Editor’s note
Leibniz now takes Descartes's seven rules for the collision of hard bodies (Principia II.46–52) one at a time. Only the first survives. Each of the others is shown to produce a jump in the outcome for an arbitrarily small change in the data, which the law of continuity of art. 45 forbids. The positive results he draws along the way, above all that a body striking an equal body at rest stops and hands over all its motion, are the standard results for elastic collision. Source: Gerhardt's printing (public domain), checked against the OCR; the OCR's "4" and "A" for "1" are silently corrected where the sense is certain.
Original (GP IV 376–380)
English
1
Ad artic. (46). Jam regulas motuum Cartesianas videamus. Intelligantur autem corpora dura nec ab aliis circumstantibus impedita. Reg. 1. Si directe sibi occurrant B et C aequalia et aequivelocia, ambo reflectentur ea qua venerunt celeritate. Haec prima regula de motu Cartesiana sola penitus vera est. Atque ita quidem demonstratur: cum par amborum corporum ratio sit, aut ambo pergent, adeoque se penetrabunt, quod absurdum est, aut ambo quiescent, sed ita periisset potentia, aut ambo reflectentur, et quidem pristina celeritate, quoniam si unius imminueretur celeritas, deberet ob rationum paritatem et celeritas alterius minui, sed amborum celeritate minuta virium quoque summa minueretur, quod fieri nequit.
On art. 46. Now let us look at the Cartesian rules of motion. The bodies are understood to be hard and not hindered by any surrounding bodies. Rule 1: if B and C, equal and of equal speed, meet head on, both will be reflected with the speed with which they came. This first Cartesian rule of motion is the only one that is entirely true. It is proved as follows. Since the case of the two bodies is the same, either both will go on, and so penetrate each other, which is absurd; or both will come to rest, but then power would have perished; or both will be reflected, and indeed with their former speed, since if the speed of one were diminished, the speed of the other would also have to be diminished because the reasons are equal, and if the speed of both were diminished, the sum of the forces would also be diminished, which cannot happen.
2
Ad artic. (47). Reg. 2. Si B et C celeritatibus aequalibus sibi occurrunt, sed B sit majus, tunc solum C reflectetur et B perget, ambo priore celeritate, atque ita ambo simul ibunt directione quam habuerat B. Haec regula falsa est, nec consentit praecedenti, ut patet per criterion paulo ante positum. Nam diminuta continue inaequalitate seu excessu ipsius B super C, donec in ipsam plane aequalitatem abeat, continue etiam eventus accedere debebit ad eventum aequalitatis, atque ita si ponamus B tanto excessu vincere occurrens C ut post occursum progredi pergat, necesse erit ut B paulatim imminuto etiam continue progressus ejus imminuatur, donec ad certam quandam ipsius B ad C proportionem deventum sit, ubi B plane sistatur, ac deinde continuata imminutione ipsius B vertatur in motum contrarium paulatim crescentem, quousque tandem omni inaequalitate inter B et C sublata desinatur in regulam aequalitatis, regressusque post ictum ex regula prima plane fiat progressui ante ictum aequalis. Itaque haec regula Cartesii secunda stare non potest, utcunque enim diminuas B, ut ad magnitudinem ipsius C accedat, atque ita quidem prope ut discrimen fiat parvitatis prope inexplicabilis, tamen semper longissime diversa, si ipsi credimus, manent eventa aequalitatis et inaequalitatis inter B et C, neque ad sese paulatim accedunt, B semper pergente in eandem plagam eadem celeritate quamdiu tantillo majus est ipso C, unde fit ut deinde uno ut ita dicam flatu corrigendus sit hic neglectus et magno in eventis saltu opus sit, minima licet mutatione adhuc facienda in datis, cum scilicet postremo excessus ipsius B plane cessat tantillo illo indefinite parvo prorsus detracto, atque ita ab omnimodo progressu ad omnimodum regressum transeatur intermediis infinitis casibus velut saltu transmissis, et fiet ut duo casus qui discrimen habent in hypothesibus seu datis infinite parvum seu quod minus sumi queat quovis dato, habeant tamen maximum et notabilissimum discrimen in eventis, neque ad se nisi ultimo momento transitus tota simul accedant seu simul incipiant appropinquare et desinant, appropinquationemque absolvant omnino coincidendo, quod cum ratione pugnat. Ita etiam fiet, ut regula aequalitatis seu inaequalitatis infinite parvae sub regula generali inaequalitatis comprehendi non possit. Itaque cum B et C aequalia et aequivelocia sibi occurrentia (per reg. 1.) ambo priori celeritate reflectantur, oportet aucto nonnihil B, vel (si maneat) imminuto C, mutationem quandam et in eventu fieri, ac nonnihil accedi ad eum eventum, qui obtineret maxima ipsius C imminutione, hoc est omnimoda sublatione. Nunc autem cum C nonnihil tantum infra B imminui incipiat, non aliter a casu omnimodae aequalitatis seu omnimodae reflexionis, ad