every component s-ad is nguished from every other component s-ad for all values of from 1 to system of this form .
shall term a discrete heap .
38 .
The other is that of a system of units which is such that every component s-ad is undistinguished from every other component s-ad for all values of from 1 to .
A system of this form I shall term a single heap .
39 .
A discrete heap may be graphically represented by a number of small separated \copy ; \copy ; \copy ; \copy ; Fig. 3 .
circles each containing a different letter ( fig. 3 ) .
The letters render the graphical units distinguishable from each other .
8 MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAh FOBM .
40 .
A single heap may be graphically represented by a number of small separated circles each containing the same letter .
( Fig. 4 .
) \copy ; Fig. 4 .
41 .
Differences arising the positions of the graphical units on th6 paper on which they are drawn are always to be disregarded .
Where positions are considered , they must be represented by separate graphical units .
42 .
Where there is occasion to employ letters outside and adjacent to the units for purposes of reference , differences arising from the use of such letters are also to be ignored .
( Sec. 173 .
) 43 .
In every system of units other then a discrete or single heap , some -ads are distinguished from each other and some are not , for all or some values of from 1 to Thus every form which a system of units can assume may be regarded as either that of a discrete or single heap or as intermediate between the two .
44 .
There are systems of units of other forms than those of discrete and single heaps which can be represented by means of graphical units alone .
These consist of one or more detached independent single heaps .
The term " " independent\ldquo ; win be fully explained presently ( secs .
117 , 118 ) ; for the present it will be sufficient to state that two independent systems are such as can be graphically represented on separate sheets of paper , without the employment of symbols on either sheet relating to those on the other .
45 .
A system of independent detached heaps will be termed an -tuple heap .
Fig. 5 represents a treble heap of seven units .
Fig. 5 .
46 .
A discrete heap of units is an -tuple heap of units .
47 .
The number of different forms which a heap of units can assume is the same as that of the partitions of 48 .
It will in some cases be convenient , instead of employing circular graphical units distinguished by internal letters , to use coloured graphical , or black spots of different sizes , or graphical units of various shapes .
49 .
Where graphical units are employed alone , like units will represent undistinguished units , but in the case of forms other than heaps it will be necessary to employ means to distinguish pairs , \amp ; c. , and these may cause like graphical units to be distinguished from each other .
No general inference must therefore be drawn from the use of like and unlike graphical units other than that unlike units represent distinguished units .
MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
Pairs.\mdash ; Graphical Representation .
50 .
There are three different forms of pairs ; viz. , if be two units , we can have : ( 1 ) distinguished from , and therefore also distinguished from ba ; ( 2 ) undistinguished from , but distinguished from ; ( 3 ) undistinguished from , and undistinguished from 51 .
The two units in case ( 1 ) belong to different single systems , and thus the pair is a pair connecting two single systems .
62 .
In cases ( 2 ) and ( 3 ) both units belong to the same single system , and the pairs are components of that system .
The units in case ( 2 ) may be said to form an unsymmetrical pair , those in case 3 ) being said to form a symmetrical pair .
It will be convenient in speaking of undistinguished unsymmetrical pairs ab , , to say that and are of the same polarity , and that and are of opposite polarities .
Of course in this case and ba are of opposite polarities .
In the case of a symmetrical pair , of and fe will be of the same polarity .
53 .
The expression " " uIlsymmetrical\ldquo ; may also be applied to pairs falling under case ( 1 ) .
We may also in the case of such pairs say that and if undistinguished are of the same polarity , and of opposite polarities .
54 .
If in a diagram consisting of a number of graphical units , some pairs of those units are joined by plain lines or links , as in fig. 6 , while others are not so joined , the Fig. 6 .
pairs will be divided into two systems .
Pairs which are thus joined by links will not be necessarily undistinguishable , nor will pairs which are not joined by links be necessarily undistinguishable , but we merely have pairs which are joined by links distinguished from pairs which are not so joined .
Thus the two systems into which the pairs are divided may each be either a single or multiple system .
55 .
A large number of systems may be completely defined by diagrams consisting of graphical units and links only .
We have already seen that an extensive class Fig. 7 .
of systems may be represented by diagrams consisting of different sorts of graphical units only ; there are also systems which may be fully defined when we indicate by links a dichotomy of their component pairs , without defining further what units are distinguished and what are not .
Thus the system of fig. 7 must be a discrete heap , each graphical unit admitting of separate definition .
By the use of links in conjunction with graphical umts of different sorts the number of representable forms is greatly increased .
I shall give in the following four sections some examples of forms which may be represented by the use of graphical units and links .
56 .
Figs. 8 and 9 illustrate the fact that a system may be defined in various ways ; Fig. 8 .
Fig. 9 .
they represent systems of the same form .
Pairs which are not joined by links in fig. 8 are joined in fig. 9 , and vice versed . .
I give the systems of figs. 10 and 11 as examples of two varieties of single systems of the same number of units .
In fig. 10 there are four sorts of pairs of which , are types .
In fig. 11 there are only three sorts , of which , and , are types .
It may be noticed that in each case the linked pairs compose a single system ; while the unlinked pairs in the former case compose a treble system , in the latter a double one .
58 .
Fig. 12 is an example of a triple system containing unsymmetrical pairs of both Fig. 12 .
sorts .
We have an unsymmetrical pair composed of two undistinguished units , and an unsymmetrical pair composed of two distinguished units .
59 .
The single system shown in fig. 13 is one of considerable interest ; it is that dealt with in the case of the theorem that if two coplanar triangles are coaxial they MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
are also copolar ( Theorem ( 1 ) , sec. ) .
The graphical units may be taken to represent either the ten straight lines of the theorem , or the ten points of intersection ; Fig. 13 .
the form is the same in either case .
Taking the former case , the pairs of graphical units which are joined by hnks correspond to pairs of lines whose points of intersection points other than the ten qidered in the theorem .
It should be noticed that the two systems into which the pairs are divided ( linked and unhnked ) , are each single .
60 .
We shall see ( secs .
81 , 82 ) , that by the addition of units to any system , i. e. , by regarding as a component of a more extensive system , we can represent the form of by the use of graphical units and links onIy .
It will , however , be convenient in many cases to employ some further devices which ] enable us , without employin additional graphical units , to represent forms which could not be exhibited by the use of plain links only .
61 .
Thus we may have lines of various sorts joining two graphical units , viz. , dotted , wavy ( fig. 14 ) , red , blue , \amp ; c. , in addition to links , which will always be underFig .
14 .
stood to be plain lines .
Pairs joined by unlike lines will be , as in the case of links , distinguished from each other .
62 .
Where a pair is unsymmetrical an arrow-head or barb may be added to the link or other line joining the two graphical units which represent the pair ( fig. 15 ) .
The arrow-head has the effect of making distinguished from ba .
In the fig. the arrow : Fig. 15 .
heads make distinguished from fe , but not from of .
The relative directions of the arrow-heads in the case of pairs joined by unlike lines are immaterial .
It is MR. A. B. KEMPE ON THE THEORY OF MATHEMATfCAL FOBM .
unnecessarv to add an arrow-head in the case of a pair of distinguished units ; the difference between the graphical units indicates the absence of symmetry .
63 .
We may in some cases of symmetrical pairs instead of employing a single line , use two barbed lines of opposite polarities , as in fig. 16 .
Fig. 16 .
64 .
In place of using a number of different sorts of lines , we may use plain lines , barbed or unbarbed according as the pairs are unsymmetrical or symmetrical , with letters alongside of them ; like letters being used where otherwise like lines would be used , and unlike where unlike lines would be used ( fig. 17 ) .
Though this method is not so good as the previous one as regards its power of enabling us to visualize the Fig. 17 .
systems represented , it has distinct advantages for purposes of description , as we can denote pairs of different sorts by the annexed letters , the sorts ( sec. 84 ) of which denote the sorts of the pairs .
65 .
If denotes an unsymmetrical pair , we may denote the pair ba by ; or , as this is in some instances awkward , by where 66 .
Although , as we have seen , it is in many cases unnecessary to draw lines connecting all pairs of graphical units in order to completely define a system ( secs .
55-59 ) , it may in some cases be desirable to do so , especially where we wish to show how many different of pairs there are under consideration .
In such cases we may connect like pairs by like lines and unlike by unlike lines .
Thus the diagram of fig. 18 , which completely defines a system of six units , may be completed as in Fig. 18 .
Fig. 19 .
fig. 19 so as to show that there are two sorts of unsymmetrical pairs and one sort of symmetrical pairs .
67 .
In general , in diagrams in which lines other than links are employed , either in conjunction with links or not , like pairs of graphical units will be joined by like lines .
68 .
It can readily be shown that every form which admits of representation by .
A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
means of graphical units and lines of diffelent sorts , can , by the introduction of additional or auxiliary graphical units , be represented by means of graphical units and links only .
It will be convenient for the purpose of cription to suppose that the lines of different sorts are of different colours .
In the place of any red line substitute a red unit ( i.e. , one coloured red , instead of having a letter inside ) , this unit being connected by two links to the bwo units which were joined by the red line ( fig. 20 ) .
Effect this change in the case of ] the red lines ; and Fig. 20 .
similarly in the place of blue , yellow , \amp ; c. , lines , substitute blue , yellow , \amp ; c. , units , joined by links to the units which were connected by the several coloured lines .
The red units and links have precisely the same effect in rendering pairs distinguishable as the red lines had .
We may , if we please , substitute letters for the various colours in the added units , and so obtain a diagram consisting of graphical units and links , representing a collection of which the original one is only a component .
This component is , however , of the same form as before .
69 .
In the case of an unsymmetrical pair , , of undistinguished graphical which would ordinarily be joined by a barbed line , we must have two additional undistinguished graphical units in some cases , as in fig. 21 , where are the auxiliary units .
Fig. 21 .
70 .
In the case of unsymmetrical pairs of distinguished units we have seen that no barbs are necessary ( sec. 62 ) , and we need therefore only one additional graphical unit , as in sec. 68 .
71 .
All systems of which the forms are determined by the distinguishableness and undistinguishableness of units and pairs only , may be represented by the methods we have been considering .
But there are also systems , the forms of which are not so determined , but depend upon the distingnishableness and undistinguishableness of triads , tetrads , \amp ; c. Such systems may at first sight require other expedients for their graphical representation , but we shal ] see , as already stated ( secs .
33 , 60 ) , that these are not necessary .
72 .
Diagrams such as those we have been discussing may come under our consideration in other ways than as representing graphically systems of different forms .
In such cases it must not be assumed that the units under consideration are the small circles , or only those ; in general this will not be so ; the units may be the links , or the ways in which a number of small circles or spots can be connected by links , \amp ; c. If we ascertain exactly what the units really dealt with are , we may construct a graphical representation of the system they constitute , and this may be of a very different character from that of the diagram from which it is derived .
Aspects .
73 .
If two collections of units are undistinguished , they may be regarded as corresponding to each other in one or more ways , in each of which correspondences to each unit , and therefore to each pair , triad , \amp ; c. , of one collection there corresponds in the other a counterpart unit , pair , triad , \amp ; c. , undistinguished from the former in dress or other circumstance .
In any one of these correspondences , two corresponding units are regarded as occupying corresponding places , or , as we may express it , places of the same sort .
Furthermore , each of the corresponding units is regarded as belonging to one or the other of the two collections , each of which ; regarded also in the aggregate as a single unit .
74 .
Now the unit A which is dealt with when we thus regard a unit as occupylng a place of a particular sort in a particular collection , is a different unit from ; it may be called an aspect of 75 .
Thus when we consider a correspondence of two undistinguished collections , . .
and , . .
where corresponds with and so on , we deal with a collection of units , . . . . .
, . . . . .
where any unit is the unit arrived at by considering the unit as occupying a place of a particular sort , the unit being that arrived at by considering the unit .
as occupying a place of the same sort ; and similarly in the case of the other corresponding units .
Each of the collections , . .
and , . .
an aspect , the former of the collection , . .
and the latter of the collection , . . .
We ako regard , . .
in the aggregate as a single unit V , which may be termed a uni.fied of , .
and so also in the case of , .
76 .
The collections , and , . .
may be different aspects of the same collection , . .
, as a collection may be self-correspondent , and the number of units , . .
be accordingly the same as that of the number of sorts of places considered .
77 .
We may have a number of undistinguished collections , each of un ts , all corresponding to each other .
If there be such collections , in any correspondence in MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
which they all correspond we have unit aspects such as sorts of places , and each sort of place occupied by units .
78 .
, the aspect of in a correspondence , bears definite relations ( sec. 79 ) to three other units , viz. :\mdash ; ( 1 .
) of which it is an aspect .
( 2 .
) The unified aspect V arrived at by regarding as a unit the aspect of which it is a component .
( 3 .
) The sort of the place occupies in the correspondence .
79 .
