BRACERS Record Detail for 53171

To access the original letter, email the Russell Archives.

Collection code
RA3
Recent acquisition no.
422
Box no.
6.51
Source if not BR
La Chaux-de-Fonds Bib.
Recipient(s)
Couturat, Louis
Sender(s)
BR
Date
1901/01/17
Form of letter
ALS(X)
Pieces
4
BR's address code (if sender)
FH
Notes and topics

Success of Peano's method. A contradiction concerning classes.

His book (briefly). Wrote article for Encyclopaedia Britannica a year ago, but not yet published. Lechalas. French politics and Christianity.

Transcription

BR TO LOUIS COUTURAT, 17 JAN. 1901
BRACERS 53171. ALS. La Chaux-de-Fonds Bib., Suisse. Russell–Couturat 1: #73
Edited by A.-F. Schmid


Friday’s Hill,
Haslemerea.
le 17 janvier 1901.

Cher Monsieur,

J’ai bien reçu votre article sur le Congrès, que j’ai trouvé admirable.

Ce que vous dîtes dans le P. S., et ce que vous m’écrivez, répond effectivement à ce que je vous ai écrit au sujet de la langue internationale. Une analyse logique des idées ne donne pas une langue qu’on peut parler dans les gares. Mais je répondrais qu’en premier lieu la commodité des voyageurs ne me paraît pas un butb très important, et qu’en second lieu on peut voyager sans inconvénient dans des pays où on ne sait pas un mot de la langue : je l’ai fait moi-même plusieurs fois. Mais je suis d’accord qu’on devrait avoir une langue internationale pour la science, puisque les Allemands et les Russes ont malheureusement cessé d’écrire en français, comme le bon Leibniz. Et c’est cela qui m’a suggéré la remarque à propos de Peano, puisque, en ce qui concerne la mathématique, son symbolisme me paraît suffisant sous ce point de vue.

Pour ce qui concerne la valeur de son symbolisme, je ne suis pas complètement d’accord avec vous. Je le trouve, au contraire, excellent du point de vue symbolique, et je trouve que c’est tout d’abord le symbolisme de Peano qui a permis aux Italiens de faire de si beaux travaux sur la logique mathématique. J’emploiec maintenant, dans tous les problèmes de ce genre, entièrement cet algorithme, que j’ai complété par une algèbre des relations différente de celle de Peano et Schröder. J’ai trouvé (1) que l’analyse logique se facilite énormément ; (2) que les paralogismes deviennent beaucoup plus rares ; (3) que les formules et les démonstrations deviennent mille fois plus faciles à comprendre. Quand je lis Cantor, par exemple, je le traduitd toujours en formules Péanesques, quoique, avant le Congrès, je n’eusse pas lu un mot de cette école. Et du point de vue de la logique formelle, je trouve qu’on a beaucoup trop insisté sur les équations, qui n’ont guère aucune importance, et qu’on a eu tort de méconnaître la distinction entre ε et — distinction indispensable, à mon avise, à la théorie de l’infini, et même à tout ce qui s’appelle mathématique. J’ai même réussi à faire de nouvelles découvertes sur le champ de la mathématique pure, ce que je n’ai jamais réussi à faire par les anciennes méthodes. Pour ces raisons, je trouve dans le symbolisme de Peano une supériorité immense à tous ses précurseurs.

Je suis d’accord avec Burali-Forti qu’il soit faux d’affirmer pour les types d’ordre : a = b . È . a < b . È . a > b. Ce qui est plus, je soupçonne qu’on ne puisse affirmer la même chose pour les nombres cardinaux. Les arguments de Cantor à ce sujet ne sont pas concluants. Je vous serai très reconnaissant si vous voulez bien me prêter l’article de Burali-Forti, que je n’ai pas vu. — Quant à la classe des classes, si vous admettez une contradiction dans ce concept, l’infini reste toujours contradictoire, et vos travaux ainsi que ceux de Cantor n’ont pas résolu le problèmef philosophique. Car il y a un concept classe et il y a des classes. Donc classe est une classe. Or on prouve (et ceci est essentiel à la théorie de Cantor) que toute classe a un nombre cardinal. Donc il y a un nombre des classes, c’est à dire un nombre de la classe classe. Mais il n’en résulte aucune contradiction, puisque la preuve que donne Cantor que