casum summae inaequalitatis seu omnimodae sublationis ipsius C adeoque ad casum omnimodi progressus ipsius B paulatim accedere incipiemus, quam si reflexionem ipsius B minuere incipiamus. Aucto deinde magis magisque discrimine inter B et C, eo tandem excessus perveniet B ut plane non reflectatur, sed inter regressum et progressum velut medium atque suspensum sistatur; porro autem auctum jam plane progredietur ea qua venerat directione, etsi nunquam usque adeo augeri possit ejus magnitudo quin progrediendi celeritas occursu contrarii nonnihil retardetur, nisi cum ratio ejus ad C fit infinita, id est cum C plane evanescit aut tollitur. Haec vera corporum inaequalium sed aequali celeritate sibi occurrentium habitudo est, per omnia consentiens rationi et sibi; praecise autem determinare celeritatum residuarum quantitates non est hujus loci, nam ea res altius repeti debet et peculiarem a nobis tractationem accepit.
On art. 47. Rule 2: if B and C meet with equal speeds, but B is larger, then only C will be reflected and B will go on, both with their former speed, and so both will go together in the direction B had. This rule is false and does not agree with the preceding one, as is clear from the criterion laid down a little earlier. For if the inequality, the excess of B over C, is continually diminished until it passes into exact equality, the outcome too must continually approach the outcome of equality. So if we suppose B to exceed the oncoming C by so much that it goes on forward after the collision, then as B is gradually diminished its going forward must also be continually diminished, until a certain proportion of B to C is reached at which B is stopped altogether; and then, as B is diminished further, its motion must turn into a contrary motion that gradually grows, until at last, when all inequality between B and C has been removed, we end in the rule of equality, and by the first rule the going back after the impact becomes exactly equal to the going forward before it. This second rule of Descartes therefore cannot stand. For however much you diminish B so that it approaches the size of C, and so closely that the difference becomes of an almost inexpressible smallness, the outcomes for equality and for inequality between B and C still remain, if we believe him, as far apart as can be, and do not gradually approach each other: B always goes on in the same direction with the same speed for as long as it is the least bit larger than C. So this neglect must afterwards be made good at a single stroke, so to speak, and a great leap in the outcomes is needed while only the smallest change is still to be made in the data, namely when the excess of B finally ceases with the removal of that indefinitely small remainder. Thus one would pass from going wholly forward to going wholly back with an infinity of intermediate cases jumped over, as it were; and two cases whose difference in the hypotheses or data is infinitely small, or smaller than any given quantity, would yet have the greatest and most notable difference in their outcomes, and would approach each other only at the very last moment of the transition, all at once, beginning and ending their approach together and completing it by coinciding outright. This is contrary to reason. It would also follow that the rule for equality, that is, for an infinitely small inequality, could not be included under the general rule for inequality. Since, then, B and C, equal and of equal speed, meeting each other, are both reflected with their former speed (by rule 1), when B is somewhat increased, or C (if B stays the same) somewhat diminished, some change must also occur in the outcome, and some approach must be made to the outcome that would hold if C were diminished to the utmost, that is, removed altogether. Now when C first begins to be diminished a little below B, the only way we can begin to move gradually from the case of complete equality, complete reflection, toward the case of greatest inequality, the complete removal of C, and so toward B's going wholly forward, is by beginning to diminish the reflection of B. Then, as the difference between B and C is increased more and more, the excess of B will at last reach the point where B is not reflected at all but is stopped, suspended as it were midway between going back and going forward. Increased further, it will now go straight on in the direction from which it came, although its size can never be increased so far that its speed of going forward is not somewhat slowed by meeting the opposing body, except when its ratio to C becomes infinite, that is, when C vanishes altogether or is removed. This is the true behaviour of unequal bodies meeting each other with equal speed, agreeing in every respect with reason and with itself. To determine the exact amounts of the remaining speeds is not for this place, for that matter must be taken further back and has received a special treatment from us.