The " " definite relations\ldquo ; of the preceding section consist merely in this , viz. , that the pairs which A makes with , and , are all distinguished from all the other pairs which A makes with other units .
( See secs .
143 ff , on Associates .
) 80 .
The tetrad , may be graphically represented as in fig. 22 ; the graphical unit A being joined to no other graphical units besides , and .
Since Fig. 22 .
the units , and , are obviously distinguished from each other , the graphical units are made so also .
81 .
It appears , then , that in studying the form of a system by means of correspondences of undistinguished components , we really regard as a component of a more extensive system , containing ( 1 .
) the system ; ( 2 .
) the system X composed of units which are conceived of as sorts of places ; ( 3 .
) the system composed of units which are conceived of as unified aspects of all the various component collections of ; ( 4 .
) the system composed of units which are conceived of as aspects of single units of S. The pairs connecting and the joint system , break up into two systems , graphically represented by linked and unlinked pairs of graphical units respectively .
82 .
A graphical representation in which the units of , are all represented by graphical units , and links connecting the graphical units are drawn in the manner indicated in the last two sections , fully defines the form of 8 , as it completely indicates what components are undistinguished and what are not .
Such a graphical representation is sometimes termed a ' linkage so that we may say that every system may be graphically represented by a linkage ( secs .
194 , 195 ) .
83 .
The system was described in sec. 81 as consisting of unified aspects of all the component collections of , i.e. , the component pairs , triads , \amp ; c. It is , however , unnecessary so far as defining the form of is concerned that this should be so .
We shall see that it is sufficient that should include a unified aspect V of the whole system , and all other unified aspects of which are undistinguished from V ( secs .
194 , 195 ) .
Letters , their Sorts and Positions .
84 .
Let us now turn aside to consider certain points with regard to the representation of units by letters .
Any letter admits of being repeated as often as we please ; each repetition being one of a number of letters of the same sort .
When in common parlance we speak of letters \amp ; c. , we are really referring , not to the letters themselves , but to their sorts .
Each letter ( not sort of letter ) occupies a different position on the paper on which it is written or printed ; as in the case of the letters , we have a number of positions of the same sort , each called ' ' the same position\ldquo ; with reference to another of the number .
For example , if we have two arrays of the same number of letters , e.g. , abcde and pqrst , we say that " " occupies the position in the first array that does in the second meaning that the sort of the position occupies is the sort of that which occupies .
85 .
The sorts of the letters and the sorts of the positions are both dealt with as units .
Further , we regard each letter as belonging to a particular collection regarded in the aggregate as a single unit , say a unifi , ed collection .
86 .
Thus a letter bears definite relations to ( 1 .
) its sort ; ( 2 .
) the unified collection to which it belongs ; ( 3 .
) the sort of its position .
Representation of Aspects of Collections by Arrays of Letters .
87 .
Thus we have an exact copy in a different dress of the sta , te of things considered in secs .
73-83 .
If then we represent ( 1 .
) the units of any collection by the sorts of letters ; we may represent ( 2 .
) the so , rts of places occupied by those units in a correspondence with an undistinguished collection by the sorts of the positions of the letters ; ( 3 .
) an aspect of the collection by a collection of letters in sorts of positions ; ( 4 .
) unified aspects by unified collections of letters ; ( 5 .
) aspects of single units by single letters .
88 .
Taking , then , arrays as our collections of letters , an aspect may be represented thus . .
( i.e. , in the manner adopted in sec. 7 ) , where the order of the letters is to be regarded as material to this extent , viz. , that though the aspect which MR. A. B. KEMPE ON THE THEO OF MATHEMATICAL FORM .
represented by abcdef .
, .
might equally well be represented by .
, , if this be done , the aspect originally represented by bdacfe . .
must be represented by debaef . .
, and that originally represented by pqrstu . .
by qsprut . .
, and similarly in the case of other aspects and other transpositions of the sorts of the positions occupied by the letters in the arrays .
Whatever change be effected in the order of the letters in one array , a precisely similar change must be made in the order of the letters in each of the other arrays of the same number of letters .
The relative order in arrays of different numbers of letters is , as such xrays represent collections of different numbers of units which are therefore distinguished , and are not regarded as corresponding .
Elementary Properties of Aspects .
89 .
The symbols will be used to represent " " is undistinguished from\ldquo ; and is distinguished from\ldquo ; respectively ; and such expressions as " " let \ldquo ; must be read , " " let be undistinguished from \ldquo ; ; and similarly in other cases .
90 .
Symbols , such , in which commas separate the letters , will always be supposed , as heretofore , to represent a collection of units without reference to particular aspects , the order of the letters being accordingly supposed to be immaterial .
91 .
A statement such as abed . . . .
implies that the components represented by taking corresponding letters on each side of the undistinguished ; e.g. , .
A statement such as does not imply that , . .
, . .
for it is consistent with the statement that abed . . . .
which implies that , . .
, .
92 .
If abed . . . .
, then . .
; and similarly in the case of any other transposition of letters .
93 .
If pqrs .
be merely . .
in a different order , so that we are considering two aspects of the same collection , and if we repeat on pqrs . .
the transposition by which it is derived from abed . . .
, we get another aspect which by the preceding section is undistinguished from pqrs . .
, and therefore from abed .
94 .
If abed , and if , . .
be units other than , . .
, there must be units , other than , such that abed . . .
wxyz 95 .
If abed . . .
each collection having the same number of units , and if , . .
be units other than , . .
there cannot be any units , . .
such that abed . . . .
96 .
Every collection of units has aspects .
If aspects of the collection are undistinguished from each other , but are distinguished from all other aspects of the MR. A. B. KEMPE ON THE THEORY OF MATHElfATICAL FORM .
collection , must be a factor of , and the aspects break up into systems , each consisting of aspects .
97 .
It follows immediately from se that if we write down the symbols abed .
\amp ; c. , representing any aspect and all other aspects which are undistinguished from it , and we transpose the letters of . .
and make precisely the same transpositions in the case of the letters of the symbols representing the other undistinguished aspects , we shall get a collection of symbols of undistinguished aspects which may or may not be the same as the former .
98 .
If abed .
be an aspect of a whole system of units , so that all the aspects undistinguished from it are aspects of the whole system , there being in all undistinguished aspects , the different transpositions of sec. 97 will give us all the different systems of aspects of referred to in sec. 96 .
99 .
If the array of letters abed .
representing any aspect of a whole system be given , and also those representing all the other aspects of which are undistinguished from obed . .
, then the form of is given ; i.e. , if pqrs . . .
be an aspect of aIly component collection of , we know what other aspects of component collections of are distinguished and what undistinguished from pqrs . .
For add letters to pqrs .
until we get an array pqrs . . . .
representing an aspect of the whole system S. If this is not the same as one of the given arrays , transpose the letters of any one of the latter so that it becomes pqrs . .
lmn . .
, and make precisely similar transpositions in the case of each of the other given arrays .
If now wxyz . . .
, it follows from sec. 94 that there be an aspect wxyz . .
of the whole system , such that wxyz . . . .
; and if wxyz . . . .
it follows from sec. 95 that there is no aspect wxyz . . .
of the whole system such that wxyz . . .
; i.e. , in the former case there must be a symbol uxyz . .
amon the transposed alrays , in the latter case there can be no such symbol .
Tabular Representation of Systems .
100 .
A convenient mode of arranging the symbols abed .
, \amp ; c. , representing a system of undistinguished aspects of a whole system of units , . .
is Fig. 23 .
to place them one above another so that letters occupying " " the same position\ldquo ; in the different rows may lie in the same .
For example , the system of fig. 23 may MR. A. B. EMPE ON THE THEORY OF MATHEMATICAL FORM .
be represented thus : \mdash ; a a a a a a 101 .
The symbol thus arrived at , of letters arranged in rows and colunms , will be termed the tabular representation of S. Each of the letters be termed an element .
102 .
The order of the rows is clearly .
Any alteration in the order of the umns merely substitutes for one system of undistinguished aspects of another system of undistinguished aspects of , and any one of these systems of undistinguished aspects the form of , thus the order of the columns is also immaterial .
The material point is that certain elements are all in the same row , and certain elements all in the same column .
103 .
Each sort of letter represents a unit of column regarded as a unit represents a sort of place ; each row represents an aspect of , or , regarded as a unit , a unified aspect of ; and each letter , or element , an aspect of a unit of S. 104 .
Any row may be regarded as derived from another row by a substitution of the letters of the row .
If the same substitution be effected upon any row , we get a row which is also a row of the table .
105 .
The substitutions by which any row of the tabular representation of is derived from another may be said to be proper to 106 .
If we confine our attention to certain rows and columns only of the tabular representation of , we get a table which will be termed a of the whole table .
The order of the rows and columns will be disregarded in a constituent as in a complete table .
107 .
If and can never appear in the same column .
Thus the columns break up into lots , each lot to a single system . .
If , then appears in every column that does , and in every column that does .
109 .
Generally if abed . . . .
we have in the ] a constituent . .
, and if abed . . . .
we have no such constituent .
Thus the forms of the various components can readily be ascertained from the table .
110 .
If we desire to consider the forms of portions only of a system , we may confine our attention to such constituents as will exhibit the form .
Thus in the case of sec. 100 the constituent shows that the pair is a symmetrical one .
111 .
A discrete heap will be represented by a table of one row ; a single heap of units by a table of rows ; and a heap consisting of single independent heaps of , . .
units respectively will have a table of . .
rows .
Correspondences of Undistinguished Collections .
112 .
The various correspondences of undistinguished collections are indicated by two-row constituents of the tabular representation of the system of which the collections are components .
11.3 .
If the two rows of the constituent contain like letters in different orders , the constituent will be said to indicate a self-correspondence , i.e. , a respondence of a collection with itself ; and not with another collection ( sec. 76 ) .
Any two-row constituent containing two complete rows of the table representing a system will dicate a self-correspondence of , and the table may accordingly be said to indicate the form of by .
all its self-correspondences .
114 .
If the two rows of the constituent representing a correspondence of with itself contain like letters in the same order , the constituent will be said to indicate an identical-correspondence of C. The identical-correspondence may be regarded as included among the self-correspondences of C. The self-correspondence of a single unit is an identical-correspondence .
115 .
In general the table ) resenting a system will have rows which are partially alike , indicating that in some of the self-correspondences of some of the components are identically-correspondent .
For example , in sec. 100 the unit is always identically-correspondent .
Thus constituents may have several rows which are duplicates .
If we merely desire to consider those correspondences which the constituent indicates , we may , of course , omit duplicate rows .
116 .
If one two-row constituent be a part of another , the correspondence indicated by the former may be said to ' occur in\ldquo ; that indicated by the latter ; as when the latter correspondence occurs the former also occurs .
117 .
lf we regard ] collections of units under consideration in an investigation as components of the universal system which comprises all units , or , as is sufficient , as components of the whole system of units considered in the investigation , we may regard every correspondence of undistinguished collections as occurring in one or more self-correspondences of the universal or more limited system .
From this point of view we see that a correspondence of two undistinguished collections , or the self- correspondence of a collection , restricts the possible correspondences and self-correspondences of other collections .
In some cases it may determine absolutely the correspondences and self-correspondences of the other collections , not permitting MR. A. B. KEMPE ON THE THEOB , OF MATHEMATICAL FORM .
alternative correspondences or self-correspondences , in others it may partially restrict them , in others it may exercise no controlling effect whatever , so that collections may go through all correspondences and self-correspondences with as much freedom as if there were no correspondence already existing .
118 .
If two cohections are such that each can through all its self-correspondences while the other remains identically-correspondent , they will be said to be independent .
If two collections are not independent , they will be said to be related .
119 .
If and be the number of self-correspondences of two systems and respectively , the table representing the joint system will have rows if and are independent .
120 .
In the self-correspondences of a system every correspondence of undistinguished components occurs ; but it is not of course in general the case that in the self-correspondences of any collection all the correspondences of undistinguished components of that collection occur .
For example , in sec. 100 the only self-corre- a spondence of , is given by the constituent and here no correspondence ' of the undistinguished units occurs .
121 .
When investigating the correspondences of a number of -ads we may for some time be occupied with the consideration of correspondences in which of the units always remain identically-correspondent .
The absence of change in the correspondence of the -units may lead us to forget or overlook the fact that we are considering correspondences of the -ads , and we may suppose that we are dealing with -ads only .
When then we find that a certain correspondence of two -ads apparently does not exist , we must look closely to see whether we are not really considering correspondences of -ads , and whether a change in the correspondence of the units may not lead to the correspondence of the which is supposed not to exist .
122 .
Units which in any correspondence of two undistinguished aspects are identically-correspondent may be termed the foci of the correspondence .
Three or more undistinguished aspects such that the foci of the correspondences are the same in the case of each pair , may be said to be confocal .
We may also use the term as applicable to the case of corresponding components of the corresponding confocal aspects .
123 .
