The philosophy of mathematics Part 10

ITS RELATIONS TO THE ORDINARY CALCULUS.

In order to make as clear as possible the philosophical character of the Calculus of Variations, I think that I should, in conclusion, briefly indicate a consideration which seems to me important, and by which I can approach it to the ordinary transcendental a.n.a.lysis in a higher degree than Lagrange seems to me to have done.[12]

[Footnote 12: I propose hereafter to develop this new consideration, in a special work upon the _Calculus of Variations_, intended to present this hyper-transcendental a.n.a.lysis in a new point of view, which I think adapted to extend its general range.]

We noticed in the preceding chapter the formation of the _calculus of partial differences_, created by D'Alembert, as having introduced into the transcendental a.n.a.lysis a new elementary idea; the notion of two kinds of increments, distinct and independent of one another, which a function of two variables may receive by virtue of the change of each variable separately. It is thus that the vertical ordinate of a surface, or any other magnitude which is referred to it, varies in two manners which are quite distinct, and which may follow the most different laws, according as we increase either the one or the other of the two horizontal co-ordinates. Now such a consideration seems to me very nearly allied, by its nature, to that which serves as the general basis of the method of variations. This last, indeed, has in reality done nothing but transfer to the independent variables themselves the peculiar conception which had been already adopted for the functions of these variables; a modification which has remarkably enlarged its use. I think, therefore, that so far as regards merely the fundamental conceptions, we may consider the calculus created by D'Alembert as having established a natural and necessary transition between the ordinary infinitesimal calculus and the calculus of variations; such a derivation of which seems to be adapted to make the general notion more clear and simple.

According to the different considerations indicated in this chapter, the method of variations presents itself as the highest degree of perfection which the a.n.a.lysis of indirect functions has yet attained. In its primitive state, this last a.n.a.lysis presented itself as a powerful general means of facilitating the mathematical study of natural phenomena, by introducing, for the expression of their laws, the consideration of auxiliary magnitudes, chosen in such a manner that their relations are necessarily more simple and more easy to obtain than those of the direct magnitudes. But the formation of these differential equations was not supposed to admit of any general and abstract rules.

Now the a.n.a.lysis of Variations, considered in the most philosophical point of view, may be regarded as essentially destined, by its nature, to bring within the reach of the calculus the actual establishment of the differential equations; for, in a great number of important and difficult questions, such is the general effect of the _varied_ equations, which, still more _indirect_ than the simple differential equations with respect to the special objects of the investigation, are also much more easy to form, and from which we may then, by invariable and complete a.n.a.lytical methods, the object of which is to eliminate the new order of auxiliary infinitesimals which have been introduced, deduce those ordinary differential equations which it would often have been impossible to establish directly. The method of variations forms, then, the most sublime part of that vast system of mathematical a.n.a.lysis, which, setting out from the most simple elements of algebra, organizes, by an uninterrupted succession of ideas, general methods more and more powerful, for the study of natural philosophy, and which, in its whole, presents the most incomparably imposing and unequivocal monument of the power of the human intellect.

We must, however, also admit that the conceptions which are habitually considered in the method of variations being, by their nature, more indirect, more general, and especially more abstract than all others, the employment of such a method exacts necessarily and continuously the highest known degree of intellectual exertion, in order never to lose sight of the precise object of the investigation, in following reasonings which offer to the mind such uncertain resting-places, and in which signs are of scarcely any a.s.sistance. We must undoubtedly attribute in a great degree to this difficulty the little real use which geometers, with the exception of Lagrange, have as yet made of such an admirable conception.

CHAPTER VI.

THE CALCULUS OF FINITE DIFFERENCES.

The different fundamental considerations indicated in the five preceding chapters const.i.tute, in reality, all the essential bases of a complete exposition of mathematical a.n.a.lysis, regarded in the philosophical point of view. Nevertheless, in order not to neglect any truly important general conception relating to this a.n.a.lysis, I think that I should here very summarily explain the veritable character of a kind of calculus which is very extended, and which, though at bottom it really belongs to ordinary a.n.a.lysis, is still regarded as being of an essentially distinct nature. I refer to the _Calculus of Finite Differences_, which will be the special subject of this chapter.

_Its general Character._ This calculus, created by Taylor, in his celebrated work ent.i.tled _Methodus Incrementorum_, consists essentially in the consideration of the finite increments which functions receive as a consequence of a.n.a.logous increments on the part of the corresponding variables. These increments or _differences_, which take the characteristic ?, to distinguish them from _differentials_, or infinitely small increments, may be in their turn regarded as new functions, and become the subject of a second similar consideration, and so on; from which results the notion of differences of various successive orders, a.n.a.logous, at least in appearance, to the consecutive orders of differentials. Such a calculus evidently presents, like the calculus of indirect functions, two general cla.s.ses of questions:

1. To determine the successive differences of all the various a.n.a.lytical functions of one or more variables, as the result of a definite manner of increase of the independent variables, which are generally supposed to augment in arithmetical progression.