α ε Nc . ⸧ . 2α > α

présuppose qu’il y ait au moins une classe contenue dans une classe donnéeg u (dont le nombre est α) qui n’est pas elle-même un individu de u, c’est à dire qu’on a 

ƎCls Ç v ɜ (vu . v  ~ ε u 

Si l’on met u = cls, ceci devient faux. Donc la preuve ne tient plus.

Pour les nombres réels, je veuth dire, si x est un nombre rationeli quelconque, une suite infinie de nombres tous égaux à x ne détermine jamais comme nombre réel x lui-même, mais bien le segment des rationnels plus petits que x. C’est à dire, les nombres réels ne sont pas des limites de suites de nombres rationnels, mais des segments de rationnels. Cette théorie se trouve suggérée par Peano, qui cependant ne l’adopte pas en définitif [RdM IV, je crois]. J’ai des raisons très-fortes pour l’adopter, mais elles sont trop longues pour les écrire.

Quant à M. Lechalas, l’homogénéité est un mot que j’ai trop employé autrefois, et qui me paraît ne servir à rien. Il ne veut dire que ceci : que telle proposition est vraie pour tous les points. Il y a donc une homogénéité projective et une homogénéité métrique : on peut avoir la première sans la seconde, mais je crois qu’on ne peut avoir la seconde sans la première. — Qu’il faille recourrirj à la troisième dimension en Géométrie projective, c’est bien quelque chose de très remarquable : mais puisque cela se démontre, on ne peut que le trouver curieux. Et qu’il soit possible d’avoir un espace projectif qui ne possède pas les propriétés métriques qu’il faut pour la mesure, cela est évident après un examen attentif. Le meilleur livre qui existe sur la Géométrie projective est celui de Pieri « I principii della Geometria di Posizione », Turin, 1898. Vous voyez que j’ai abandonné beaucoup de thèses que je soutenai dans mon livre. J’ai écrit un article sur la géométrie non-euclidienne il y a un an, pour l’Encyclopaedia Britannica, mais on ne l’a pas encore publié. Quand il paraît, je vous l’enverrai, si on me permet des tirages à part, ainsi qu’à M. Lechalas. Pour le moment, il est difficile de discuter sur mon livre, qui ne représente plus mes opinions actuelles. J’ai oublié d’écrire ceci sur une feuille à part, mais je vous prie d’avoir la bonté d’en communiquer au moins la porték à M. Lechalas. Je renouerai volontiers la discussion aussi tôtl que mon article sera paru.

Nous regrettons beaucoup que vous ne pouvezm nous faire une visite à Cambridge, mais nous espérons que ce plaisir pour nous ne sera que renvoyé à une autre occasion.

Que pensez-vous de la politique contemporaine en France ? J’espère que M. Waldeck-Rousseau réussira dans son conflit avec Rome, mais c’est un adversaire formidable. Je suis toujours contre l’Eglise. Quoique je préfère le Christianisme aux autres religions, je le déteste cordialement. Pourquoi êtes-vous honteux de l’amnestien ? N’était-elle pas nécessaire ?

Veuillez agréer, de ma part et de celle de ma femme, l’expression de nos sentiments bien cordiaux.

Bertrand Russell

Notre adresse jusqu’au 1er Mars sera West Lodge, Downing College, Cambridge.