3
Ad artic. (48). Reg. 3. Si B et C sint aequalia sibique occurrant motibus contrariis inaequalibus, tunc celerius B secum abripiet tardius C et dimidia differentia celeritatum detracta celeritati ipsius B addetur celeritati ipsius C, ut ambo eant simul celeritate aequali. Haec regula non minus falsa est quam praecedens, nec minus cum rationibus quam experimentis pugnat. Nam ut nostrum Criterium applicemus, B celerius secum abripiat tardius modo dicto, ex hypothesi, et decrescat continue celeritas ipsius B, donec fiant ambae aequales, seu quod eodem redit; donec excessus celeritatis B super C fiat incomparabiliter parvus, itaque tunc ambo ferentur simul celeritate ipsius B nulla assignabili quantitate diminuta, quod est absurdum, et contra regulam primam, quae recte statuit tam in magnitudine quam in velocitate in casu perfectae paritatis vel saltem in casu ab hoc inassignabiliter differente ambo sua aut saltem inassignabiliter aucta vel diminuta celeritate reflecti. Nec fieri potest ut eventus inaequalitatis evanescentis in eventum aequalitatis non evanescat.
On art. 48. Rule 3: if B and C are equal and meet with unequal contrary motions, then the faster B will carry the slower C along with it, and half the difference of the speeds will be taken from the speed of B and added to the speed of C, so that both go together with equal speed. This rule is no less false than the preceding one, and conflicts no less with reasons than with experiments. To apply our criterion: let the faster B carry the slower along with it in the way stated, by hypothesis, and let the speed of B decrease continually until the two become equal, or, what comes to the same, until the excess of B's speed over C's becomes incomparably small. Then both will be carried along together with B's speed diminished by no assignable amount. This is absurd, and contrary to the first rule, which rightly states that in the case of perfect parity in both size and speed, or at least in a case that differs from it by an unassignable amount, both are reflected with their own speed, or with a speed increased or diminished by an unassignable amount. Nor can it be that the outcome of a vanishing inequality does not vanish into the outcome of equality.
4
Ad artic. (49). Reg. 4. Si B sit minus quam C, et B moveatur, C vero quiescat, B reflectetur qua venit celeritate, C vero manebit quiescens. Haec regula in eo vera est, quod minus semper a majore quiescente reflectitur, sed non ea qua venit celeritate, nam quo magis decrescet excessus ipsius B, eo magis decrescet repulsa, donec tandem perveniatur ad casum aequalitatis, de quo reg. sexta; et absurdum est, hypothesibus paulatim ad casum aequalitatis corporum accedentibus, eventus quoque non paulatim ad eum accedere, sed semper eosdem manere, donec postea uno ictu velut per saltum ad eum transeatur. Et sane facile quivis intelligit etiam illud a ratione alienum esse, ut hypothesis continue variata nihil variet eventum, excepto casu singulari determinato, cum contra potius in omnibus rerum exemplis variatio hypotheseos variare eventum debeat, exceptis casibus determinatis, ubi diversae forte variationes complicatae se mutuo compensant.
On art. 49. Rule 4: if B is smaller than C, and B is moving while C is at rest, B will be reflected with the speed with which it came, and C will stay at rest. This rule is true in this respect, that the smaller body is always reflected by a larger one at rest; but not with the speed with which it came. For the more the excess of C decreases, the more the rebound will decrease, until at last we arrive at the case of equality, which rule 6 deals with. And it is absurd that, while the hypotheses gradually approach the case of equal bodies, the outcomes should not also gradually approach it but always stay the same, until afterwards they pass into it at one stroke, as by a leap. Anyone can readily see that it is also contrary to reason that a continually varied hypothesis should make no variation at all in the outcome, except at one particular determinate case; whereas, on the contrary, in all examples in nature a variation of the hypothesis must vary the outcome, except in determinate cases where different variations happen to be combined so as to compensate each other.