In cases in which , as mentioned in sec. 121 , we consider -ads as if they were n-ads , we really pass from the consideration of the original unit , to that of other units which are arrived at by taking the foci with each of the other successively , and , then regarding the resulting -ads as single units .
124 .
Differences between things can always be ignored , and thus we may at one time regard two collections of units as distinguished , and at another , by ignoring differences between them , as undistinguished .
Here we are dealing with certain units , upon which the differences depend , in addition to those of the two collections .
125 .
For example in the case of the system of fig. 24 , although all the graphical units are alike if taken apart from the barbed lines joining them , the two triads Fig. 24 .
and are because plain lines are distinguished from wavy ones .
If we ignore the difference between straightness and waviness , we are really dealing with the eight units , and instead of dealing with units , \amp ; c. , are really dealing with the triads , \amp ; c. When we regard the two triads and as distinguished , we deal with spondences represented by the constituent consisting of the first , or last , nine rows of the table at the end of the section , in which it will be seen that and remain identically-correspondent , and the correspondences are such as would exist if and were distinguished .
When we ignore the difference between straightness and abcdefLW waviness , we admit correspondences such as defabcW The table representing the whole system of eight units is as follows:\mdash ; beefd LW abcfde LW aefd beafdeLW a bdef LW cabefdLW LW WL of bc WL efdabcWL WL efdcab fdeabcWL fdebcaWL fdecabWL .
A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
126 .
Many of the relations which collections of units hold to each other are apparently independent of the form of the collections , or of the form of the system , of which they are uents .
In dealing with such relations we ignore differences , and regard the units composing the collections as single heaps , ) , we do not deal with the units , . .
of , but with others , S , .
which constitute a single heap , but have the same self-correspondences as , . .
have as long as certain other units remain identically-correspondent .
127 .
Since we can in all cases ignore differences , any system may be regarded as a single heap , the peculiarities of form which it possesses in any particular investigation being regarded as due to the fact that we are not considering alone , but in conjunction with other units , those correspondences only of being dealt with which admit of the additional units remaining identically-correspondent .
All statements , therefore , as to t , he distinguishableness and undistinguishableness of components of and as to their being of particular forms may be taken as relative , viz. , as sts , tements that the components have the correspondences characteristic of those forms as long as certain units detached from remain identically-correspondent .
128 .
It may be laid down generally , that in almost every instance where we seem to investigate a base system which may be regarded and spoken of as being of units and of a particular form , we really deal with a single heap system of units and a system which remains identically-correspondent while the units of go through the correspondences characteristic of the system S. 129 .
Any collection of units which while another collection remains identicallycorrespondent has self-correspondences characteristic of a collection of a special form may be said to be of that form relatively to Sets .
130 .
If abed .
pqrs .
, are undistinguished components of a collection , . .
, . .
, .
, then the units , . .
which are such that . .
lmno . . .
wxyz .
may or may not be units of the collection , and in some cases cannot be selected so as to be units of the collection .
Ifthe collection be such that whatever undistinguished components abed . .
, we select , and whatever other component lrnno .
we select , .
can always be selected from the collection , then the collection will be termed a set .
131 .
A system is obviously a set .
A set is not necessarily a system ; it may be one of a number of undistinguished sets which together compose a single or multiple system .
In most investigations our inquiries are directed towards the discovery of the forms of component sets of the base system .
As far as the distribution of its distinguished and undistinguished components is concerned , a set in no way differs from a system .
132 .
In the self-correspondences of a set every correspondence of its undistinguished components occurs .
Thus the } ituent which represents a set has sukconstituents indicating all the correspondences of the components of the set .
133 .
In a graphical representation of a set the diagram not only indicates what units , \amp ; c. , are distinguished and what qhed , but is such that if we ignore geometrical position in accordance with sec. 41 , and the reference letters in accordance with sec. 42 , the units , \amp ; c. , will be actually distinguished and undistinguished , and the diagram will represent the form of the set by bein actually of that form . .
Any collection of units which is not a set may be said to be imperfect . .
If any collection of units is such that its component pairs are all distinguished from the pairs connecting it with units detached from it , the collection is a set .
For let lmn . .
and . .
be any two undistinguished components of the collection , then if be any unit of the collection , a unit which is such that . .
is undistinguished from . .
must aJso be a component of the collection , otherwise the pairs connecting and , . .
will be pairs connecting a unit not of the collection units of the collection , and they will be undistinguished from the pairs connecting and , . .
which are all pairs of the collection .
136 .
Let be any unit of a single set of units ; consider the pairs formed by with other ullits of ; take any one of these xp ; let the number of pairs xp , which are undistinguished from xp , be ; then in the case of any other unit of the number of pairs yi , yj , , . .
which are undistinguished from xp is also .
Further the number of pairs , . .
which are undistinguished from is also ; fur the number of such pairs must be the same in the of each unit , and thus we have , i.e. , spects unique respect to Collections .
137 .
If .
abc . . .
, then the aspects xyz . .
and uvw may be said to be duplicates with respect to the collection , .
If there is no duplicate of . .
with respect to , . .
then .
may be said to be unique with respect to , . .
It , should be observed that if . .
is unique with respect to , . .
there may be an aspect . .
such that . .
abc . . .
cab 138 .
If the aspect abc .
be unique with respect to the collection .
and if the apsect def . .
be unique with respect to the collection , . .
; then abc . .
is unique with respect to , . .
I give the proof in the case in which the three collections are detached ; the proof in other cases is somewhat longer but presents no difficulty .
If .
is not unique with respect to , . .
there are units such that . .
, and there are ( sec. 94 ) units .
such that abc .
def . . . . . .
; but def .
unique with respect to , . .
therefore . .
is def . . .
, therefore abc . . .
def MR. A B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
25 . . .
def .
, therefore .
is not unique with respect to , .
contrary to the hypothesis ; i.e. , . .
is unique with respect to , .
139 .
If a single unit be unique with respect to a collection , . .
we may represent by the symbol ( abc ) or [ abc . .
] \amp ; c. The two symbols ( abc . .
) and will , where the different sorts of brackets are distinguished from each other , represent two units which are distinguished from each other , and each unique with respect to , . .
; or we may use the two symbols to represent guished units which are each unique with respect to .
, the difference between the brackets indicating that the units are not identical .
140 .
If abc . . .
then ( abc .
) .
where ( abc . .
) and ( pqr . .
) may be the same or different units ; and ( pqr .
) . .
If . .
then ( ubc . .
) and ( pqr . .
) may or may not be distinguished units .
141 .
If we consider aspects such as Pabc . .
, Qabc .
, Pdef .
\amp ; c. , where in each aspect one only of the units \amp ; c. , appears , we may represent \amp ; c. , by brackets of different sorts , and write the units which are unique with respect to , . .
, \amp ; c. , thus ( abc . .
) , [ abc .
( def . .
) , \amp ; c. , where is represented by the brackets , by the brackets 142 .
Suppose , , , then we may represent by the symbol ( .
or , each symbol representing and at the same time an aspect of a collection of which is a component .
We may have such symbols with various different sorts of brackets , e.g. , if will be represented by [ mn ] e ) ) .
143 .
If , .
be any collection of units , and if be another unit , such that the pairs , . . .
are distinguished from all pairs which makes with units which are not components of the collection , may said to be an associate of the collection , . .
144 .
If the pairs , . .
are all undistinguished from each other , may be termed a single associate .
145 .
If some of the pairs are distinguished from each other , may be termed a multiple puir associate .
146 .
In the special case in which all the pairs are distinguished from each other a multiple pair associate may be termed a discrete associate .
147 .
A single pair associate of a collection can only exist if all the units of the collection are undistinguished from each other .
Also in the cabe of a multiple pair associate of a collection units of the ] ] ection which make pairs with the associate which are undistinguished must themselves be undistinguished .
148 .
An associate of any description may be graphically represented by a unit connected by links with each of the units of which it is an associate .
149 .
If we desire in the case of a ] tiple pair associate to indicate the disisheda pairs , we may employ lines of various sorts in lieu of links , thus : Fig. 25 .
150 .
Many of our conceptions and definitions of systems of units involve the idea of associates , and the graphical defirlition of systems by means of links is very convenient as enabling this to be visually represented .
151 .
In many cases units which we might at first sight regard as associates of a system of some definite form , are really associates of a single heap system accompanied by a em which remains self-correspondent while goes through correspondences characteristic of the form of ( sec. 128 ) .
Uni.fied Aspects . .
The unit arrived at by regarding any aspect of a collection , . . .
as ngle unit , is unique with respect to the collection and may be represented by the symbol ( abc . .
) .
153 .
Two aspects when unified are two distinct units , and noone unit .
Thus if .
and . .
are different aspects ( abc .
) is not ( lmn . . .
) .
154 .
Every aspect is unique with respect to its unified aspect ; for if ( abc ) .
then .
and then abc . . .
( lmn .
) ; thus we have . .
( abc . .
) .
( lmn .
) ; and ( lmn .
) is not abc . .
) ( sec. 153 ) , thus is not unique with respect to . .
, which is contrary to sec. 152 ; therefore , \amp ; c. .
If and be two unified aspects of the same collection they are unique with respect to each other .
For , by sec. 152 , is unique with respect to the collection , and , by sec. 154 , the aspect of the collection which is when unified is unique with respect to , thus , by sec. 138 , is unique with respect to imilarly is unique respect to 156 .
The pairs which a unified aspect makes with , .
respect ve .are all distinguished from each other .
For if , then there are units , . . .
such that ( abc .
) .
i.e. , such that abc . . . . . .
, i.e. , such that ( abc .
) ( blm . . .
) .
, i.e. , such that i.e. , such that ( blm . . .
) is not unique with respect to , contrary to sec. 152 , therefore , \amp ; c. .
A. B. KEMPE ON THIJ THEORY OF MATHEMATICAL FORM .
157 .
In the same way it may be shown that the pairs which ( abc .
) makes with , . .
respectively are all distinguished from the pairs which makes with any other units , . .
not components of the collection , .
158 .
A unified aspect ( abc ) is accordingly a discrete pair associate of the collection , . . .
( sec. 146 ) .
159 .
It should be observed that in the case of a number of undistinguished aspects no unit can be said to be the unit arrived at by regarding any particular one of those aspects as unified .
All we can say is that any one of a number of units may be so regarded .
If , however , one of that number is regarded as representing a particular one of the unified aspects , some other definite unit of the number must be regarded as the unit which represents another given unified aspect , i.e. , the two unified aspects will be such that they can only be represented by certain pairs of the units and not by any pair .
160 .
When we represent aspects by single letters , those letters really represent the unified aspects .
Whatever relations as to distinguishableness or undistinguishableness exist between the aspects , there will be precisely the same relations between the uIlified aspects ; so that we may deal with either the unified or non-unified pects .
Thus if , be the aspects abed , efgh , ijkl , uvwx , when regarded as units , then if abcdefgh ijkluvwx , we have AB 161 .
The conception of a unified aspect is an " " accidental\ldquo ; one ; for the units representing unified aspects of components of a system may be regarded as representing any other things holding ] relations to the units of S. method of defining systems by regarding their units as unified aspects of components of other systems is , however , so convenient and simple that it will be frequently employed , and the accidental part of the definition being borne in mind , no danger can arise from the employment of this method of arriving at systems .
Correspondences of Systems of like Forms .
162 .
We may consider correspondences of any two independent ystems S and of the same form , in which every component of corresponds to a component of of the same form , and we may regard these correspondences as units .
The number of such correspondences is in general greater than the number of selfcorrespondences of each system , as in the case of the latter we only suppose undistinguished components to correspo1ld , and do not admit correspondences of distinguished components of like form .
163 .
If , however , we start with a correspondence of and and then restrict ourselves as regards other correspondences to those in which each component of corresponds only with such components of as are undistinguished from that with which it corresponded in the first instance , we shall get the same number of respondences as there are self-correspondences of each collection 164 .
If we then take another correspondence not included among the former , we shall get a second collection of correspondences ual in number to the former , and ) on until the correspondences are all exhausted .
165 .
For example the six graphical units of the system of fig. 26 constitute two Fig. 26 .
abc independent tems of the same form .
Starting with the correspondence which abc we may represent as a unit , we may add correspondences and fde which we may represent by units and respectively .
Here we have only correspondences such as those referred to in sec. 163 .
It is to be observed that we cannot properly say that represents any particular one of the three correspondences , for they are undistinguished ; the three units , together represent the three correspondence ; but if is arded as representing any definite correspondence and will each represent definite correspondences , for , are each unique with respect to each other .
If we now consider the correspondence which we have components of abc ab the same form corresponding as before , we get the correspondences fed and .
We may represent these by the units , N. The whole system of units considered is represented by the following table \mdash ; abcefdQRPNLM ab.cfdeRPQMNL beadefRPQNLM beafdeQRPLMN cabefdRPQLMN 166 .
We may of course consider self-correspondences of a system in which components of like form correspond which are distinguished from each other .