2. Reciprocally, to start from these differences, or, more generally, from any equations established between them, and go back to the primitive functions themselves, or to their corresponding relations.

Hence follows the decomposition of this calculus into two distinct ones, to which are usually given the names of the _Direct_, and the _Inverse Calculus of Finite Differences_, the latter being also sometimes called the _Integral Calculus of Finite Differences_. Each of these would, also, evidently admit of a logical distribution similar to that given in the fourth chapter for the differential and the integral calculus.

_Its true Nature._ There is no doubt that Taylor thought that by such a conception he had founded a calculus of an entirely new nature, absolutely distinct from ordinary a.n.a.lysis, and more general than the calculus of Leibnitz, although resting on an a.n.a.logous consideration. It is in this way, also, that almost all geometers have viewed the a.n.a.lysis of Taylor; but Lagrange, with his usual profundity, clearly perceived that these properties belonged much more to the forms and to the notations employed by Taylor than to the substance of his theory. In fact, that which const.i.tutes the peculiar character of the a.n.a.lysis of Leibnitz, and makes of it a truly distinct and superior calculus, is the circ.u.mstance that the derived functions are in general of an entirely different nature from the primitive functions, so that they may give rise to more simple and more easily formed relations: whence result the admirable fundamental properties of the transcendental a.n.a.lysis, which have been already explained. But it is not so with the _differences_ considered by Taylor; for these differences are, by their nature, functions essentially similar to those which have produced them, a circ.u.mstance which renders them unsuitable to facilitate the establishment of equations, and prevents their leading to more general relations. Every equation of finite differences is truly, at bottom, an equation directly relating to the very magnitudes whose successive states are compared. The scaffolding of new signs, which produce an illusion respecting the true character of these equations, disguises it, however, in a very imperfect manner, since it could always be easily made apparent by replacing the _differences_ by the equivalent combinations of the primitive magnitudes, of which they are really only the abridged designations. Thus the calculus of Taylor never has offered, and never can offer, in any question of geometry or of mechanics, that powerful general aid which we have seen to result necessarily from the a.n.a.lysis of Leibnitz. Lagrange has, moreover, very clearly proven that the pretended a.n.a.logy observed between the calculus of differences and the infinitesimal calculus was radically vicious, in this way, that the formulas belonging to the former calculus can never furnish, as particular cases, those which belong to the latter, the nature of which is essentially distinct.

From these considerations I am led to think that the calculus of finite differences is, in general, improperly cla.s.sed with the transcendental a.n.a.lysis proper, that is, with the calculus of indirect functions. I consider it, on the contrary, in accordance with the views of Lagrange, to be only a very extensive and very important branch of ordinary a.n.a.lysis, that is to say, of that which I have named the calculus of direct functions, the equations which it considers being always, in spite of the notation, simple _direct_ equations.

GENERAL THEORY OF SERIES.

To sum up as briefly as possible the preceding explanation, the calculus of Taylor ought to be regarded as having constantly for its true object the general theory of _Series_, the most simple cases of which had alone been considered before that ill.u.s.trious geometer. I ought, properly, to have mentioned this important theory in treating, in the second chapter, of Algebra proper, of which it is such an extensive branch. But, in order to avoid a double reference to it, I have preferred to notice it only in the consideration of the calculus of finite differences, which, reduced to its most simple general expression, is nothing but a complete logical study of questions relating to _series_.

Every _Series_, or succession of numbers deduced from one another according to any constant law, necessarily gives rise to these two fundamental questions:

1. The law of the series being supposed known, to find the expression for its general term, so as to be able to calculate immediately any term whatever without being obliged to form successively all the preceding terms.

2. In the same circ.u.mstances, to determine the _sum_ of any number of terms of the series by means of their places, so that it can be known without the necessity of continually adding these terms together.

These two fundamental questions being considered to be resolved, it may be proposed, reciprocally, to find the law of a series from the form of its general term, or the expression of the sum. Each of these different problems has so much the more extent and difficulty, as there can be conceived a greater number of different _laws_ for the series, according to the number of preceding terms on which each term directly depends, and according to the function which expresses that dependence. We may even consider series with several variable indices, as Laplace has done in his "a.n.a.lytical Theory of Probabilities," by the a.n.a.lysis to which he has given the name of _Theory of Generating Functions_, although it is really only a new and higher branch of the calculus of finite differences or of the general theory of series.