5
Ad artic. (50). Reg. 5. Si B sit majus quam C et B moveatur, C vero quiescat, tunc B perget, et ambo simul ferentur celeritate eadem, at quantitate motus priore. Haec regula et ipsa erronea est, nam erratur in vera quantitate celeritatis cujusque definienda, dum ponitur omnia post concursum simul incedere, quod nunquam in durorum concursu contingere potest; recte tamen pronuntiatum est, omne majus, quod in quiescens incurrit, pergere post ictum. Quod autem hoc loco non possint simul moveri, etiam ex criterio nostro intelligitur. Nam B esse tantillo majus quam C, et C esse tantillo majus quam B, sunt duo casus, qui sibi admoveri possunt ad differentiam usque incomparabiliter parvam, ergo fieri non potest, ut eventa eorum tunc quoque longissimo intervallo dissideant, ita ut priore casu simul eant directione quam habuerat B, posteriore autem casu B in contrarium repellatur tota sua celeritate.
On art. 50. Rule 5: if B is larger than C, and B is moving while C is at rest, then B will go on, and both will be carried together with the same speed, but with the former quantity of motion. This rule too is wrong, for it errs in defining the true amount of each body's speed, when it supposes that everything moves on together after the collision, which can never happen when hard bodies collide. It is rightly stated, however, that every larger body that runs into one at rest goes on after the impact. That the two cannot move together here is also understood from our criterion. For B being the least bit larger than C, and C being the least bit larger than B, are two cases that can be brought together until their difference is incomparably small. So it cannot be that their outcomes are then still as far apart as can be, so that in the first case the two go together in the direction B had, and in the second B is driven back with its whole speed.
6
Ad artic. (51). Reg. 6. Si B et C sint aequalia, et B moveatur, C vero quiescat, tunc B reflectetur tribus quadrantibus velocitatis qua venit, C autem ibit directione priore ipsius B residuo velocitatis quadrante. Haec auctor; sed nescio an quicquam in hac materia excogitari possit a ratione alienius, ut satis mirari nequeam talia in mentem egregii Viri venire potuisse. Sed patiamur Cartesianos quaerere rationem dictorum Magistri; nobis sufficerit, inconsistentiam regularum demonstrare. Si B et C sint aequalia, et occurrant sibi velocitatibus aequalibus, reflectetur tam B quam C velocitate qua venit per reg. 1. Diminuta jam continue velocitate ipsius C, manente velocitate ipsius B, necesse est B minus reflecti adeoque C magis quam ante, cum quantum uni aequalium celeritatis detrahitur, tantum alteri addatur. Evanescente jam velocitate ipsius C seu quiescente C, quaeritur quantum celeritatis detractum sit reflexioni ipsius B. Regula Cartesii praesens pronuntiat tantummodo detractum esse quadrantem. Jam pergendo minuamus nonnihil C quiescentis magnitudinem, utique ex reg. praecedenti B perget; ergo per mutationem casus utcunque parvam, fit ingens mutatio in eventu, seu fit saltus, nam in eventu ex magna celeritate reflexionis B (tribus scilicet quartis pristinae celeritatis) cum C quiescens ei aequaretur, nunc C tantillum diminuto fit subito destructio omnimoda reflexionis, imo quod plus est conversio ejus in contrarium, nempe in progressum, casibus intermediis saltu transmissis, quod absurdum est. Dicendum igitur est in casu aequalitatis inter B et C, quiescente C ante ictum, quiescere B post ictum, et totam suam celeritatem transferre in C. Idque etiam colligitur ex eo quod verum inest regulis 4 et 5. Nam per reg. 4, incurrente B in quiescens C majus, utique reflectetur B; rursus per reg. 5. incurrente B in quiescens C minus, utique perget B: ergo incurrente B in quiescens C aequale, nec perget B, nec reflectetur, sed (quod inter haec medium est) quiescet, tota sua vi translata in C.