Thus the case of the system , of fig. 27 we may consider correspondences such MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
abcde abcde as or It is , however , to be understood that the term " " self-correspondence\ldquo ; aedcb acebd will always be used , as heretofore , with reference to self-correspondences which involve the correspondence of undistirJguished collections only .
Fig. 27 .
167 .
The two systems and X of sec. 81 are of the same form .
In the tabular representation of either the columns will represent the units of the other .
Further , and X are independent systems , no special relation except that of similarity of form exists between them .
Each aspect of is arrived at by con .
a correspondence of and X ; it may in fact be regarded as such a correspondence , so that a unified aspect is a unified correspondence , and the whole system of unified aspects of represent a system of unified correspondences of two systems of the same form , in which we are restricted to such correspondences as those referred to in sec. 163 .
168 .
When two systems are regarded as corresponding they may be spoken of as being projections of each other in as many ways as there are unified correspondences .
169 .
If we represent the units of one of two independent systems of the same form by the ( Aa ) , , , \amp ; c. , we may represent those of the other by the symbols , , , \amp ; c. , where in one correspondence we may conveniently suppose that corresponds to , to , and so on ; but it is not to be supposed that this correspondence is distinguished from others .
Replicas .
170 .
If , . .
and , S , . . .
be two systems of units such that and are unique with respect to each other , as also and and \amp ; c. , and if , and being any two units of the first system , when we have also then , S , . .
may be called a replica of , . .
171 .
The replica of a system is of the sam form as , and same relations to other systems as has .
In the tabular representation of and its ] , whatever transpositions of the letters representing an aspect of takes place in ssing from one row to another , precisely the same transposition takes place in the of the letters representing an aspect of the replica .
MR. A. : KEMPH ] ON THE THEORY OF MATHEMATICAL FORM . .
The replica of a multiple system may have some single systems in cOmmon with ; some of the single systems may be their own replicas .
173 .
The letters placed adjacent to graphical units for the purpose of reference compose a system which is a replica of the graphical system .
Independent and Related Systems .
174 .
Let be a system consisting of two detached systems and ; let , . .
be the units of the number of those units , and the number of the self-correspondences of S. Let , . . .
be the units of the number of those units , and the number of the self-correspondences of .
Let be the num- ber of self-correspondences of the system : cannot be otherwise there would be ]icate rows in the tabular .representation of .
Consider now a set of rows in the tabular representation of obtained by taking any row and all others in which the letters , . .
remain untransposed from the order they have in R. Let the number of rows in the set be .
A transposition indicated by two rows of the set if made to operate on any row of the set will clearly give a row also of the set .
Now take any fresh row not of the set ; this must also be one of a set of rows detached from the former set .
Thus all the rows divide up into sets of rows , and we have .
If be the corresponding number in the case where , . .
remain untransposed , we have .
Thus we have or 175 .
Let , then , and the two systems are independent .
176 .
We must have unless and have a oommon integral factor ; that two systoms and must be independent unless the numbers of their respective self-correspondences have a common factor .
The two systems may of course be independent when and have a common factor .
177 .
If and are prime to each other and may be said to be prime to each other ; so that systems which are prime to eaoh other are independent .
178 .
If and must have a oommon factor , and the two systems nd will be related .
If , , x'y'z ' .
are undistinguished components of , and abed . .
be an aspect of the whole system , then the aspects .
abed .
and x'y'z ' . . . .
will not in general be undistinguished each other ; and any graphical diagram representing the system must have lines , or successions of lines , connecting units of to units of so that and cannot be drawn on separate sheets of paper , 179 .
If we have three detached systems , ; and and are related , and also and , it does not follow that and are related ; the single consideration that and may have a common factor , and also and , without , and having one , shows this .
180 .
If , then are replicas of each other .
MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
181 .
If , and will be a factor of , and may be said to be a factor system of S. We have here 182 .
If be a factor system of we have ; i.e. , there is no self-correspondence of , other than the identical-correspondence , in which is identicallycorrespondent .
If is not a factor system of , we have , and there are self-correspondences of , other than the identical-correspondence , in which is identically-correspondent .
183 .
Hence if be a factor system of , there are no units of which are duplicates with respect to S. And if there are units of which are duplicates with respect to , then is not a factor system of S. Further , if be not a factor system of , there are units of which are duplicates with respect to , unless and are replicas of each other .
And if there are no units of which are duplicates with respect to , f , hen is a factor system of , unless and are replicas of each other .
184 .
If be a factor system of , each single system of must be so also .
For if be the number of self-correspondences of a single system of , we have , i. e. , , i. e. , 185 .
If be a factor system of , it is also a factor system of , the single system composed of the unified aspects of S. This follows immediately from the fact that the combined system , has the same number of self-correspondences.as has .
186 .
Hence also every single system of , a factor system of , must be a factor system of 187 .
A system which has no factor systems , except systems of one unit , may be termed a prime system .
188 .
If is a factor system of , and is a factor system of , then is a factor system of S. For let units of be duplicates with respect to ; then we have .
Now if be any unit of , there must be a unit such that a\ldquo ; c'S ; and since is a factor of , we cannot have , unless is .
Thus we have , i.e. , we have , whatever of we take , i. e. , is and there are no units of duplicates with respect to , i. e. , is a factor system of S. Three Modes of Compounding Systems .
189 .
There are three modes of deriving a system from two or more other systems to which a passing reference may here be made .
In the first the compound system is arrived at by regarding the -ads connecting independent or related systems as units .
190 .
Again , we may have a system which may be regarded as composed of independent ( and therefore detached ) undisti1i sets , each of a given form and such that their unified aspects compose a system of the form 191 .
An important special case of such a compound system is that in which the form is that of a single heap of units .
Here being the form of any system , the compound system differs from in that we have in lieu of any unit of a single heap of units , each unit of which may be called , the form relations between the connecting -ads of of the heaps being the same as those existing between the -ads of S. Thus if in we have , in we shall also have pqrs , whatever units of the heaps . .
, .
\amp ; c. , we select .
192 .
If we combine two connected component collections and of , the resulting collection cannot contain those units which are common to and twice over ; i. e. , the result of combining the collections , and is the collection If in we select collections and so that the units of the first are not the same units as of the second , the sum of the two collections will be the collection 193 .
The third mode of composition is that in which the derived system has all the self-correspon dences which each of a number of systems , of the same number of units , has .
We may represent the units of such a system by symbols , ( Bb . .
) , ( Cc .
) \amp ; c. , where , C. .
are units of one of the compounded systems , those of another , and so on .
General Method of Representing a System .
194 .
Let each unit of a system , each unified column , each unified row , and each element of its tabular representation , be represented by a graphical unit , using different kinds of graphical units in the case of the units of , the rows , columns , and elements respectively , four kinds in Connect each graphical unit which represents an element by links to ( 1 .
) the graphical unit representing the unit of of which the element represents a unit aspect ; ( 2 .
) the graphical unit representing the uni-fied row in which the element lies ; ( 3 .
) the graphical unit representing the unified column in which the element lies .
We get a graphical representation of a system of units , of which is a component system .
195 .
It is obviously not always necessary to employ the somewhat cumbrous mode of graphical representation here given ; simpler methods can be adopted in special cases .
Frequently , as we have already seen , it will not be necessary to have any auxiliary graphical units , but merely those representing the units of itself Networks .
196 .
The pairs of any single collection of pairs are either all component pairs of a single system of units , or all connecting pairs of two single systems of units . .
A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
197 .
Where a single system of pairs connects two single systems and 2 , the number of pairs of which connect any unit of to units of 2 is the same as the number of pairs of which connect any other unit of to units of 2 .
The number in the case of units of is the same as the corresponding number in the case of units of 2 if and 2 contain the same number of units , but not otherwise .
198 .
Where the pairs of a single system of pairs .
are components of a single system of units .
the number of pairs of the former system which connect any unit of to other units of is the same as that in the case of any other unit of S. Also the number of such pairs connecting units of to is the same as that of those connecting units of to any other unit of , being in both cases aJso ( sec. 136 ) .
199 .
A single system of component pairs of any single system of units constitutes a simple network of which the number of the last section may be called the waynumber .
This network consists of one or more portions each continuous and detached from the other portions .
Where there are two- or more detached continuous portions of a simple network , each is undistinguished from the others , for component pairs of distinguished portions would be distinguished .
200 .
Every simple network is accompanied by one in which the pairs connect the same units but in the reverse order , so that if is a pair of one network , ba is a pair of the other .
The two networks may be called the reverses of each other .
If we represent the unified pairs of the one by single let .ers , we may represent those of the other by .
Here the sorts of the letters may be regarded as representing the networks when regarded as units .
201 .
We may use the symbol ( a ) to represent the network , of which is a unified pair .
Here ( a ) takes the place of the sort of .
Or we may use ( ab ) ( sec. 263 ) .
202 .
The units of a continuous detached portion of a simple network form a set , for any two are connected by a chain of one or more pairs of the network , while no component unit is connected by such a chain with a detached unit , so that all component pairs are distinguished from allconnecting pairs of the collection ( sec. 135 ) .
203 .
Thus a simple network divides the units of a single system into one or more detached sets , each of the same number of units .
If , then , be the number of units iu each set , and the number of sets , the number of units in the system .
Thus and must each be integral factors of 204 .
Equations such as where , may be employed to denote the fact that there is a simple network which divides the units of into sets of units , another which divides them into sets of and so on .
We may have equations such as aaa ' showing that there are three different simple networks which divide into of units , and so on .
We may in such a case write the equation thus : 205 .
Two distinguished simple networks , components of a single system , either divide into the same detached sets , or the pairs of one of the simple networks are pairs connecting the detached continuous portions of the other .
For pairs belonging to one simple network must either be component pairs of the sets into which is divided by the other simple network or all pairs connecting them .
206 .
A network composed of two or more simple networks may be called a compound network , and may be called double , treble , \amp ; c. , according to the number of the component simple networks .
A compound network will consist of one or more undistinguished detached continuous compound portions .
The units of each of these portions constitute a set .
207 .
Where a network simple or compound consists of only one detached portion , it may be said to be complete .
208 .
If the pairs of each simple network of a compound network connect detached portions of the compound network composed of the remalmng simple networks of , then the compound network may be said to be pure .
209 .
Any two simple networks of a pure compound network may be said to be outside each other .
210 .
If each pair of a simple network connects units which are both components of the same detached portion of a simple or compound network , the latter network may be said to enclose the former .
Chains .
211 .
A succession of undistinguished pairs , .
may be termed a simple .
Where the chain has no terminal units it may be said to be dosed .
Every simple chain is a portion of a closed chain , unless it contains one pair only , when it may be a pair connecting two systems .
212 .
ompound c are such as contain distingu shed airs , or undistinguished pairs of opposite polarities .
213 .
In the case of the pairs connecting two sy stems we may have networks consisting entirely of undistinguished pairs ; but every chain in each consists successions of pairs alternately of opposite polarities .
214 .
A complete network of a system contains simple or compoun chains connecting every two units of the system .
215 .
If a chain of pairs of a pure compound network conta ns one tains one air only from one of the component simple networks , it cannot be closed , for if it were sng the le MR. .
B. KEMPE ON THE THEORy OF MATHEXIATICAL FORM .
pair would not connect two detached portions of the compound network composed of the simple networks from which the remaining pairs are taken , but two units of a continuous portion .
216 .
The number of pairs in a closed chain may be termed its period .
The period of a closed chain is of course the same as the number of units it connects .
217 .
A pair of units may be regarded as constituting a simple closed chain of period 2 ; thus Fig. 28 .
218 .
It may also be convenient in some cases to regard each unit as connected to itself by a single barbed line constituting a simple closed chain of period 1 , thus Fig. 29 .
219 .
We may represent a chain of pairs thus , where represent units , and the sorts of the networks of which the pairs , are respectively components ; \amp ; c. , being the letters placed alongside of the plain lines if we employ the graphical method of sec. 64 .
In some cases it will not be necessary to employ specific letters to represent the units , they may be represented by the spaces between the letters representing the unified pairs , e.g. , we may represent the chain given above thus , where only the terminal units are given .
If we have and , we see that the two units are connected by the two different chains and .
We may have equations such as denoting the fact tha6 the three chains can have the same terminals , and we may write a chain such as for shortness , and similarly in other cases .
221 .
We may employ as the symbol to be attached to the barbed line of sec. 218 in accordance with the method of sec. 64 the symbol 1 .
An equation such as will then show that the chain on the left hand side of the equation is closed .
Groups\mdash ; Circuits .
222 .
Any single set which is such that each component unit is unique with respect to each of the others , may be termed a group .
223 .
The single system of unified aspects of any system ( sec. 185 ) is a group , as each unified aspect is unique with respect to each of the others .
224 .
Every component set of a group is a group .
S6 MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL .
The table , or constituent of a table , representing a group of units has rows , a letter of any one sort appearing once and once only in each column .
