A PHILOSOPHICAL ESSAY
ON
PROBABILITIES.
BY
PIERRE SIMON, MARQUIS DE LAPLACE.
TRANSLATED FROM THE SIXTH FRENCH EDITION
FREDERICK WILSON TRUSCOTT,
PH.D. (HARV.), Professorof GermanicLanguagesintheU'est Virginia.University,FREDERICK LINCOLN EMORY,
M.E. (WoR. POLY.INST.), Professorof MechanicsandAppliedMathematicsin the WestVirginiaUniversity;Mem.Amer.Soc.Mtch. Eng.
FIRST EDITION.
FIRST THOUSAND.
NEW YORK:
JOHN WILEY &
SONS.LONDON:
CHAPMAN
& HALL, LIMITED.1902.
BY F.W.TRUSCOTT
F. L.EMORY.
ROBERT DRUMMOND PRINTER,NEW YORK
Annex
TABLE OF CONTENTS.
PART
I.A PHILOSOPHICAL ESSAY ON PROBABILITIES.
CHAPTER
I.PAGE
Introduction i
CHAPTER
II.ConcerningProbability 3
CHAPTER
III.GeneralPrinciples of theCalculusof Probabilities 11
CHAPTER
IV.ConcerningHope 20
CHAPTER
V.AnalyticalMethodsof theCalculusof Probabilities 26
PART
II.APPLICATION OF THE CALCULUS OF PROBABILITIES.
CHAPTER
VI.GamesofChance 53
CHAPTER
VII.Concerning the Unknown Inequalitieswhich may Existamong
ChancesSupposedtobe Equal 56
iii
CHAPTER
VIII.PACK ConcerningtheLawsof Probabilitywhichresultfromthe Indefinite
Multiplication ofEvents 6
CHAPTER
IX.Application of theCalculusof ProbabilitiestoNatural Philosophy.. 73
CHAPTER
X.Applicationof theCalculus of Probabilities to theMoralSciences.. 107
CHAPTER
XLConcerningthe Probability ofTestimonies 109
CHAPTER
XII.Concerningthe SelectionsandDecisionsofAssemblies 126
CHAPTER
XIII.Concerningthe Probability of theJudgmentsofTribunals 132
CHAPTER
XIV.Concerning Tablesof Mortality,and theMeanDurations ofLife,
Marriage, andSomeAssociations 140
CHAPTER
XV.Concerningthe Benefitsof Institutionswhich Depend upon the
Probability ofEvents 149
CHAPTER
XVI.ConcerningIllusionsintheEstimationof Probabilities 160
CHAPTER
XVII.ConcerningtheVariousMeansofApproaching Certainty 176
CHAPTER
XVIII.HistoricalNoticeof theCalculusof Probabilities to1816 185
ERRATA.
Page 89, line 22, for Pline read Pliny
"
102, lines 14, 16, " minutes "
days
"
143, line 25, "
sun soil
"
177, lines15, 17, 18,21, 22, 24,forprimaryread prime
"
182, line 5, for conjunctions read being binary
A PHILOSOPHICAL ESSAY ON
PROBABILITIES.
CHAPTER
I.INTRODUCTION.
THIS
philosophical essay is the development of a lecture on probabilities which I delivered in 1795 to the normal schools whither Ihad been called, by a decree of the national convention, as professor of mathematicswithLagrange. Ihaverecentlypublished upon the same subjectaworkentitled TheAnalytical Theory ofProbabilities. I present here without the aidofanalysistheprinciplesandgeneralresultsofthis theory,applying themtothe mostimportant questions oflife,whichareindeedforthemostpartonlyproblems of probability. Strictlyspeaking itmay
even be saidthatnearlyall ourknowledgeisproblematical; andin the small
number
of things whichwe
areabletoknow
with certainty, even in the mathematical sciences themselves, the principalmeans
forascertaining truth induction and analogy arebased on probabilities;A
PHILOSOPHICAL ESSAYON
sothatthe entiresystemof
human
knowledge is con- nected withthetheorysetforth in this essay. Doubt- lessitwillbe seen here withinterest thatinconsidering, even in the eternal principles of reason, justice, and humanity, only the favorable chances which are con- stantlyattachedtothem, there isa greatadvantage in followingthese principlesand seriousinconveniencein departing from them: their chances, likethosefavor- able to lotteries,alwaysend byprevailing inthemidst of the vacillationsof hazard. I hope that thereflec- tions given in this essaymay
merit the attention ofphilosophers and direct it to a subject so worthy of engagingtheirminds.
CHAPTER
II.CONCERNING PROBABILITY.