On art. 51. Rule 6: if B and C are equal, and B is moving while C is at rest, then B will be reflected with three quarters of the speed with which it came, and C will go in B's former direction with the remaining quarter of the speed. So the author. But I doubt that anything more contrary to reason could be devised on this subject, so that I cannot wonder enough that such things could enter the mind of so distinguished a man. But let us leave the Cartesians to look for the reason of their Master's sayings; it will be enough for us to show that the rules are inconsistent. If B and C are equal and meet with equal speeds, both B and C will be reflected with the speed with which they came, by rule 1. Now if the speed of C is continually diminished while that of B stays the same, B must be reflected less and C therefore more than before, since as much speed as is taken from one of two equal bodies is added to the other. When the speed of C vanishes, that is, when C is at rest, we ask how much speed has been taken from B's reflection. Descartes's present rule declares that only a quarter has been taken. Now, going further, let us diminish a little the size of the resting C: by the preceding rule B will certainly go on. So a change in the case, however small, produces a huge change in the outcome, that is, a leap. For in the outcome, from a great speed of reflection of B (three quarters of its former speed) when the resting C was equal to it, there now comes, with C diminished the least bit, a sudden and complete destruction of the reflection, and, what is more, its conversion into its contrary, namely into going forward, with the intermediate cases jumped over. This is absurd. We must say, therefore, that in the case of equality between B and C, with C at rest before the impact, B is at rest after the impact and transfers its whole speed to C. This can also be gathered from what is true in rules 4 and 5. For by rule 4, when B runs into a larger C at rest, B is certainly reflected; and by rule 5, when B runs into a smaller C at rest, B certainly goes on. Therefore when B runs into an equal C at rest, B neither goes on nor is reflected, but (what lies between these) comes to rest, its whole force transferred to C.
7
Ad artic. (52). Reg. 7. Si B et C moverentur versus eandem partem, B quidem insequens et celerius, C vero antecedens et tardius, essetque C majus, sed minor ratio ipsius C ad B quam celeritatis B ad celeritatem C, irent ambo simul directione priore ac celeritate, quae faceret eandem quae prius quantitatem motus. Sin C majore existente major esset ratio C ad B quam velocitatis B ad velocitatem C, reflectetur B celeritate qua venit, et perget C celeritate qua prius. Ita autor noster. Sed haec quam incongrua sint, facile intelligitur, nam dudum notavimus corpora dura (qualia hic ponuntur) nunquam post ictum simul incedere, quod tamen hic fit casu priore; et nihil est a ratione alienius, quam corpus B agendo in corpus C nihil in ipso immutare, et tamen plurimum ab eo pati, quod fit in casu posteriore. Pugnant ista ni fallor cum Metaphysica (ut ita dicam) naturali, quam nobis praebet lumen rationis. Etiam alia insunt prioribus contraria, nam cum C est infinite parvo majus, id est aequale ipsi B, et celeritate incomparabiliter parva antecedit, id est revera quiescit, locum habet casus prior regulae hujus septimae, adeoque sequetur ex hac regula ambo ire simul, cum tamen B ad quietem redigi, et vim totam transferre in C antea quiescens sibi aequale, ostensum sit fine dictorum ad regulam sextam. Ut alia brevitatis causa taceam non minus inconsistentia. Postremo admonendum est, oblitum esse Autorem casus intermedii, cum ratio corporum est reciproca rationis celeritatum, nec apparet, quid vi regulae sit dicendum; deberet enim eventus quoque esse in medio, et omnino in ipso confinio casus utriusque; sed casus prior et posterior regulae, etsi confinium habeant in hypothesi, non tamen confinium habent in eventu, quod iterum pugnat cum Criterio nostro. Sed etiam dissimulatus est casus, cum B est major quam C. Adjicienda jam fuisset regula 8, qua explicuisset autor, quid contingat cum duo corpora inaequalia, inaequalibus celeritatibus in contrarias partes tendentia inter se concurrunt. Distinguendum quoque erat inter incursum centralem et eccentricum, inter perpendicularem et obliquum; sed finis tandem examinis faciendus est, nec doctrinae jacenti ac deploratae ultra insistendum.