226 .
The simple networks of a group are all one-way ( sec. 199 ) networks .
The closed chains which constitute the detached portions of any network of a group may be called simple circuits .
Any closed compound chain of a group may be called a compound circuit .
227 .
If the component pairs of one simple circuit are undistinguished from those of another , the two circuits must have the same period .
228 .
If the period of a simple circuit be where and are integers , there will be simple circuits of period , and of period , connecting the units of the circuit .
229 .
Now the number of units in a simple circuit of a group of units must be a factor of ( sec. 203 ) ; so that if be a prime number , there is only one form of group ; for one simple circuit must contain all the units , and this being given all the other circuits are given ; they are all of period 230 .
In a group if the pairs of one circuit are distinguished from those of another , but both circuits are of the same period , they may be said to be .
The palrs of such circuits may also be said to be similar .
231 .
If , . . . .
, , . . . .
be units of a group , and if , . .
constitute a simple circuit in the order given , and if the pairs , .
be all undistinguished each other , then the units , . . .
const tute a similar circuit in the order given .
232 .
In the case of a group the terminals of any chain constitute a pair of a definite simple network , i.e. , we cannot have , and 233 .
We may use the symbol to represent the network ( ) where 234 .
The network of which the two terminals of a chain of pairs are a component pair , may be termed the procluct of the networks of which the pairs composing the chain are components ; the order of the networks in the product being of course the same as the order of their component pairs in the chain .
235 .
A pure complete network furnishes chains connecting every pair of units of a group , and thus if , 8 , . . .
be unified pairs of the simple networks of such a network , we can with them make chains \amp ; c. , whose terminals const tute pairs of every simple network of the group .
A group is accordingly fully defined if one of its pure complete networks is given .
236 .
We have , whatever pair may be , ( sec. 200 ) , and if be a symmetncal pair , and thus 237 .
If , the two networks ( a ) and ( ) may be said to be commutat ve .
238 .
If a simple network is commutative with each of a number of others , commutative with every simple network which is ehclosed in the compound networ composed of the latter .
239 .
The system arrived at by compounding toget er a num her a number of independent groups in the mode described in sec. 189 is a group having component sets orm fthe MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
of each of the original groups .
We may fully define the group by a pure complete network composed of a number of pure networks that each detached portion of one is composed of units constituting a group of the form of one of the original groups .
Groups containing from one to twelve Units , 240 .
In the following twelve sections I shall denote the number of units in a group by will denote the number of forms of groups for any value of the graphical represe1ltation ; the tabular ; and ( S ) will be the symbol described in section 204 .
Where is greater than umity , the symbols , and ( S ) will have numerical suffixes corresponding to the different forms .
In the graphical representation , only such lines will be drawn as are necessary to completely define the group , and in some cases alternative representations will be given , obtained by taking different circuits .
The values of taken are from 1 to 12 inclusive .
241 .
If we have 242 .
If we have 243 .
If we have 244 .
If we have Fig. 30 .
Fig. 31 .
a Fig. 32 .
a a Fig. S3 .
a a a 245 .
If we have 246 .
If we have Fig. 36 .
Fig. 38 .
247 .
If we have 248 .
If we have a a a Fig. 37 .
a a a a a MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
Fig. 46 .
cdefgh hgfedc a ghabed a hdbcdef cdefgh a behefg ghabed a a a a 249 .
If we have 250 .
If we have bedefghi caefdhig abfdei gh of a fdhigbc deighcab igbcaefd ghcabfde .
a bedefgh jab defghi .
MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
Fig. 52 .
Fig. 51 .
251 .
If we have cdefghij hgfedc efghij a bajihgfe ghij bcd abcdef hgfedcbaji ij abcdefgh Fig. 53 .
252 .
If we have ( S ) tDCCCLXXXVI .
a a a ghij kabcdef h. ijk kabcdefgh kabcdefghi kabcdefghij .
bedefghijkl a defghij klab defghij kl fghijkl abed : bed klab cdef labc defg a bedefgh a cdefghi klab cdef ghij efghijk Ci ( S ) ( S ) .
A. B. KEMPE ON THE THEORY OF MATHEMATIOAL FORM .
Fig. 59 .
efghijkl jkdefhig hi ljkefd bejkl ghi ghabcklj ighb abefdljk higk ljcabfde ghefdkljabc jklfdebcaigh kljh igfdec ljkbcahi gdef .
Some Forms of 253 .
The graphical representations of the preceding forms suggest others which exist in the case of larger numbers of units .
Thus the form in which a simple circuit passes through all the units of the group appears for all the given values of ; it obviously also exists as one form of group for every other value of 254 .
So if be even , a form such as that given by the first figure of when , 6 , 8 , and 10 , and by the first figure of when , in which all the are connected by a continuous chain of non-polar lines of two kinds , clearly exists whatever value has .
255 .
If we have the form such as when in which ] pairs are symmetrical .
This form is closely related to the important one considered in Logic ( sec. 381 ) , which may be derived from it by ignoring the differences between pairs constituting a pure complete network .
A Family of Groups . .
Let us consider groups in which every circuit is of period 2 or 4 .
Some of the symmetrical pairs of these groups are component pairs of sets of four units composing simple circuits of period 4 , say are ( a in fig. 43 is a diagonal pair ) ; while some are not diagonal pairs of any simple circuit .
We might study these groups generally , but I propose here to restrict myself to groups in which all diagonal pairs undistinguished from each other , so that they all belong to the same one-way simple network , say the network .
The investigation of sections 256-269 was suggested by ths late Professor CLIFFORD'S paper ' ' GRASSMAN'S Extensive Algebra in the ' American Journal of Mathematics , ' vol. , pp. 350-358 .
257 .
Every unit is accompanied by one unit , which makes with it a diagonal pair .
We may term the companion of .
We have 258 .
Let each pair of the diagonal network be represented by a , then if be an unsymmetrical pair of any simple network of unsymmetrical pairs , we have Also if be a symmetrical pair we have 2.59 .
The network is commutative with all the others of the group .
For if be any unsymmetrical pair we have , and therefore ; and if be any symmetrical pair , either there is no unsymmetrical pair in the group , every circuit bein of period 2 , in which case every pair is commutative with every other , or else there are unsymmetrical pairs .
\amp ; c. , in which case let then 260 .
If be any pairs , or .
For let , and .
then or But or 1 , therefore or 1 .
261 .
We have or aS or according as i.e. , if we start , from any given unit and proceed along a chain composed of pairs belonging to a given collection of networks , and then starting from the same un proceed along a chain composed of pairs of the same kind but in a different or er , der we shall either arrive at the same unit as in the previous case or at compan on . .
If we substitute for the symbol , so that we have , if be a symmetrical pair , ; if unsymmetrical , .
Also , if and be any two ) airs , we have or MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
263 .
The group ( No. 5 of sec. 248 ; see fig. ) of 8 units given by the following table :\mdash ; Fig. is one of the species we are considering .
The laws of multiplication of the component simple networks ( sec. 234 ) , viz. of ( ux ) , ( ux ' ) , ( uy ) , ( uy ' ) , ( uz ) , ( uz ' ) , ( uu ) , ( uu ' ) , ( sec. 201 ) are the same as those of the quaternion expressions For example , we have and , and thus , just as we have , so we have ( ux)(uy)(uz ) .
I shall accordingly term the group a quaternion group .
264 .
A double network of which the detached portions ( sec. 206 ) are each such as that shown in fig. may be termed a pure quaternion network .
The two component networks are non-commutative ; each may be termed the conjugate of the other , and we may represent the conjugate of by .
265 .
I proceed to show that every group of the family now under consideration may be defined by a pure complete network ( sec. 235 ) consisting of pairs of conjugate networks ( i.e. distinct quaternion networks ) , and other simple networks having no conjugates in the network , ( simple in all ) ; the simple networks being such that each two are commutative unless they be a conjugate pair .
266 .
Consider a pure compound network consisting of distinct pure quaternion networks .
If be any chain of the pairs composing the simple networks , each sort of pair entering only once , if at all , into the chain , we have or i.e. , non-commutative or commutative with , according as the chain contains or does not contain a pair .
Now if be any simple network which is not one of the win be commutative with some of the , and non-commutative with others .
Let be a chain containing one pair from each of the conjugates of those of the networks with which is non-commutative , then the nple network ( Nv ) ( sec. 233 ) MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORlf .
is commutative with each of the .
For if ( a ) one of the is non-commutative with , the chain contains the pair , and thus ( a ) is non-commutative with ( N ) also , and therefore commutative with ; and if ( a ) is commutative with , does not contain , and thus ( a ) is commutative with ( N ) also , and therefore also with .
267 .
Suppose now we have any pure complete network consisting of networks such as the of the preceding sections , and also of other simple networks ( ) , ( ) , ; we can substitute for the latter the simple networks , , where the ( L ) , ( M ) , ( N ) . . .
are networks such as the ( N ) of the last section , and we shall still have a pure complete network , and it will be such that the networks other than the are commutative with the .
Now these networks other than the are either all commutative with each other , or else there are two at least which are not ; in the latter case we may add such two to the and get such as the ; we may then as before substitute for the remainder simple networks which are all commutative with the , and may repeat the process continually until the remaining networks other than the conjugate pairs are all commutative with each other .
268 .
Any group of the family we are considering is such that all the simple networks are commutative , or that two at least are not ; in the latter case , if we take the two as conjugates , and take in others so as to constitute a pure complete network , we can proceed to deal with this in the mode we have just indicated .
Thus in every case we can obtain a pure complete network containing lots of two conjugate networks , and one lot of commutative networks , the networks of each of the lots being commutative with those of the others .
Either or may vanish .
269 .
If we make and we have a group of 16 units given by the following table :\mdash ; xyz uz'y X ' XU U'XyZ XU xu X ' xu ux ' yx ' xu ux ' MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
The component simple networks of this group , .
:\mdash ; ( ux ) , , ( uy ) , ( uy ' ) , , ( uz ' ) , ( uu ) , ( uu ' ) , , are subject to the same laws of multiplication as the bi-quaternion expressions , 1 , Groups .
270 .
We may have a set such that the unified component .-ads constitute a group .
We may call the set an -adic group .
If the set is an ordinary group .
271 .
In an -adic group any r-ad may be made to correspond to any other r-ad , but the correspondence of two -ads completely determines a self-correspondence of the set .
Thus if the form of the set be known , any self-correspondence is fully represented by a two-row -column constituent of the table .
272 .
In an -adic group of units the tabular representation has rows .
For all -ads are undistinguished , and there are of these , and while any one remains identically-correspondent the remaining -ad has self-correspondences characteristic of agroup , i.e. , has self-correspondences .
Thus there are in all self-correspondences of the group including the identical correspondence .
273 .
A correspondence of two undistinguished aspects of an -adic group has not more than foci ( sec. 122 ) .
274 .
Let be a system of which the units are the whole collection of points lying on a straight line , viz. , , . . .
Any aspect abed . .
of is what is usually termed the " " range abed . .
For the range abed . .
we can by a homographic transformation substitute another range of the same points , .
, i.e. , another aspect of S. Employing all the various homographic transformations we get a set of aspects of S. Now , if we assume that this set is a single system , i. e. , that all aspects of derived from . .
by homographic transformations are undistinguished from abed . . .
and each other , but are distinguished from all aspects which cannot be derived from .
by such transformations ; then will be a triadic group .
For we can by a proper homographic transformation substitute for any three units of any other three units of , but when this substitution is made , every other unit of has a definite unit of substituted for it .
MR. A. B. KEMPE ON THE THEORY OF HEMATICAb F Substitutions .
275 .
Each row of the tabular representation of a system is derived from each of the others by definite substitutions .
Instead of writing down the valious rows , we may give one only , and state the law or laws according to which the other rows are derived .
Thus the form of a system , .
is given by merely writing down the letters representing the units of the system ( since each letter appears only once , we may represent the units by the letters in place of their sorts ) , and stating the laws of substitution , in other words , the substitutions proper to the form ( sec. 105 ) .
276 .
Conversely , in considering a system of letters , or other things , admitting of certain substitutions , we are considering a system of units of a definite form .
277 .
The various arrays of lettei .
S considered when dealing with substitutions are thus aspects of a system .
278 .
When the arrays are regarded as units the substitutions will be represented by pairs ; substitutions of the same sort by undistinguished pairs ; similar substitutions by similar pairs ( sec. 230 ) .
279 .
The substitution of one aspect of a system for another , substitutes for a component collection of any form another component collection of the same form .
Algebras .
280 .
Consider the system V of units arrived at by regarding as units the pairs connecting the units , . .
of a system of units with the three distinguished units .
If we adopt the method of representation of secs .
139 and 152 the units of will be represented by\mdash ; \amp ; c. It will , however , conduce to clearness in the present instance to write them thus\mdash ; \amp ; c. The symbols of the three rows represent the units of three different systems , each of which systems is a replica of the others and of S. We may denote the three systems when regarded as units by , respectively .