ALL
events, even thosewhich on account of their insignificance donotseem to followthe great laws of nature, arearesultofitjustas necessarily astherevolu- tions ofthesun. In ignorance of thetieswhichunite such events totheentire systemofthe universe, they have beenmade
to dependuponfinal causes or upon hazard, accordingas they occur andare repeated with regularity, orappear without regardtoorder;but these imaginary causes have gradually receded with the widening boundsof knowledge anddisappearentirely beforesound philosophy, which seesinthem
only the expression of ourignorance of thetrue causes.Present events are connected with preceding ones
by
atie basedupon theevident principle thata thing cannot occur without a causewhich producesit. This axiom,known
bythename
ofthe principleofsufficient reason, extends even to actionswhich are considered indifferent; thefreestwill is unable without a determi- native motive to give
them
birth; ifwe
assume two positions with exactly similar circumstances and find that thewill is active inthe one and inactive in the3
4
A
PHILOSOPHICAL BBSAY
other,
we
saythatitschoiceisaneffectwithout acause.It is then, says Leibnitz, the blind chance of the Epicureans.
The
contrary opinion isan illusion ofthe mind, which, losing sightof the evasive reasons of the choice ofthe will in indifferent things, believes that choiceis determinedofitselfandwithout motives.We
ought then to regard the present state of the universe as the effect ofits anterior stateand as the cause of the one which is to follow. Given for one instantanintelligencewhichcouldcomprehend allthe forcesby which natureis animated andthe respective situationofthebeingswho
compose it anintelligence sufficiently vast to submit these data to analysis it would embraceinthe same formula themovements
of the greatest bodies of the universe and those of the lightest atom; forit, nothing would be uncertain and thefuture, as the past, would be present to its eyes.The human
mindoffers, inthe perfectionwhich ithas beenable to give toastronomy,a feebleideaofthis in- telligence. Its discoveriesinmechanicsandgeometry, added tothat of universal gravity, have enabled it to comprehend in the same analytical expressions the past and future states of the system of the world.Applyingthesame methodto
some
otherobjects ofitsknowledge,ithas succeededinreferring togeneral laws observed
phenomena
and in foreseeing those which given circumstancesoughttoproduce. Alltheseefforts inthesearchfortruthtend tolead it backcontinually to the vast intelligence whichwe
havejust mentioned, butfromwhichitwillalways remaininfinitelyremoved.This tendency, peculiar to the
human
race, is that which rendersitsuperior toanimals;andtheirprogress
in thisrespect distinguishes nationsandages andcon- stitutestheirtrue glory.
Letus recall that formerly, andatno remoteepoch, anunusualrainoranextremedrought, acomet having
in train a very long tail, the eclipses, the aurora borealis, and in general all the unusual
phenomena
were regarded as somany
signs of celestial wrath.Heaven
was invoked in order to avert their baneful influence.No
one prayedtohave the planetsandthe sun arrested in their courses: observation had soonmade
apparent the futility of such prayers. But as these phenomena, occurring anddisappearingat long intervals, seemedtooppose theorderof nature, itwas supposed that Heaven, irritated by the crimes of the earth, had created them "to announce its vengeance.Thus the longtailofthe cometof1456spread terror through Europe, alreadythrown into consternation
by
the rapid successes of the Turks,who
had just over- throwntheLower
Empire. Thisstarafterfour revolu- tions has excitedamong
usa very different interest.The
knowledge ofthelaws of the systemof theworld acquired in the interval had dissipated the fears begotten bythe ignorance of the true relationship ofman
to the universe; and Halley, having recognized the identity ofthiscometwith those of the years 1531, 1607, and 1682, announced itsnextreturn forthe end of the year 1758 or the beginning of the year 1759.The
learned world awaited with impatiencethisreturn which was to confirm one ofthe greatest discoveries that have beenmade
in the sciences, and fulfil the prediction of Senecawhen
he said, in speakingofthe revolutions ofthosestarswhich fall from an enormous6
A
PHILOSOPHICAL ESSAYON
height:
"The
daywillcome
when, bystudypursued through several ages, the thingsnow
concealed will appear with evidence; andposteritywillbe astonished that truths so clear had escapedus.''
Clairaut then undertooktosubmitto analysis theperturbationswhich the comet had experienced by theaction of the two great planets, Jupiterand Saturn; after
immense
cal- culations he fixed its next passage at the perihelion towardthebeginningof April, 1759,whichwasactuallyverifiedbyobservation.
The
regularitywhich astronomy shows us in themovements
of the comets doubtless existsalsoinallphenomena. -The
curve described bya simple molecule of airor vapor is regulated inamanner
just as certain as the planetaryorbits; the onlydifference betweenthem
isthatwhichcomesfrom our ignorance.
Probabilityis relative, in part to this ignorance, in part to our knowledge.