On art. 52. Rule 7: if B and C are moving in the same direction, B following and faster, C ahead and slower, and C is larger, but the ratio of C to B is less than the ratio of B's speed to C's speed, then both will go together in their former direction with a speed that makes the same quantity of motion as before. But if, C being larger, the ratio of C to B is greater than the ratio of B's speed to C's speed, B will be reflected with the speed with which it came, and C will go on with its former speed. So our author. How unfitting this is can easily be seen. We noted long ago that hard bodies (such as are supposed here) never move on together after an impact, yet that is what happens in the first case; and nothing is more contrary to reason than that body B, acting on body C, should change nothing in it and yet suffer a great deal from it, which is what happens in the second case. Unless I am mistaken, these things conflict with the natural metaphysics, so to speak, that the light of reason gives us. There are also other things in them contrary to what came before. When C is larger by an infinitely small amount, that is, equal to B, and goes ahead with an incomparably small speed, that is, is in fact at rest, the first case of this seventh rule applies, and so it will follow from this rule that the two go together; whereas it was shown at the end of what was said on rule 6 that B is reduced to rest and transfers its whole force to C, which was at rest before and equal to it. I pass over other things, no less inconsistent, for brevity's sake. Finally, it must be pointed out that the author forgot the intermediate case, when the ratio of the bodies is the reciprocal of the ratio of the speeds, and it is not clear what is to be said about it by force of the rule. The outcome too ought to lie in the middle, at the very boundary of the two cases. But the first and second cases of the rule, though they share a boundary in the hypothesis, do not share a boundary in the outcome, which again conflicts with our criterion. The case in which B is larger than C has also been passed over in silence. A rule 8 ought to have been added, in which the author would have explained what happens when two unequal bodies moving in opposite directions with unequal speeds collide. A distinction also needed to be drawn between central and eccentric impact, and between perpendicular and oblique. But the examination must be brought to an end at last, and we must not dwell any longer on a doctrine that lies fallen and past saving.
Notes to text 107
- Descartes, Principia philosophiae II, arts. 46–52 (1644), give seven rules for the collision of two hard bodies B and C moving along one line, in a perfect vacuum, with no other bodies in the way. They follow from his laws of the conservation of the quantity of motion (size times speed, without regard to direction) and of the persistence of motion against rest. Leibniz quotes each rule in summary before answering it.
- Leibniz accepts Rule 1 alone. His proof of it already uses the conservation of "the sum of the forces" in his own sense (mass times the square of speed, art. 36 of this text, text 105), not Descartes's quantity of motion.
- "Peculiarem a nobis tractationem accepit" (art. 47): Leibniz had worked out the exact speeds after elastic collision in unpublished studies from the late 1670s and in the Dynamica de potentia (written about 1689–1690, GM VI). He also gave rules in the Specimen dynamicum (1695) and in an essay on the laws of collision sent to the Journal des savants [to be checked]. He does not say which treatment he means.
- Rule 6 (art. 51) is the most striking failure: Descartes has a body striking an equal body at rest rebound with three quarters of its speed. Leibniz's correction, that B stops and C goes on with B's whole speed, is the familiar result for perfectly elastic equal masses (as in a row of pendulum balls). He reaches it by continuity alone, as the midpoint between rules 4 and 5.
- Throughout, the OCR prints "reg. 4" or "reg. A." where the sense requires rule 1 (arts. 48 and 51, first mention), and "©" for C; these are corrected. In art. 49 the printed text reads "excessus ipsius B", which makes no sense when B is the smaller body; the Latin column keeps Gerhardt's reading, and the translation follows the required sense, "the excess of C" [?: a slip in Leibniz's manuscript or in Gerhardt's printing].
- "Doctrinae jacenti ac deploratae": the language of a lost cause, "a fallen and despaired-of doctrine". By 1692 Huygens, Wren and Wallis had published the correct rules (Philosophical Transactions, 1668–1669), and Leibniz treats Descartes's rules as already refuted by experiment as well as by his criterion.