 

Textual Notes

a Adresse imprimée      b [j’emploi]  c sic             d {à mon avis,}             e-n sic

 

BR TO LOUIS COUTURAT, 17 JAN. 1901
BRACERS 53171. ALS. La Chaux-de-Fonds Bib., Suisse. SLBR 1: #92
Translated and edited by N. Griffin 


Friday’s Hill1
17 01 1901

Dear Sir,

I am pleased to receive your article about the congress,2 which I found excellent. What you say regarding in the P.S. and what you write to me, effectively reply to what I have written to you about our international language. A logical analysis of ideas does not give a language which can be spoken on the street. But I would reply, in the first place, that the convenience of Travellers does not appear to me to be a very important end, and in the second place, one may travel without inconvenience in countries where one doesn’t speak a word of the language. I have done it myself several times. But I agree that we must have an international language for science, since the Germans and the Russians have unfortunately stopped writing in French, like dear Leibniz. And it is this which suggested to me the observation about Peano, since, insofar as mathematics is concerned, his symbolism strikes me as sufficient from this point of view.

As concerns the value of his symbolism, I do not agree completely with you. I find on the contrary, that it is excellent as regards to symbolism, and I find that is above all Peano’s symbolism which permits the Italians to do such good work on mathematical logic. I now use, in all problems of this kind, entirely this algorithm which I completed with an algebra of relations different from that of Peirce and Schröder.3 I have found (1) that logical analysis was made very much easier; (2) that paralogisms became much more rare; (3) that formulas and proofs became a thousand times more easy to understand. When I read Cantor, for example, I always translate it into Peanoesque formulas, although before the Congress I had not read a word of that school. And from the point of view of formal logic I found them to hardly have any importance; and we were wrong to ignore the distinction between ε et — in my opinion a distinction which is indispensable to the theory of the infinite, and even to all of what is called mathematics. I have even succeeded in making new discoveries in pure mathematics, something which I never succeeded in doing with the old methods. For these reasons I find Peano’s symbolism immensely superior to all its precursors.

I agree with Burali-Forti that it is false to affirm for order types:4 a = b . È . a < b . È . a > b. What is more, I suspect that one cannot affirm the same thing for the cardinal numbers. Cantor’s arguments on the subject are not conclusive. I would be very grateful if you would bring me the Burali-Forti article that I have not seen.5

About the class of classes, if you admit a contradiction in this concept, infinity will remain forever contradictory, and your works as well as Cantor’s have not resolved the philosophical problem. For there is a concept class and there are classes. Therefore, class is a class. Now it is proven (and this is essential to Cantor’s theory) that all classes have a cardinal number. Thus, there are a number of classes, that is, a number of the class class. But it does not result in any contradiction since Cantor’s proof gives that:6

α ε Nc . ⸧ . 2α > α  

presupposes that there is at least one class contained in a given class u (whose number is α) which is not itself a member of u, that is, one has:7

ƎCls Ç v ɜ (vu . v  ~ ε u)

If we put u = Cls, this becomes false. Thus the proof doesn’t hold. 

For the real numbers, I would say that if x is any rational number, an infinite sense of real numbers all equal to x never determines x itself as a real number, but only the segment of rationals smaller than x. That is, real numbers are not limits of sequences of rationals, but segments of rationals. This theory is found suggested by Peano, who, nevertheless, does not adopt it definitively (R d M.IV, I believe). I have very strong reasons for adopting it but they are too long to write you. 

As to Mr. Lechalas, homogeneity is a word which I have used too many times and which strikes me now as worthless. It means no more than this: that such and such a proposition is true of all points. Thus there is a projective homogeneity and a metrical homogeneity: we can have this without the second, but I do not think we can have the second without the first. But it fails to appeal to the third dimension in projective Geometry, it is certainly remarkable but since it can be proven, we cannot find it curious. And it is evident after a careful examination that it is possible to have a projective space which does not have the metrical properties needed for measurement. The best book on projective geometry is Pieri’s I prinicipi della Geometria di Posizione, Turin, 1898. You see that I have abandoned many of the theories that I maintained in my book. I wrote an article on non-Euclidean geometry a year ago for the Encyclopaedia Britannica, but it is not yet published. When it appears and if they let me have some offprints I shall send it to you as well as to Mr. Lechalas. For now, it is difficult to discuss my book, which no longer represents my true opinions. I have forgotten to write this on a separate sheet, but please be good enough to communicate it nonetheless to Mr. Lechalas. I shall willingly remember to resume the discussion as soon as my article has appeared.