We know
thatofthree or a greater numberofevents asingle one oughttooccur;but nothing inducesus to believe thatoneof
them
will occurratherthan theothers. Inthisstateofindecisionitis impossibleforus toannouncetheiroccurrence with certainty. It is, however, probable thatone of these events, chosen atwill, will not occur because
we
see several casesequallypossiblewhich excludeitsoccur- rence, while only asingleonefavors it.The
theory of chance consists in reducing all the events ofthe same kind to acertainnumber
ofcases equallypossible, that isto say, tosuchaswe may
beequally undecided about in regard to their existence, and in determining the
number
of cases favorable to the eventwhose probability is sought.The
ratio ofthis
number
to that of all the cases possible is the measure of this probability, which is thus simply a fraction whose numerator is thenumber
offavorable casesandwhose denominatoris thenumber
ofallthe cases possible.The
preceding notion of probabilitysupposes that, in increasingin thesame ratiothenumber
of favorable casesandthatofallthe cases possible, theprobability remains the same. In ordertoconvince ourselveslet us take two urns,A
and B, the first containing four whiteandtwoblack balls, and the second containing only two white balls and one black one.We may
imagine thetwo blackballsof the firsturn attachedby a thread which breaks at the
moment when
one of them is seized in ordertobe drawn out, andthefour whiteballsthus formingtwosimilarsystems. All the chanceswhich willfavorthe seizureofone oftheballs of the black systemwill lead toa black ball. Ifwe
conceive
now
thatthe threadswhichunitethe ballsdo notbreak atall, itisclearthatthenumber
ofpossible chances will not change any more than that of the chances favorableto theextractionof the blackballs;buttwoballs will be drawnfrom the urn atthe same time; the probabilityof drawingablack ballfrom the urn
A
will thenbe the sameas at first. Butthenwe
have obviously the case of urnB
with the singlediffer- ence that the three balls of this last urn would be replacedbythreesystems oftwo balls invariablycon- nected.When
all the cases are favorable to an event the probability changes to certainty and its expression becomesequalto unity.Upon
thiscondition, certaintyA
PHILOSOPHICAL ESSAYand probabilityarecomparable, althoughthere
may
bean essential difference between the two states of the mind
when
atruthis rigorouslydemonstrated toit, crwhen
itstillperceivesa small source oferror.In thingswhich are only probable the difference of thedata,whicheach
man
hasinregardtothem,isone oftheprincipalcausesofthe diversity ofopinionswhich prevailinregard tothe sameobjects. Let ussuppose, forexample, thatwe
havethree urns, A, B, C, one of which contains only black ballswhile thetwo others contain only white balls; a ballis to be drawn from the urnC
and the probabilityisdemanded
that this ballwill be black. Ifwe
do notknow
which ofthe threeurnscontainsblackballs only,sothatthereisno reasontobelieve thatitisC
ratherthanB
orA, these threehypotheseswillappear equallypossible,andsince a blackball can bedrawnonlyinthe firsthypothesis, the probability ofdrawingit isequalto onethird. If itisknown
that the urnA
containswhite balls only, the indecisionthen extends onlytothe urnsB
andC, andthe probability thattheball drawnfrom the urnC
will be black is one half. Finally this probability changesto certaintyif
we
are assuredthat the urnsA
and
B
containwhiteballs only.It is thus that an incident related to a numerous assemblyfinds variousdegrees of credence, according to the extent of knowledge of the auditors. If the
man who
reportsitis fullyconvinced of it and if, byhispositionandcharacter,heinspiresgreatconfidence, his statement, however extraordinary it
may
be, will have for the auditorswho
lack information thesame
degree of probability as an ordinary statementmade
bythesame man, andtheywill have entire faithinit.
Butif
some
one of themknows
that the sameincident is rejected by other equally trustworthymen, hewill be in doubt andthe incidentwillbediscredited bythe enlightened auditors,who
will reject it whether itbe inregard tofacts well averred orthe immutable laws of nature.It is tothe influence of the opinion ofthose
whom
themultitude judgesbestinformed andto
whom
ithas been accustomed to give its confidence in regard to themost important matters oflifethatthe propagation ofthoseerrors is due whichintimes of ignorancehave covered the face of the earth. Magic and astrology offerus twogreat examples. These errors inculcated in infancy, adopted without examination, and havingfor a basis only universal credence, have maintained themselves during a very long time; but at last the progress ofsciencehas destroyed them inthe mindsof enlightened men, whose opinion consequently has caused
them
to disappear evenamong
thecommon
people, through the powerof imitationandhabit
w
rhich had so generally spreadthem
abroad. This power, the richestresourceofthe moralworld,establishesand conserves ina whole nationideas entirelycontraryto those which it upholds elsewhere with the same authority.What
indulgence oughtwe
not then to have foropinionsdifferent from ours,when
this differ- enceoftendependsonlyuponthevarious pointsofview where circumstanceshave placedus! Letus enlighten thosewhom we
judgeinsufficiently instructed; butfirst let us examinecriticallyourown
opinions and weigh withimpartiality theirrespective probabilities.A
PHILOSOPHICAL ESSAYThe
difference of opinions depends, however, upon the manner in which the influence ofknown
data isdetermined.