281 .
Now consider the system arrived at by regarding the triads connecting the three systems as units .
This will contain units , which may be written thus\mdash ; , \amp ; c. MR. A. B. KEMPE ON THE THEORY OF MATHEblATICAL FORM .
We have colections of triads connecting in which each pair connecting and appears once in a triad , and once only , , if contains two units , only , we have such collections as or or The number of such collections in the general case is , each containing triads .
When the triads are regarded as units we get collections each containing urlits of E. 283 .
When these collections are unified we get a system A of units , which will be a multiple system .
284 .
The form of A is in part independent of that of and in part dependent , celtain components being distinguished whatever be the form of , while it depends on what the form of is whether others are distinguished or not .
285 .
When considering the multiplicity of the system , or those peculiarities of form it possesses which arise from the mode of its construction apart from any special peculiarities in the form of , we regard as a single heap system .
286 .
If be a single heap system : certain units of A are undistinguished from each other , while others are distingllished .
If be not a single heap system .
so that its self-correspondences are more restricted than they would be in the former case , some units of A originally undistinguished become distinguished , so that a unit which was one of several undistinguished units , may become unique .
Thus certain units of A have definite relations to , and are of use as auxiliaries to the latter .
287 .
A convenient mode of conceiving of the system is to regard its units as small cubes composing a big one , the axes of the cube being represented by and .
Any layer of the cube normal to , say a layer , contains small cubes senting triads ( unified ) all containing the same unit of , and similarly in the cases of and .
If be the unit of in the triads which are represented when unified by a layer , the layer may be termed the layer , and when unified represents .
We may also conveniently term a row of cubes parallel to the axis a row ; it contains cubes representing triads all of which contain the same connecting pair of and say the pair , and when unified is represented by .
The layers thus represent the system , and must be supposed to admit of substitutions among themselves characteristic of the form of ( sec. 276 ) , and similarly in the case of and .
Here the collection of units , which when unified gives a unit of , is represented by cubes * In strictness this should be ' triads connecting the systems which when are respectively but the abbreviation is convement and will not lead to misconception ( see sec. 20 ) .
CCCLXXXVI .
MR. A. B. KEMPE ON THE THEORY OF rICAL FORM .
such that each row contains one cube and one only of the collection .
If be a single heap system , the multiplicity of A will be the number of the systems into which the collections of cubes thus constituted break up .
288 .
We may represent any one of the collections which when unified is a unit of A by a square diagram .
Take , for example , the case ; we have such a diagram as the following:\mdash ; a aHere the square may be supposed to be a face of the cube normal to and being supposed to be vertical and horizontal respectively .
The left hand letters indicate whether the adjacent layers are or or \amp ; c. , respectively , the top letters whether the adjacent layers are or or \amp ; c. , respectively ; and the letters contained in the squares , each of which squares is supposed to represent a cube , i.e. , a unit of , indicate which layers the ious cubes lie in .
289 .
By regarding the pairs connecting the system and a system of units \amp ; c. , which are all distinguished from each other , as unified , we may get a number of systems , \amp ; c. , replicas of each other and of , and we may consider systems arrived at by taking one or more component systems of each of the systems \amp ; c. Similarly in the case of A. 290 .
Now let be one of the units of A arrived at by regarding as unified a collection of units of of which is one .
This colection of units of which when unified is , contains no other IJified triad of the sort , e.g. , , so that is unique with respect to ; for when is changed to and remaining unchanged , is changed to a unit arrived at by regarding a collection of units of containing as unified , which cannot be .
We may , therefore , write .
We may deal with the other triads concerned in arriving at in the same way , and get a collection of equations , .\mdash ; \amp ; c. .
\amp ; c. \amp ; c. \amp ; c. .
A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
where the letters on the right-hand side of the equations depend on the particular unit selected .
291 .
Now we may make , and similarly in the case of every other unit of and ; also we may represent the units and by two positions , and write thus , where the order of and is aterial , or merely where the order is material ; the equations then become \amp ; c. \amp ; c. If then we have two equations such as and we may write , and similarly in other cases : so that we obtain complex expressions representing the form relations which the units of hold to each other and to the units ; which expressions , however , admit of considerable simplification in certain cases .
292 .
An equation such as I shall term a primitive equation .
The whole collection of triads which are concerned in arriving at will be termed a primitive algebra , being termed a unified primitive algebra .
An equation , such as , where only one primitive algebra is involved , may be termed a complex primitive equation .
293 .
Those components of which are represented when unified by the units of A furnish us , by the application of the preceding methods , with every possible primitive algebra , associative or non-associative , commutatiye or non-commutative , \amp ; c. , and we can discuss the number of the forms , and the relations of the various algebras , by discussing in the case in which is a single heap system ( sec. 285 ) .
294 .
We may have any number of unified primitive algebras , S , . .
which may be undistinguished from each other or not .
If the unified primitive algebras are all selected from one system ( sec. 289 ) , they may be said to be algebras of the same operation , if from different systems , they may be said to be algebras of different operations , e.g. , multiplication , addition , \amp ; c. 295 .
We may deal with each of these algebras as with and we may also deal with them in combination and obtain equations such as , giving complicated expressions for the units of , which will represent the form relations they hold to each other and to the units \amp ; c. * We might regard the primitive equations as units , and apply the term ' unified primitive algebra not to , but to the unit arrived at by regarding as unified the collection of ullified primitive equations derived from the triads concerned in arl.iving at ti .
Where two or more primitive algebras are considered , we may the unit arrived at by regarding them as unified as a unified compound tgebra .
297 .
The discussion of form is very generally carried on by the help of these auxiliary algebras .
In some cases the units , . .
of the base system undec consideration are regarded as unified pairs , . . .
where X an primitive or compound , and , . .
are units constituting a single heap which have the correspondences characteristic of the system , . . .
as long as X remains identically correspondent .
But val'ious devices exist which must be examined in each case .
298 .
Where only one unified primitive algebra is considered , we may regard the unit as expressed by the brackets , and write an equation simply as , or we may omit the brackets and write the equation thus 299 .
If in an algebra every equation is accompanied by an equation the algebra is said to be commutative .
300 .
If in an algebra any three equations such as are accompanied by the equation the algebra is said to be associative .
Since either expression may be written ( abc ) , and no ambiguity arises .
301 .
If and two algebras are such that the equations are accompanied by the equations i. e. , if and then the algebra is said to be distributive with respect to the algebra We may take to be , and to be ; we then have 1[R .
A. B. KEMPE ON THE THEORY OF hlATHEbfATICAL FORIL 302 .
An algebra may be , i.e. , such that .
( See sec. 349 .
) 303 .
An important class of primitive algebras is that in which or except in the case of one , which such that , and also such that whatever unit of may be and .
Here every unit is accompanied by another such that .
Ordinary addition and multiplication two such algebras .
In the former is is ; in the latter is is 1 , and is 304 .
The algebras which are such that in the case of each unit we have ( aa ) also of considerable importance .
We have such an algebra in the case of ordinary logic ( see .
360 ) .
305 .
We may in ordinary parlance speak of three units , composing the triad giving rise to a primitive equation , as if they were units , of , without any ambiguity arising .
The unit when dealt with as a multiplier will be , and so on ; but it not be necessary to be continually pointing this out .
We may thus speak of products when there is really only one 306 .
If in an system of units every pair is accompanied by one or more units unique with respect to the pair , we may select one of these accompanying units the case of each pair and term it the product of the pair .
Where there is only one such accompanying unit in the case of each pair , this will be the product , but if there be several such , any one of these be chosen as the product .
The most natural mode of selection in the case of undistinguished pairs is to select products which will with the pairs make distinguished triads .
( Secs .
343 , 349 .
) 307 .
In the case of distinguished pairs we cannot of do this , but we may in some cases choose products so that the resulting triads will all be of the same form , this not being so if other products be chosen ; or we may choose the products , so that by ignoring some difference the triads all become undistinguished from each other , the products still being unique with respect to the pairs .
308 .
In cases in which some pairs are not accompanied by a unit unique with respect to them , or even in other cases , we may add to a single system of one unit , and call this the product of the pairs ; being also considered as the product of each unit of when multiplied by or into Z. ( Sec. 319 , and cf. sec. 343 .
) 309 .
In certain cases the products considered may not be unique with resp the multiplier and multiplicand , but with respect to a collection containing them and certain other units , which remain the same in the case of every multiplier and multiplicand of the system , and may therefore be termed constants .
An instructive example of this is furnished by a system of collinear points .
If , be five points of such a system , there is one point , and one only , which is that the six points lie respectively on the six lines PQ , , PS , passing rough four coplanar points : i. e. , the unit is unique with respect to the collection .
Now let , be three constants in the foregoing sense .
Then we may call the product of and .
The primitive algebra which thus arises may readily be shown to be associative and commutative .
Again , there is one point , and one only , which is such that the points lie respectively on the six lines , NO , , MO , , MN , passing through four coplanar points : i.e. , the unit is unique with respect to the collection .
Here , if and constants , may be regarded as a second species of product of and , and may be termed the surn of and .
The primitive algebra thus arising is also associative and commutative .
The first of these two primitive algebras is distributive as regards the second .
In fact the compound algebra composed of the two primitive algebras is of the same kind as the ordinary algebra of quantity , the units , corresponding respectively to the , and 1 of such algebra .
* 310 .
We may speak of the collection consisting of the products of the pairs connecting two component collections of a system as the product of those collections .
If the number of units in the multiplier collection be , and the number of those in the multiplicand be , the number in the product will not necessarily be , for the product of a connecting pair may be the same unit as the product of another connecting pair .
We may speak of the collection consisting of the products of all the pairs of a collection . .
together with the products ( aa ) , ( bb ) , ( cc ) . . .
as the square of the collection .
In the same way we may have cubes , \amp ; c. , of a collection .
311 .
If we substitute for the system of sec. 191 , we may regard products and as two different units ( sec. 192 ) , and if this be done in the case of each product , we shall have units in the product .
312 .
We may regard the components of or as units , and then the products will be units , the products of pairs of units .
Quadrates .
313 .
Let , be two single heap systems , replicas of each other .
Let be units of . .
be corresponding units of .
Considering the connecting and ; these may be written a aper by the writor ' On an siou of Ordinary Algebr\ldquo ; in the Messenger of Mathematics .
Vol. XV .
New Series , p. 188 .
MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
\amp ; c. \amp ; c. , \amp ; c. \amp ; c. If it be understood that we are throughout with pairs connecting we may write any aspect thus , the position of a letter in the pair taking the place of the unit represented by the subscript number .
We then have under consideration the aspects .
( X ) These will be in number , where is the number of units each of the systems .
The aspects aa , bb , , . . .
are undistinguished from each other , but are distinguished from all the other aspects ab , , bc .
which again are all undistinguished from each other .
* 314 .
We may regard the aspects as unified , and shall then get a double system composed of the system of unified aspects ( aa ) , ( bb ) , ( cc ) .
, and the system of unified aspects ( ab ) , , ( bc ) .
315 .
The mode of arranging the aspects adopted at X in sec. 313 suggests a mode of clothing the units of which is very convenient for purposes of description , and for emphasizing the peculiarities of , that of regarding the units as arranged in rows and columns , the order of which is to be disregarded , so that the rows as units form a single heap system , as also the columns .
316 .
In we have a system of the species considered , in which ; the asterisks represent units of , the dots those of * * * * Fig. 60 .
317 .
The pairs of form a system of multiplicity 13 ; viz. , we have the sorts of pairs given by the following diagrams , where in the last seven cases the top unit in the first column , and the bottom in the second are those composing the pair considered .
( 1 ) * , ( 2 ) * , ( 3 ) ( 4 ) * ( 5 ) * , ( 6 ) .
( 7 ) , ( 8 ) * ( 9 ) * , ( 10 ) ( 11 ) * , ( 12 ) * , ( 13 ) * If we confine ourselves to the latter system of aspects , we might regard them as aspects of either of the systems and ; but the aspects aa , bb , .
must be regarded as connecting pairs of the two systems and We should obtain a system of the form as if we took , and the unified pairs of .
Hore takes the place of , and the unified pairs of take the place of The pairs ( 1 ) , ( 5 ) , ( 10 ) , and ( 12 ) will be termed joined pairs ; each is such that the column in which the first unit lies and the row in which the second unit lies intersect at an asterisk .
In the literal mode of representation they are such as ( 1 ) , ( 5 ) ( ba)(aa ) , ( 10 ) , ( 12 ) ( ab)(bc ) , , such that the last letter in the first bracket is of the same sort as the first letter in the last bracket .
We may add to these the pairs such as ( aa)(aa ) which are really connecting pairs of two systems and .
The pairs ( 2 ) , ( 3 ) , ( 4 ) , ( 6 ) , ( 7 ) , ( 8 ) , ( 11 ) , may be called unjoined pairs .