These general views which I have indicated give only an imperfect idea of the truly infinite extent and variety of the questions to which geometers have risen by means of this single consideration of series, so simple in appearance and so limited in its origin. It necessarily presents as many different cases as the algebraic resolution of equations, considered in its whole extent; and it is, by its nature, much more complicated, so much, indeed, that it always needs this last to conduct it to a complete solution. We may, therefore, antic.i.p.ate what must still be its extreme imperfection, in spite of the successive labours of several geometers of the first order. We do not, indeed, possess as yet the complete and logical solution of any but the most simple questions of this nature.

_Its ident.i.ty with this Calculus._ It is now easy to conceive the necessary and perfect ident.i.ty, which has been already announced, between the calculus of finite differences and the theory of series considered in all its bearings. In fact, every differentiation after the manner of Taylor evidently amounts to finding the _law_ of formation of a series with one or with several variable indices, from the expression of its general term; in the same way, every a.n.a.logous integration may be regarded as having for its object the summation of a series, the general term of which would be expressed by the proposed difference. In this point of view, the various problems of the calculus of differences, direct or inverse, resolved by Taylor and his successors, have really a very great value, as treating of important questions relating to series.

But it is very doubtful if the form and the notation introduced by Taylor really give any essential facility in the solution of questions of this kind. It would be, perhaps, more advantageous for most cases, and certainly more logical, to replace the _differences_ by the terms themselves, certain combinations of which they represent. As the calculus of Taylor does not rest on a truly distinct fundamental idea, and has nothing peculiar to it but its system of signs, there could never really be any important advantage in considering it as detached from ordinary a.n.a.lysis, of which it is, in reality, only an immense branch. This consideration of _differences_, most generally useless, even if it does not cause complication, seems to me to retain the character of an epoch in which, a.n.a.lytical ideas not being sufficiently familiar to geometers, they were naturally led to prefer the special forms suitable for simple numerical comparisons.

PERIODIC OR DISCONTINUOUS FUNCTIONS.

However that may be, I must not finish this general appreciation of the calculus of finite differences without noticing a new conception to which it has given birth, and which has since acquired a great importance. It is the consideration of those periodic or discontinuous functions which preserve the same value for an infinite series of values of the corresponding variables, subjected to a certain law, and which must be necessarily added to the integrals of the equations of finite differences in order to render them sufficiently general, as simple arbitrary constants are added to all quadratures in order to complete their generality. This idea, primitively introduced by Euler, has since been the subject of extended investigation by M. Fourier, who has made new and important applications of it in his mathematical theory of heat.

APPLICATIONS OF THIS CALCULUS.

_Series._ Among the princ.i.p.al general applications which have been made of the calculus of finite differences, it would be proper to place in the first rank, as the most extended and the most important, the solution of questions relating to series; if, as has been shown, the general theory of series ought not to be considered as const.i.tuting, by its nature, the actual foundation of the calculus of Taylor.

_Interpolations._ This great cla.s.s of problems being then set aside, the most essential of the veritable applications of the a.n.a.lysis of Taylor is, undoubtedly, thus far, the general method of _interpolations_, so frequently and so usefully employed in the investigation of the empirical laws of natural phenomena. The question consists, as is well known, in intercalating between certain given numbers other intermediate numbers, subjected to the same law which we suppose to exist between the first. We can abundantly verify, in this princ.i.p.al application of the calculus of Taylor, how truly foreign and often inconvenient is the consideration of _differences_ with respect to the questions which depend on that a.n.a.lysis. Indeed, Lagrange has replaced the formulas of interpolation, deduced from the ordinary algorithm of the calculus of finite differences, by much simpler general formulas, which are now almost always preferred, and which have been found directly, without making any use of the notion of _differences_, which only complicates the question.

_Approximate Rectification, &c._ A last important cla.s.s of applications of the calculus of finite differences, which deserves to be distinguished from the preceding, consists in the eminently useful employment made of it in geometry for determining by approximation the length and the area of any curve, and in the same way the cubature of a body of any form whatever. This procedure (which may besides be conceived abstractly as depending on the same a.n.a.lytical investigation as the question of interpolation) frequently offers a valuable supplement to the entirely logical geometrical methods which often lead to integrations, which we do not yet know how to effect, or to calculations of very complicated execution.

Such are the various princ.i.p.al considerations to be noticed with respect to the calculus of finite differences. This examination completes the proposed philosophical outline of ABSTRACT MATHEMATICS.

CONCRETE MATHEMATICS will now be the subject of a similar labour. In it we shall particularly devote ourselves to examining how it has been possible (supposing the general science of the calculus to be perfect), by invariable procedures, to reduce to pure questions of a.n.a.lysis all the problems which can be presented by _Geometry_ and _Mechanics_, and thus to impress on these two fundamental bases of natural philosophy a degree of precision and especially of unity; in a word, a character of high perfection, which could be communicated to them by such a course alone.

BOOK II.

GEOMETRY.