The
theory ofprobabilitiesholdstocon- siderationsso delicate thatitisnotsurprising thatwith the same datatwo persons arrive at different results, especially in very complicated questions. Let us examinenow
thegeneralprinciples ofthistheory.CHAPTER
III.THE GENERAL PRINCIPLES OF THE CALCULUS OF
PROBABILITIES.FirstPrinciple.
The
firstof these principles isthe definition itselfofprobability,which,ashas been seen,isthe ratioof the
number
of favorable cases to thatof allthe casespossible.Second Principle. But that supposes the various cases equally possible. If they are not so,
we
will determine first their respective possibilities, whose exact appreciationisone ofthe mostdelicate points of the theory of chance.Then
the probability will be thesum
of the possibilities of each favorable case.Letusillustrate thisprincipleby an example.
Let us suppose that
we
throwinto the air a large and very thin coin whose two large opposite faces, whichwe
will callheadsandtails, are perfectly similar.Let us find the probabilityof throwing headsat least one time intwo throws. It is clearthat fourequally possible cases
may
arise, namely, heads at the firstandatthe secondthrow; headsatthe firstthrowand
tails at the second; tails atthe firstthrow andheads at the second; finally, tailsatboth throws.
The
first12
A
PHILOSOPHICAL ESSAYthree cases are favorable tothe eventwhoseprobability issought; consequentlythis probabilityis equal to|;
so that it isa bet ofthree to onethat headswill be thrownatleastonceintwothrows.
We
can countatthisgame
only threedifferent cases, namely, headsatthefirstthrow, which dispenses with throwing a second time; tails at the first throw and headsatthesecond; finally, tailsatthe firstandatthe second throw. Thiswould reduce the probability to| if
we
should consider with d'Alembert these three cases asequally possible. But itis apparentthat the probabilityof throwing heads at the first throwis f, while that of thetwo other cases is J, the first case being a simple eventwhich corresponds totwoevents combined: headsatthefirstandatthe second throw, and heads at the firstthrow, tails at the second. Ifwe
then, conforming tothe second principle, addthe possibilityf ofheads at the first throwto the possi- bility J of tails at the first throw and heads at the second,we
shall have f for the probability sought, which agrees with what is found in the suppositionwhen we
playthetwo throws. This'suppositiondoes not changeat allthe chanceof thatonewho
bets onthis event; itsimplyserves toreduce thevarious cases tothe casesequally possible.
ThirdPrinciple.
One
of themost importantpoints ofthetheoryof probabilities and thatwhich lends the most to illusions is themanner
inwhich these prob- abilitiesincrease or diminishbytheir mutualcombina-tion. Iftheeventsareindependent ofoneanother, the probabilityoftheircombined existence is the product oftheirrespective probabilities.
Thus
the probability13 of throwing one ace with a single die is ^; that of throwing two aces in throwingtwo dice at the same time is --$.
Each
face of theonebeing able tocom- bine with the six facesofthe other, there are in fact thirty-six equally possible cases,among
which onesingle casegivestwo aces. Generally the probability that a simple event in the same circumstances will occur consecutively a given
number
oftimesis equalto the probabilityofthissimple eventraised tothe power indicated bythisnumber. Havingthusthe successive powers of afractionlessthanunitydiminishing without ceasing, an eventwhich depends upon aseriesof very great probabilitiesmay become
extremely improbable.Suppose then an incident be transmitted to us by twenty witnesses in such
manner
that the first has transmitted it to the second, the secondtothe third,andsoon. Suppose againthatthe probability ofeach testimony be equal to the fraction T9 ; that of the incident resulting from the testimonies will be less than .
We
cannotbetter comparethisdiminution of the probability than with the extinction ofthelightof objects bythe interposition of several pieces ofglass.A
relatively smallnumber
of pieces suffices to takeaway
the viewofan object that a single piece allows us toperceive inadistinctmanner.The
historiansdo not appear to have paid sufficient attention to this degradation of the probability of eventswhen
seen acrossa greatnumber
ofsuccessivegenerations;many
historical events reputed as certainwould beat least doubtfuliftheyweresubmitted tothis test.
Inthepurely mathematicalsciencesthe mostdistant consequences participateinthe certainty of the princi-
M A
PHILOSOPHICAL ESSAYpiefromwhich theyare derived. In the applications of analysis tophysics theresultshave allthecertainty of facts or experiences. But inthe moral sciences, whereeach inferenceis deduced fromthatwhich pre- cedesitonlyina probable manner, however probable thesedeductions
may
be, thechanceof error increases with theirnumber and ultimatelysurpasses the chance of truth in the consequences very remote from the principle.FourthPrinciple.