318 .
Each joined pair is accompanied by a unit unique with respect to it , represented by the dot or asterisk in the same row as the dot or asterisk representing the first unit of the pair , and in the same column as the dot or asterisk representing the second unit of the pair .
This unit may be taken as the product of the two units composing the pair .
In the literal mode of representation we shall have where , or any two of them may be letters of the same sort .
319 .
The product of two unjoined pairs is taken to be a unit or , and if ( ab ) be any unit of , we take ( sec. 308 ) .
320 .
The units ( ab ) , , ( bc ) , . .
, when thus dealt with as multipliers .
multiplixands , and products , the products being those specified in the preceding sections , may be called quadrates .
321 .
We may consider products of component collections of or ( sec. 311 ) ; and we may regard these as unified , and thus arrive at various primitive algebras .
322 .
These quadrate algebras are all associative .
323 .
It will be convenient in many cases to write a quadrate ( ab ) thus , assimilating it to the ordinary algebraic fraction , and a collection of quadrates ( see .
321 ) thus , 324 .
Quadrates have been arrived at by considering unified pairs in the case of singleheap systems .
We might consider unified pairs in the case of systems of any other form , but in such case would not have the characteristic form which we .
have been considering , and we should not get products .
Thus if we considered discrete heaps , all the unified pairs would be distinguished from each other .
We , of course , can and do consider , unified pairs of any system , and can take their products as if they were See PIERCE on " " Linear Associative Algebras in the 'American Journal of Matbematics , ' vol. MR. A. B. KEMPE ON THEORY OF MATHEMATICAL FORM .
quadrates ; but in such case we ignore differences , and do not deal with , but with a single heap system of the same number of units .
The special form of has its effect as the selector of special quadrates , or collections of quadrates ( sec. 321 ) .
325 .
rfhus if be a group the unified pairs divide up into systems , and we naturally select these .
The product of two such systems of unified pairs is another such system .
The unified systems ( which are represented by the sorts of the pairs ) give us a primitive associative algebra .
The equations we obtain are precisely the same as those we should get if we considered chains in the manner indicated in section 234 .
In both instances the equations exhibit relations existing between sorts of letters , which in both instances represent simple networks .
In the one instance we arrive at these relations by considering individual component pairs of simple networks , in the other by considering the whole systems of pairs composing those networks .
326 .
The unified aspects of any system are a group .
The unified pairs of these are substitutions by which any aspect of is substituted for another .
The product of any two of these is a substitution proper to ( sec. 105 ) , viz. , that which results from operating with the multiplier and multiplicand in succession .
327 .
A substitution proper to is represented by a two-row constituent of the tabular representation of S. Each column of this constituent may be regarded as a quadrate , the whole constituent being thus regarded as a collection of quadrates .
Thus a substitution may be represented either by a collection of quadrates obtained by taking pairs of , or by a collection of quadrates obtained by taking pairs of the group of which the units are unified aspects of S. Isolated tisfied Collections .
328 .
A collection of units which is such that each unit of the collection is unique with respect to the residue of the collection may be said to be isolated .
Each unit of the collection may be termed the residual of the rest .
329 .
The residual of a pair may be written ( ab ) ( sec. 139 ) , and similarly in the case of triads , \amp ; c. 330 .
We may graphically represent an isolated triad as in fig. 61 , where ' is an Fig. 61 .
auxiliary graphical unit ; a non-isolated triad being represented as in fig. 62 , without links or auxiliary units .
same mode of representation may also be applied in the case of isolated -ads where Mli .
A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
Fig. 62 .
331 .
Where we are residuals of collections of units , if a collection of units is such that every component collection of units has one and only one residual , which residual is also a component of the collection , then may be said to be satisfied .
Somve Systems.\mdash ; Family No. 1 .
332 .
There are certain systems in every component colection of two units has one residual and one only .
One such system is met with in the case of the points of intersection of a plane cubic with coplanar straight lines ; these make isolated triads of collinear points such that if are isolated triads , is one also ; i.e. , we have ( de ) ( be ) .
333 .
In another which I shall now consider the law of distribution of the isolated triads is such that if , be a non-isolated triad ; and if be the residual of of of ; then is an isolated triad ; i.e. , we have .
334 .
It is an immediate consequence of the law of distribution of sec. 333 , that if be the residual of ( fig. 63 ) , it is the residual of and .
Thus we arrive at a collection of 7 units ( comprising the isolated triads ; Fig. 64 .
, which contains the residual .each of its component pairs and is therefore satisfied .
The whole collection is shown fig. 64 . .
A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
335 .
In a system of this species , if we add a unit to a satisfied collection of units , we must also add others , getting in all , before we can get a satisfied collection .
This may be thus shown : , . .
be the units of , and let a unit be added to , and let , . .
be the residuals of ; ; . .
respectively ; then , . .
must all be different units ; for if and were identical we should have both , and , isolated triads .
If then a unit be added to others , . .
must also be added , making in all .
This collection of units be satisfied , so that we need not add any more units .
To prove this we have ] to show that the residual of , any two added units is a unit of the enlarged collection .
Let the residual of , which will be a unit of be , then we have the non-satisfied triad , and the residuals of its pairs are , and which by sec. 3:33 form a satisfied , or isolated triad , for they are the same thing , so that the residual of , is 336 .
Now every non-satisfied triad is a component of a satisfied collection of seven units ; and by the preceding section if there be another added we get a collection of 15 units , if again another be added one of 31 units , and generally a satisfied collection consists of units .
Every system of the species considered is of course satisfied , and therefore contains units .
337 .
If to any component collection of one of these systems we add such residuals of the different component pairs as are not already in , and repeat the on the enlarged collection we geb a definite satisfied component .
The componenb which consists of all the added units may be termed the complement of C. The complement of any pair is their residual .
If be itself satisfied there will be no complement .
It does not follow if be the complement of , that is the complement of , though the complement of must be a component of C. Thus the complement of , in fig. 64 is , but the complement of is 338 .
Consider a system of the family containing units .
Take any pair of , add to it any unit of not the residual , add to the resulting triad any unit of not in its complement , add to the resulting tetrad any unit of not in its complement , and so on .
Proceeding in this way we get an n-ad , one of several , such that its complement comprises an the remaining units of S. The components of three , four , \amp ; c. , units of the n-ad may be termed principal triads , tetrads , \amp ; c. , of S. A principal n-ad has the same self-correspondences as a single heap of units , 339 .
Let be a principal component of a system , substitute for their residual , for and their residual S , for S and their residual , and so on until we get the pair , for which substitute their residual , which will be unique with respect to .
We should arrive at if we took the units of the component in any other order .
To show this it is on necessary to prove it for an interchange of any two units , for by a succession of such interchanges we can pass to any fresh order .
Let us then interchange any two units and , and let be the residual of , then we have to show that is the residual of .
This MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FOBM .
follows at once from the law of sec. 333 , for is a non-isolated triad and S , , and are isolated ; thus is isolated and is the residual of We should arrive at if we were at any .
to substitute the residual of any two units of the reduced component , instead of always taking the last substituted residual as one of the two ; e.g. , after substituting S for , we might substitute the residual of for .
The ullit is clearly a component of the complement of .
It is obvious that if we started with the component we should arrive at the unit , and generally any other unit of would be arrived at from the rest .
In fact the whole collection is isolated , and each unit is thus a residual of the remainder , the only one that there is .
340 .
The process of the preceding section may be applied to any component whatever of , with this modification in the case of non-principal components , viz. , that where the residual of any pair is included in the component so that we cannot add it to the latter in substitution for the pair , in such case the whole triad must be removed from the component .
This amounts to removing any isolated r-ad which is contained in the component .
If in the result one unit remains , it will be the only residual ; if no unit remains the component is isolated .
341 .
We may classify the units of relatively to a principal component of units thus , viz. , we have ( 1 ) The principal component of . . .
( 2 ) The residuals of the pairs of these ( 3 ) , , , , triads ( 4 ) , , , tetrads units in number , , . . . . . . . . . . . . . . . . .
( r ) , , , , -ads , , ( n ) The residual of the ilnits 1 \ldquo ; Total units .
342 .
If we represent the units of the principal component ( 1 ) by the letters . . .
, we may represent every other unit of as a residual of two or more of these in the manner indicated in sec. 829 .
The order of the letters in the brackets will be immaterial .
The symbol representing the residual of any two units of will be arrived at by aggregating the letters contained in one or other of the two symbols MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
representing the two units , but not in both ; e.g. , the product of and ( cde ) will be ( abe ) .
343 .
If we call the residual the product of and , and add to a system of one unit 1 such that whatever unit of may be ( cf. sec. 308 ) , we have an algebra subject to the assoative law , viz. , we have Some -Triad Systems.\mdash ; Family 344 .
Another distribution of isolated triads which may exist is this : be an isolated triad , and if are also isolated triads , then is an isolated triad .
Thus .
345 .
Here , if . .
be any satisfied component of the system , and a unit be added , the residua]s ( oa ) , ( ob ) , ( oc ) .
will also compose a satisfied collection , the triad ( oa ) , ( ob ) , ( oc ) , being isolated or not according as is an isolated triad , or not .
346 .
The various residuals or , ( .
are all distinct units , no two of these symbols denoting the same unit ; and they constitute a satisfied collection , the triad , , being isolated or not as ( ob ) , ( oc ) , ( od ) is isolated or not , i. e. , according as , is isolated or not ; and similarly in the case of the residuals , 347 .
We have ( sec. 344 ) , and thus the united collections ( oa ) , ( ob ) , ( oc ) constitute a satisfied collection .
If the number of units in the collection , .
is , the total number in the satisfied collection arrived at by adding is 348 .
Thus the total number of units in any satisfied collection of the species now under consideration is 349 .
If we call the residual ( ab ) the product of and , and write it for simplicity , we see that , so that the associative law does not hold , as in the preceding systems .
But we have , so that the algebra is self-distributive .
We can have , as in the case of the preceding family of systems , Geometry in General .
350 .
In most geometrical investigations the units constitute a system of a high order of multiplicity ; we have points , straight lines , conics , cubics , \amp ; c. , unified collections of two , three , or more of these , operators such as \amp ; \amp ; c. , \amp ; c It will , however , be sufficient for the purpose of illustration to refer briefly to some comparatively simple systems .
35l .
It is to be understood that points and the lin at are not regarded as distinguished from other points and lines , and consequently parallelism is looked upon merely as intersection on a particular line .
System of Points Straight Lines .
352 .
In a general consideration of coplanar points and straight lines the points compose a single system , being undistinguished from each other , and the same thing is the case w]traight 1ines.onnecting pair o-drawn in the other case .
354 .
The number of points on each line is infinite , and an infinite number of lines pass through each point .
Every two points lie on one line and one only , so that if and are any two points , the line which is linked to both is unique with respect to ( fig. 65 ) .
Pairs of points are accordingly all undistinguished from each other .
Fig. 65 .
Fig. 6 355 .
Every two lines pass through one common point , and one only ; so that if there is a point linked to two lines and is unique with respect to PQ ( fig. 66 ) .
356 .
Every form which component collections of lines and points possess is possessed by component collections of points and lines ; the two component single systems being of precisely the same form .
357 .
We may take as fundamental laws defining the distribution of the links , and therefore the form of the double system the two well-known theorems : MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORhl .
( 1 .
) If the three points in each of the nine triads of points ; ; ; ; ; , fl ; ; ; are collinear , the points of the triad are also collinear ( fig. 67 ) .
( 2 .
) If the three points in each of the eight triads of points ; ; ; ; ; ; are collinear , the points in the triad are also collinear ( fig. 68 ) .
Fig. 67 .
Fig. 68 .
Coplanar , Lines , Conics .
358 .
In the case of the treble system composed of coplanar points , straight lines , and conics , each line touches , cuts , or does not touch or cut each conic , and each point lies on , in , or outside each conic ; we thus have three sorts of connecting pairs in each case .
We may conveniently connect the graphical unit representing a conic by links the graphical units representing points lyin on the conic , and to the graphical units representing lines touching the conic ; no being drawn in the other oeses ; so that the mdinked pairs are a double system .
359 .
If be any point , any conic , of the system , there is one line and one only of the system which is unique with respect to the pair , viz. , the polar with respect to C. Likewise is the only point unique with respect to , C. Logic .
360 .
It will be convenient to approach the consideration of Formal Logic from the standpoint of the algebraist .
Classes are denoted by terms , . .
The class which just contains , and the class which is just contained in , all the classes , are denoted respectively by the sum , and the product ABC of the terms denoting the several classes .
Whatever class A may be , we have , and .
Addition and multiplication are each commutative and ) are also distributive as regards each other , so that we have If , which nplies also , then A is contained in B. 361 .
Two classes and are considered which are such that whatever class may be , we have , and therefore also , and therefore also ZA so that contains every class , and is contained in every class .
is the class called the " " universe is that called the " " non-existent\ldquo ; class .