We very much regret that you will not be able to visit Cambridge, but we hope to have the pleasure of returning your hospitality on another occasion.

What do you think of current politics in France? I hope that M. Waldeck Rousseau will be successful in his fight with Rome, but that is a formidable opponent.8 I’m always against the church. Though I prefer Christianity to the other religions, I thoroughly detest it. Why are you ashamed of the amnesty? Wasn’t it necessary?

My wife joins me in sending our best wishes.

Bertrand Russell

Our address until the 1st of May will be West Lodge, Downing College, Cambridge.

  • 1

    [translation] Two paragraphs on real numbers and on Lechalas are omitted.

  • 2

    article about the congress “Les Mathématiques au Congrès de philosophie”, L'Enseignement mathématique (1900), pp. 394–410. The postscript Russell refers to in the next line is an addendum to this paper about international languages.

  • 3

    Peirce and Schröder Charles Sanders Peirce (1839–1914), American logician and philosopher, and Ernst Schröder (1841–1902), German logician and mathematician. Both developed logics of relations which were based on the work of George Boole (1815–1864). They differed from Russell’s Peanoesque theory in several respects. e.g. in not distinguishing between class-inclusion and class-membership (represented in Russell’s system by ⸧ and ε respectively). Russell remarks on the importance of this distinction below.

  • 4

    I agree with Burali-Forti that it is false to affirm for order types The formula states the so-called “trichotomy law” in Peano’s notation: for arbitrary order-types, a and b, either a is identical with b or a is less than b or a is greater than b. Cantor had stated the law for cardinal numbers but had not been able to prove it. Not surprisingly, the difficult cases concerned infinite order-types and cardinals. Russell, following Peano, uses the sign “∪” for disjunction rather than “v” which is now standard.

  • 5

    Burali-Forti article that I have not seen “Una questione sui numeri transfinite”, Rendiconti del circolo matematico di Palermo (1897).

  • 6

    it does not result in any contradiction since Cantor’s proof gives that The formula essentially gives Cantor’s power-set theorem in Peano’s notation, where α is the cardinal number of the class and 2α  the cardinal number of its power-set.

  • 7

    one has: The formula which follows translates Russell’s previous sentence into Peanoese. The point of the remark which follows the formula is that if u is the class of all classes then every class that is “contained” in it (i.e. every subclass of it) is also a member of it. Before the month was out Russell would realize that the paradox was not so easily avoided.

  • 8

    M. Waldeck Rousseau … is a formidable opponent Through much of 1899 the French Republic had been in a state of turmoil centred around the Dreyfus affair. In 1894 Alfred Dreyfus, a hitherto obscure Jewish officer in the French War Ministry, had been gaoled for treason. The evidence was unconvincing and it subsequently came to light that key documents had been forged. The case polarized French political opinion: intellectuals, anti-clericals, republicans, liberals and the left supported Dreyfus, while conservatives, anti-republicans, nationalists, Roman Catholics and anti-Semites were opposed. In June 1899 Dreyfus was granted a retrial, an event which led to much agitation and disorder on the part of the anti-Dreyfusards. The disorder brought Pierre Waldeck-Rousseau (1846–1904) to power to form a “government of republican defence” intended to save the Republic from right-wing subversion. Anti-clericalism, designed to restrict the political power of religious orders, formed a key part of his strategy. Meanwhile, in September 1899, the court which retried the Dreyfus case persisted in finding him guilty despite the absence of evidence (though extenuating circumstances were now admitted and his sentence was commuted). The government, however, pardoned him and, in December, announced an amnesty for all those involved, thereby ensuring that those responsible for the original miscarriage of justice would go unpunished. The intention of the amnesty was to close the case, but the Dreyfusards were not satisfied and the affair dragged on until Dreyfus was finally rehabilitated in 1906.

Publication
SLBR 1: #92 (English trans.)
Schmid, Russell—Couturat 1: #73
Re B&R B2
Permission
Everyone
Transcription Public Access
Yes
Record no.
53171
Record created
May 29, 2014
Record last modified
Nov 10, 2025
Created/last modified by
blackwk