When
two events depend upon each other, the probabilityof thecompound
event istheproductofthe probabilityofthefirsteventandthe probabilitythat, thiseventhavingoccurred, the second will occur. Thus in the preceding case of the three urnsA, B, C,ofwhich two contain only white balls and one contains only black balls, the probability of drawing a white ball fromtheurn
C
is f, since of the three urns onlytwo contain ballsof that color. Butwhen
a whiteballhas beendrawn fromthe urn C, the indecision relative to thatoneofthe urnswhichcontain only black balls extends onlyto the urnsA
and B;the probability ofdrawinga whiteball from the urn
B
is ;the productof\by , or ,isthen the probability ofdrawing twowhite balls at one time from the urns
B
andC.We
seebythisexample theinfluence ofpastevents uponthe probabilityoffuture events. Forthe prob- abilityof drawinga white ball from the urn B, which primarilyis f,becomes \when
a white ball has been drawn fromtheurnC
; itwould change to certaintyif ablackball hadbeen drawn from thesame urn.We
will determine this influence by means ofthe follow-
ing principle, which is a corollary of the preceding one.
Fifth Principle. If
we
calculate apriori the prob-abilityof the occurred eventandthe probability ofan eventcomposed ofthatone andasecond onewhichis
expected, the second probability divided by the first
will be the probability of the event expected, drawn from the observedevent.
Here is presented the question raised by
some
philosophers touching the influence of the past upon the probability of the future. Let us suppose at the playofheads andtailsthatheads has occurred oftener than tails.By
this alonewe
shall be led to believethat in the constitution of the coin there is a secret cause which favors it.
Thus
in the conduct of lifeconstant happiness is a proof of competency which should induce us to employpreferablyhappy persons.
Butifbythe unreliability ofcircumstances
we
arecon- stantlybrought backtoa state of absoluteindecision,if,forexample,
we
changethe coinateachthrowatthe play ofheadsandtails, thepast can shednolightupon the futureanditwould beabsurdtotake account ofit.Sixth Principle.
Each
of the causes to which an observed eventmay
beattributedis indicatedwith just asmuch
likelihood as thereisprobability that theevent will take place, supposing the event to be constant.The
probabilityof the existence of any oneof these causesisthen afractionwhose numerator isthe prob- abilityoftheeventresultingfrom thiscause andwhose denominator is thesum
of the similar probabilities relative to allthe causes; ifthesevarious causes, con- sideredapriori, areunequally probable,itisnecessary,16
A
PHILOSOPHICAL ESSAYin placeofthe probabilityof the eventresultingfrom each cause, to employthe productofthis probability bythe possibility of thecauseitself. Thisisthe funda- mentalprincipleofthisbranchoftheanalysisofchances whichconsistsin passing from eventstocauses.
This principle gives the reason
why we
attributeregulareventsto a particular cause.
Some
philosophers have thought that these eventsare lesspossible than others and that at the play of heads and tails, for example,thecombinationinwhich headsoccurs twenty successive times is less easyin its nature than those whereheadsandtailsaremixedinanirregularmanner.But this opinion supposes that past events have an influence on the possibility of future events, which is
notatalladmissible.
The
regular combinations occur more rarelyonly because theyareless numerous. Ifwe
seek a causewhereverwe
perceive symmetry, itisnotthat
we
regard a symmetrical eventaslesspossible thanthe others, but, sincethis event ought to be theeffect ofaregular causeor that of chance, the firstof these suppositions is more probable than the second.
On
a tablewe
see letters arranged in this order,Constantinople,
andwe
judgethatthisarrange- mentisnotthe resultof chance, not because itisless possible than the others, for if this word were not employed in any languagewe
should not suspect itcame
from anyparticular cause, butthiswordbeing in useamong
us, itis incomparablymore
probable that some person has thus arranged the aforesaid letters thanthatthisarrangementisdue tochance.This is the place to definethe word extraordinary.
We
arrangeinour thoughtallpossibleeventsinvariousclasses; and
we
regardas extraordinary those classes which include a very smallnumber.Thus
atthe play ofheads and tails the occurrence ofheads a hundred successivetimes appearstousextraordinary becauseof thealmost infinitenumber
of combinationswhichmay
occurina hundredthrows; andif
we
divide thecom-
binations into regular seriescontainingan order easy tocomprehend, and into irregularseries, the latterare incomparably more numerous.