The class need not be taken as the actual universe of thought , but as that class which contains classes considered in an investigation .
So may be taken to be that class which in any investigation is contained in all others considered , and is ignored , or regarded as having the quality of " " non-existence.\ldquo ; 362 .
Every class A is accompanied by a class which may be called its obverse , and is such that AA ' so that A is the obverse of is " " A and are obverses of each other .
We have ( ABCD 363 .
Now , consider a collection of classes , . .
such that the product of every two is a class ; i.e. , such that no two classes of the collection have any part in common except that which is common to all the classes , and may therefore , when our attention is confined to the classes of the collection and their aggregate regarded as immaterial , and therefore be treated as non-existent .
Such a collection may be said to consist of separated classes .
Taking these , their sums two , three , four , \amp ; c. together , and the class , we have a collection of classes , which may be called a full set , and may be said to be derived from the collection of separated classes .
* An the classes of the set are contained in a class of the set which is the sum of the separated classes from which the full set is derived , and may be called the universe of the set .
The class may be called the zero of the set .
The whole system of classes involved in any inquiry is a system of classes where is the number of separated classes , the product of each two of these being the non-existent class Z. In discussing a system of classes we have under consideration a number of full sets with their accompanying universes and zero classes .
364 .
If be any class of a full set of which ?
is the universe and the zero class , we have and and more general meaning is given to the expression ' fnll set\ldquo ; in sec. 881 , and .
it must not therefore assumed that conversely every set can be derived in the manner just indicated .
MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
and is accompanied by a class such that Also we have where , . .
are any classes of the full set .
Thus the relations of the component classes of a full set are similar to those of the component classes of a whole system .
The class may be called the " " obverse of the full set It is not of course the true obverse of , i.e. , is not , unless the full set is the whole system .
365 .
Any collection of classes , .
which are such that they and their various sums and products are all different classes may be said to be a collection of unrestricted classes .
The minimum full set of which a collection of unrestricted classes is a component contains classes ; for the separated classes of such full set are abed . .
, . .
, a'b'cd . .
, a'bc 'd . .
, . .
, .
, \amp ; c. , which are in number .
This full set may be said to be derived from the collection of unrestricted classes .
366 .
If , be any collection of unrestricted classes , we can always find a class of the derived full set , such that the classes , . .
have any desired class relations to each other .
* 367 .
Thus if we have a collection of unrestricted classes , . .
, there will be a class such that , . .
constitute a full set , and another class such that , . .
are a collection of separated classes .
368 .
Suppose we have under consideration a full set of classes .
We may , if we please , ignore the relations of inclusion , \amp ; c. , which they have to each other , and treating them as a collection of separated classes constitute a derived full set of classes containing them , their aggregates , and a zero class .
Here the units of are the products , . .
of a class and a collection , . .
of unrestricted classes .
When we ignore the relations of inclusion , \amp ; c. , which , . .
have to each other , we no longer deal with , . .
but with a collection , . .
of separated classes .
369 .
The obverse in ( sec. 364 ) of any class is , the class containing all classes of which have only the zero class in common with .
The obverse in of is , the class which contains all the separated classes , . .
; viz. , all those classes arrived at by ignoring the of inclusion , \amp ; c. .
of , .
*Thus Mr. in his ' Symbolic Logic , ' at chap .
, employs intersecting ellipses to represent unrestricted classes .
He then rogards parts of these as eliminated , and so makes the ellipses represent classes having any desired relations .
Here represents the r , lass composed of the uneliminated spaces .
MDOCCLXXXVL 370 .
When any collection , . .
of classes of a full set have certain relations of inclusion , \amp ; c. , to each other ; instead of saying that , . . .
have these relations , we may say that , . . .
have the relations " " in A class may be completely defined by stating that it is the most extensive class in which certain relations hold between other classes .
Thus the class of a full set in which and are unrestricted is fully by saying that it is the most extensive class of the set in which " " all is or in which .
Further , instead of using this definition , we may simply employ the equation without other words ; it will then assert that the class is that under consideration and will be a mark or term denoting that class .
The equation will be a term denoting the class of the set .
371 .
We might use to denote , not the class , but the class which contains all those classes arrived at by the relations of inclusion , \amp ; c. , between and the various classes contained in It is in this sense that such a mark is usually employed .
372 .
Taking as used in the last paragraph , the term , i. e. , " " some only is will represent the class arrived at by ignoring the relations of inclusion , \amp ; c. , between all those classes of which are not included in .
Here is the obverse of in the full set derived from the full set by regardin the classes of the latter as separated .
373 .
Terms such as , are called propositions .
The difference between propositions and other terms is accidental and not essential matter of exact thought ; propositions , like other terms , merely denote classes .
374 .
When we consider a full set of classes , . .
, and the related propositional classes , we consider a system of classes , viz. , that derived from the unrestricted classes , . . .
; the classes and bein other classes of S. If as is generally the case is regarded as derived from unrestricted classes , represented by single letters , . .
, the number of classes in will be , in will be , in will be 375 .
In a syllogism we have two classes called the premises , e.g. , ( 1 ) , ( 2 ) .
We consider the class which is the product of these , , i.e. , that which contains all the separated classes of , which both ( 1 ) and ( 2 ) do .
This class includes some classes , is included in others , \amp ; c. It is included in .
The syllogism indicates this fact , the class being called the conclusion .
376 .
We have then under discussion in an investigation of classes a system of units consisting of If the proposition ' all is \ldquo ; iS taken to imply the existence of its subject , we musl exclude the zero class of T. Similarly in the case of the proposition " " some only is \ldquo ; ( sec. 372 ) .
MR. A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
( 0 ) a class ( 1 ) separated classes each containing ( 2 ) classes each containing two separated classes , ( r ) classes each containing separated classes , ( m ) a class containing all the separated classes .
Each class of the collection ( r ) may be derived by adding to some class of the collection a class of ( 1 ) not contained in 377 .
We have apparently as an essential accompaniment of the idea of classes the notion of the inclusion of one class in another .
Inclusion is usually regarded as a relation between two classes .
This is not , however , really so .
The notion of inclusion essentially involves that of a chain of classes with two terminals , , a class which contains all classes , and a class which is contained in all classes , of the chain .
In dealing with a class A we really deal also with the classes and ; and in dealing with two classes A and we deal with the four classes .
When this fact is recognised , inclusion , as commonly conceived , is seen to be a relation which , as far as the processes of exact thought are concerned , is accidental ; that which is essential in it depending upon the places occupied in a system of classes by two classes relatively to two others ( sec. 388 ) .
378 .
The classes and are not , apart from accidents , distinguished from any other classes .
The system of which the units are classes is a single one .
The reason why and seem to hold exceptional positions is that when we discuss classes we consider their positions with reference to and , which , being constantly under consideration , acquire an accidental importance .
379 .
There are two distinct sorts of inclusion , , direct and indirect .
If is contained in , and in , then the inclusion of in may be said to be indirect ; if is contained in , and there is no class under consideration such that is contained in and in , then the inclusion of in may be said to be direct .
The two classes and may in the latter case be said to be djacent , in the former non-adjacent .
The relations of adjacency and non-adjacency are relations between two classes , and do not depend upon the consideration of the classes and Z. In a system of classes every class is adjacent to classes .
Thus in the case of the collections of sec. 376 every class of the collection is adjacent to classes of the collection and of the collection .
380 .
The form of a system of classes is completely defined by the division of the pairs of classes into adjacent and non-adjacent pairs .
It may therefore be graphically represented by graphical units , of which pairs representing pairs of adjacent classes are linked , and pairs representing pairs of non-adjacent classes are not linked .
381 .
A convenient mode of building up a graphical representation of such a system is the following :Suppose we have a graphical representation of any collection of units , say a .
To this we can add a precisely similar diagram , say we can repeat the diagram .
The two diagrams will correspond unit to unit in one or more ways .
Taking one of these ways only , we can ] each graphical unit of one diagram to the corresponding unit of the other , say we can connect the two diagrams .
Now draw a single graphical unit , repeat it , and connect the two diagrams ( fig. 69 ) .
Repeat the resulting diagram and connect the two diagrams ( fig. 70 ) .
Repeat the resulting diagram and connect the two diagrams ( fig. 71 ) .
Repeat the resulting diagram and connect the two diagrams .
This process can be carried indefinitely , the number of units being each time doubled , so that the number of units after processes is Fig. 69 .
Fig. 70 .
Fig. 71 .
Fig. 72 .
382 .
Each of the diagrams may be said to represent a full set of the whole system finally arrived at .
There are of course other full sets besides these in the system .
Thus in fig. 72 there are 8 full sets represented , such as that in fig. 71 ; and generally in any system of units there are full sets of units .
383 .
We can pass from one graphical unit to another along links .
We can do this by passing along a minimum number of links , or by more circuitous routes .
The number of links passed over in the case in which we pass over a minimum number may be termed the distance between the two units .
Adjacent units are at a distance 1 apart .
384 .
In the followin sections it is to be understood that when chains connecting two units are spoken of , those containing the same number of links as the two units are distant apart are referred to .
We can , of course , by ging about , make the chains much longer , but such chains are not those considered , which are only { he chains of minimum ength .
385 .
Let be any unit of a class system of units .
There are units linked to , composing a collection which may be called , every unit of which is at a distance 1 from .
We have also a colection of units other than and , each of the units of which is at a distance 2 from .
So , generally , we have .
A. B. KEMPE ON THE THEORY OF MATHEMATICAL FORM .
a collection of units , each of which is at a distance from .
When we get a single unit which is the obverse of , and is unique with respect being the only unit of the system which is so .
The units of any collection are each linked ) units of and to of , but to no other units of the system , so that none of them are linked to each other .
We see here that the relation of a class to its obverse is one not depending on the relations of and to and , though expressible in terms of such relations .
386 .
Consider now two collections of chains of links , ( 1 ) consisting of chains from to a unit , ( 2 ) consisting of chains from to a unit ( see sec. 384 ) .
There are some graphical units through which chains of both collections pass .
If we classify these according to their distances from , we find that the number in each class is always greater than 1 , except in the case of the greatest distance , when there is one only , which is therefore unique with respect to .
If the chain from to passes through , it is clear that the unit is itself In the same way , if we take any number of units , S , .
, there is a unit unique with respect to , S , which may be one of the units , S , . .
We may obtain a like unit unique with respect to , S , . .
We may term the in the former case the product of , S , . .
with respect to ; and in the latter case the product with respect to 387 .
If be , then the product of any units with reference to will be what has been spoken of in sec. 360 , simply as the product of the classes represented by those units .
The product with respect to , which will be , will in such case be that which was defined in sec. 860 as the of classes .
388 .
If a chain from to passes through two units and , first through and then through , then and represent classes which are such that contains If no chain from to passes through both and , then neither class is contained in the other .
This makes it clear that the relation of inclusion is one in essentials depending merely on the form of a system of classes , and the position two units occupy in it relatively to two others ( and ) ; or rather , as and are unique with respect to each other , to one other , viz. , either or 389 .
Suppose we have a system containing the unrestricted classes , . . .
, the obverses and , and collection of separated classes abed , a'bcd .
, ab'c 'd . . .
\amp ; c. Here has the same self-correspondences as a single heap .
Now , if we confine our attention to those self-correspondences of in which the collection of units consisting of , . .
, and their obverses is self-correspondent , the collection have those self-correspondences only which are characteristic of a full set ; and will not have any correspondences.with other collections ; also the other umits of ( except Z ) which may then be regarded as aggregates of the units of taken two , three , \amp ; c. , together will correspond only if they are aggregates of component collections of which correspond ; i.e. , they and will have correspondences as if they composed as many systems as there are types of component collections of a full set of classes .
390 .
If , so that has 16 component units , the number of these types will , so that a class system containing 16 claLsses has 397 different sorts of component collections .
* It must be noted that it is distinguished collections , and not distinguished aspects of collections , with which we are here dealing .
391 .
The number of in a system of 4 units is 5 .
In one of 8 units it , is 21 .
In both cases components containing all units of the systems are included .
Professor " " On the Types of Compound Statement involving Four Classes in the ' Proceedings of the Manchester Philosophical Society , ' vol. , Third Series .
In this paper Professor CLIFFOID denotes the classes of a system derived from four unrestricted classes , by statements or propositions .
He regards the classes , and their obverses , as distinguished from all other classes of the system , and thus regards the separated classes as constituting a full system , and consequently gets the division of all the classes into 396 sorts , not including the aggregate of all the classes .
In the view taken in this memoir , the classes should all be regarded as undistinguished from each other , the distinction raised between the classes , and others , being due to the accident of particular classes being selected to be denoted by single letters , and not to any essential differences .
This does not , of course , affect the conclusion that the number of distinct types of component collections of a system of 16 classes is , if we do not include the collection which is the aggregate of all the classes , 396 , or if we do include such collection , 397 .