The
drawing of a white ball from an urnwhichamong
amillion ballscontains onlyoneofthiscolor,the others being black, wouldappearto us likewise extraordinary, because
we
form only two classes of events relative to the two colors. But the drawing of thenumber
475813, forexample, from anurn thatcontains a millionnumbers seems to us an ordinary event; because, comparing individually the numbers with one another without dividing them into classes,
we
have no reason to believe that one ofthem
will appear sooner than the others.From
what precedes,we
oughtgenerallytoconclude thatthe more extraordinary the event, the greater the need of its being supported by strong proofs. For thosewho
attest it, being able todeceive or to have been deceived, these two causes are asmuch more
probableasthereality of the eventisless.We
shall see this particularlywhen we come
to speak of the probabilityof testimony.SeventhPrinciple.
The
probabilityofafutureeventis the
sum
of the products of the probability of each cause, drawn from the event observed, by the prob- ability that, this cause existing, the future eventwillA
PHILOSOPHICAL ESSAYoccur.
The
following example will illustrate thisprinciple.
Letusimagine an urnwhichcontains onlytwoballs, each ofwhich
may
be eitherwhite or black.One
of these ballsis drawn andis putback intothe urn before proceeding to anew
draw. Suppose that inthe first two draws white balls have been drawn; the prob- abilityof again drawingawhite ballatthe thirddrawisrequired.
Only two hypotheses can be
made
here: either one of the balls is whiteandthe other black, orbothare white. In the first hypothesis the probability ofthe event observed is J; it is unity or certainty in the second. Thus in regarding these hypotheses as somany
causes,we
shall have for the sixth principle% and | for their respective probabilities. Butifthe
first hypothesis occurs, the probability of drawing a whiteball atthethirddrawis^; itisequalto certainty in the second hypothesis; multiplying then the last probabilitiesbythose of the corresponding hypotheses, the
sum
oftheproducts, orT9^, will be the probability ofdrawinga whiteball atthethird draw.When
the probabilityof a single eventisunknown we may
suppose it equal to anyvalue from zero to unity.The
probability ofeach of these hypotheses, drawnfrom the event observed, is, by the sixth prin- ciple, a fractionwhose numeratoris the probability of theeventin thishypothesisand whose denominator isthe
sum
of thesimilar probabilities relative to all the hypotheses.Thus
the probability thatthepossibility of theeventiscomprised within given limitsisthesum
of the fractionscomprisedwithin these limits.
Now
ifwe
multiply each fraction by the probability of the future event,determined inthe corresponding hypothe-sis,the
sum
ofthe productsrelativetoallthe hypotheses will be, bythe seventh principle,theprobability ofthe future event drawn from the event observed.Thus we
find thatan eventhavingoccurred successivelyanynumber
of times, the probability that it will happen again thenexttime is equalto thisnumber
increased by unity divided by thesame
number, increased by twounits. Placing the mostancientepoch of history atfivethousand years ago, or at 182623 days, andthe sun having risen constantly in the interval at each revolution of twenty-four hours, itis a betof1826214 to one that it will rise again to-morrow. But thisnumber
isincomparablygreaterforhim who,recogniz- inginthe totality ofphenomena
theprincipalregulator of days and seasons, sees thatnothing atthe presentmoment
canarrest the course ofit.BuffoninhisPoliticalArithmeticcalculates differently the preceding probability.
He
supposes thatitdiffers from unity onlybyafractionwhose numeratoris unity and whose denominator is thenumber
2raised to a powerequaltothenumber
ofdayswhich haveelapsed since the epoch. But the truemanner
of relating pastevents with the probabilityofcausesand of future eventswasunknown
to this illustriouswriter.CHAPTER
IV.CONCERNING HOPE.
THE
probability ofevents serves to determine the hope or the fear ofpersons interested in their exist- ence.The
word hope has various acceptations; it expresses generally the advantage of that onewho
expects acertain benefitinsuppositionswhich areonly probable. This advantage in the theory ofchanceis
a product of the
sum
hoped forbythe probabilityof obtainingit; itisthe partialsum
which oughtto resultwhen we
do not wish torun the risks ofthe eventin supposingthat the divisionismade
proportional tothe probabilities. This divisionis the onlyequitable onewhen
allstrange circumstancesareeliminated; because an equal degreeof probability givesan equal right to thesum
hoped for.We
will call this advantage mathematicalhope.EighthPrinciple.
When
the advantagedepends on severaleventsit is obtained bytaking thesum
ofthe products oftheprobabilityofeach eventby
thebenefit attached toits occurrence.Letus applythis principle to
some
examples. Letus suppose that at the play of heads and tails Paul receivestwo francsifhethrows headsatthefirstthrow and five francs if he throws it only at the second.
Multiplyingtwofrancs bythe probability of the first case, andfive francs
by
theprobability of the second case, thesum
of the products, or two and a quarter francs, willbe Paul's advantage. It isthesum
which he oughtto giveinadvance tothatonewho
has given him this advantage; for, in order to maintain the equalityof the play, the throw ought to be equal to theadvantage whichitprocures.IfPaulreceives two francs
by
throwingheadsatthefirstandfive francsby throwing itatthesecond throw, whether hehasthrownit ornotatthe first, the prob- ability ofthrowing headsatthe secondthrowbeing , multiplyingtwo francs andfive francsby the
sum
of these products will give threeand one half francsfor Paul's advantage andconsequentlyfor his stakeatthe game.Ninth Principle. In a series of probable events of which the onesproducea benefitandthe others aloss,
we
shallhave the advantage which results from itby
makingasum
of the products of the probabilityofeach favorable event bythe benefit which it procures, and subtracting from thissum
that ofthe products of the probability ofeach unfavorable eventby
the loss whichisattachedto it. Ifthe second
sum
isgreater than thefirst,thebenefitbecomesa lossand hope is changed to
fear.
Consequently
we
ought always inthe conduct oflife tomake
the product ofthe benefit hoped for, by its probability,atleastequaltothesimilarproductrelativeA
PHILOSOPHICAL ESSAYto theloss. Butitisnecessary, in orderto attain this, to appreciate exactly the advantages, the losses, and their respective probabilities. Forthisa great accuracy of mind, adelicate judgment,and a great experience in affairs is necessary; itis necessaryto
know how
to guard one's selfagainstprejudices, illusions offearor hope, and erroneousideas, ideasoffortune andhappi- ness, withwhichthe majorityofpeople feed theirself- love.The
application ofthe preceding principles to the following question has greatly exercised the geometri- cians. Paul playsatheadsandtailswith the condition ofreceivingtwo francsifhe throws headsatthe first thro\v, four francs ifhe throws it onlyat the second throw, eight francs if he throws it onlyat the third,and so on. His stakeatthe playought to be, accord- ing to the eighth principle, equal to the
number
of throws, so that if thegame
continues to infinitythe stake ought to be infinite. However, no reasonableman
wouldwish toriskat thisgame
evena smallsum,for example five francs.
Whence
comes this differ- encebetweenthe resultofcalculationandthe indication ofcommon
sense?We
soon recognizethatitamounts tothis: thatthe moral advantage which a benefit pro- curesforusisnot proportionaltothis benefitandthatit depends upon athousand circumstances, often very difficult to define, but ofwhichthe most general and mostimportantisthat of fortune.
Indeeditisapparentthatonefranchas
much
greater value forhimwho
possessesonly ahundredthan for a millionaire.We
ought then to distinguish in the hoped-for benefit its absolute from its relative value.Butthe latteris regulatedbythe motiveswhich
make
it desirable, whereas the first is independent of them.
The
general principle for appreciating this relative value cannot be given, but here is one proposed by Daniel Bernoulliwhichwill serveinmany
cases.TenthPrinciple.
The
relativevalueof aninfinitely smallsum
isequal toits absolutevalue dividedbythe total benefit of the person interested. This supposes that everyone has a certain benefitwhosevalue can never beestimatedas zero. Indeedeventhatonewho
possesses nothing always gives to the product ofhis labor andtohis hopes a value at least equal to that which isabsolutelynecessary tosustainhim.
If
we
applyanalysis tothe principle justpropounded,we
obtain the followingrule: Letusdesignatebyunity the part of the fortuneof an individual, independent of his expectations. Ifwe
determine the differentvalues thatthisfortunemay
have byvirtueof these expecta- tionsandtheir probabilities,the product of these values raised respectively to the powers indicated by their probabilitieswill be the physical fortunewhich would procure fortheindividual thesame
moral advantage whichhereceivesfrom the part of his fortunetakenas unityand fromhis expectations; bysubtracting unity fromtheproduct, thedifference willbe the increase of the physical fortunedue toexpectations:we
will call thisincreasemoralhope. Itiseasyto seethat itcoin- cideswith mathematical hopewhen
the fortune taken as unity becomesinfinite in reference to the variations which it receives from the expectations. Butwhen
these variations are an appreciable partofthis unity24
A
PHILOSOPHICAL ESSAYthetwo hopes
may
differvery materiallyamong
them-selves.
This rule conduces to results conformable to the indications of
common
sensewhich can bythismeans
be appreciated withsome
exactitude.Thus
in the preceding question it is found that if the fortune of Paul istwo hundredfrancs, he oughtnot reasonablyto stake morethan ninefrancs.The same
rule leads us again to distribute the danger over several partsof a benefitexpectedratherthantoexposethe entire benefit to thisdanger. It results similarly that atthe fairestgame
the lossis always greater than the gain. Let ussuppose, forexample,thata playerhavingafortune ofonehundredfrancsrisksfiftyatthe play of headsandtails; his fortune after his stake at the play will be reducedtoeighty-seven francs, thatisto say, this last
sum
would procure for the player thesame
moral advantage as the stateofhis fortuneafter the stake.The
play is then disadvantageous even in the case where the stake is equal to the product of thesum
hoped for,by its probability.We
can judge bythis oftheimmoralityofgames
inwhich thesum
hopedfor is below this product.They
subsist only by false reasoningsand bythe cupiditywhichthey excite and which, leading the people to sacrificetheir necessaries to chimerical hopes whose improbabilitythey arenot in conditionto appreciate, arethe source of aninfinity ofevils.The
disadvantage ofgames
of chance, theadvantage ofnot exposingtothe same dangerthe whole benefit thatisexpected, andallthesimilar results indicatedbycommon
sense, subsist,whatevermay
be the function25 of the physical fortune which for each individual expresses his moral fortune. It is enough that the proportion of the increase of this function to the increase of the physical fortune diminishes in the measurethatthelatterincreases.
CHAPTER
V.CONCERNING THE ANALYTICAL METHODS OF THE CALCULUS OF
PROBABILITIES.THE
application of theprinciplewhichwe
havejust expounded to the various questions of probability requires methods whose investigation has given birth to several methods of analysis and especially to the theory of combinations and to the calculus of finite differences.If
we
form the productofthe binomials, unity plus thefirst letter, unity plusthe second letter, unityplus the third letter, and so on up to n letters, and sub- tract unity from this developed product, the result willbe thesum
ofthe combination ofall these letters taken one byone, two by two, threeby
three, etc., each combination having unity for a coefficient. In order to have thenumber
ofcombinations ofthese n letterstakensbys times,we
shall observethatifwe
supposethese lettersequal
among
themselves, thepre- ceding product willbecome
the nth power of the binomialone plus the firstletter; thus thenumber
of combinations ofnletters takens bystimeswillbe the coefficient ofthe sth power of the first letter in the27 development in this binomial; and this
number
isobtained by meansof the
known
binomial formula.Attention mustbe paid tothe respective situations of theletters ineach combination, observing thatifa second letteris joined tothefirst it
may
be placed in the firstorsecond position which gives two combina-tions. If
we
join to thesecombinations a third letter,we
can giveit ineach combination thefirst,the second, and the third rank which forms three combinations relative to each of thetwoothers, in all sixcombina-tions.
From
this it is easy to conclude that thenumber
ofarrangementsofwhichs lettersare suscepti- bleisthe product of thenumbersfrom unity tos. In orderto payregard to the respective positions ofthe lettersitisnecessary then to multiplybythis product thenumber
of combinations of nletters sbys times, which is tantamountto takingaway
the denominator ofthe coefficient of thebinomialwhich expressesthisnumber.
Let usimagine a lotterycomposed of nnumbers, of which rare drawnat each draw.
The
probabilityisdemanded
ofthe drawing of s given numbersin one draw.To
arrive at this letus form a fractionwhose denominatorwill be thenumber
ofall the cases possi- ble or ofthe combinations of n letters taken rby
rtimes, andwhose numeratorwill be the
number
ofall the combinationswhich contain the givens numbers.This last
number
isevidentlythat of the combinations of the othernumbers taken nless sbynless stimes.This fraction will be the required probability, and
we
shall easilyfind that it can be reducedto a fraction whose numerator is the
number
ofcombinations of r28
A
PHILOSOPHICAL ESSAYnumberstakensbystimes,and whose denominatoris the number of combinations of n numbers taken similarlysbystimes.
Thus
in thelottery ofFrance, formed as isknown
of 90 numbers of whichfive are drawnateach draw, the probabilityofdrawinga given combinationis -&> or TV
;the lotteryought thenforthe equalityof the playto give eighteen times the stake.The
totalnumber
of combinations two by two of the 90 numbersis4005, andthatof the combinations two by two of 5 numbers is 10.The
probability of the drawingof a given pairis then3-^-5-, andthe lottery oughttogivefourhundred andahalftimesthe stake;
itoughttogive 11748 timesfora given tray, 511038 timesfora quaternary,and 43949268 timesforaquint.
The
lotteryisfar from giving the player these advan- tages.Suppose in an urnawhite balls, bblack balls, and
after havingdrawn aballit is put back intothe urn;
the probabilityis asked thatin
number
of drawsm
whiteballs and n
m
blackballs will be drawn. It is clear that thenumber
of cases thatmay
occur ateach drawing is a-j-b.
Each
case of the second drawingbeingable tocombinewithallthe cases of thefirst, the
number
of possible cases in two drawingsis the square ofthebinomial a-\-b. In thedevelopment ofthissquare, the squareofaexpresses thenumber
of casesinwhichawhite ballistwicedrawn, the double product ofa by b expresses thenumber
of cases inwhichawhiteballandablackballaredrawn. Finally, thesquare ofbexpresses the