University of Pennsylvania Press

Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

Leibniz's Theory of Space and Time

Author(s): Kaith Emerson Ballard


Source: Journal of the History of Ideas, Vol. 21, No. 1 (Jan. - Mar., 1960), pp. 49-65
Published by: University of Pennsylvania Press
Stable URL: http://www.jstor.org/stable/2707998 .
Accessed: 04/01/2014 13:14

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

University of Pennsylvania Press is collaborating with JSTOR to digitize, preserve and extend access to
Journal of the History of Ideas.

http://www.jstor.org

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ'S THEORY OF SPACE AND TIME

BY KAITH EMERSON BALLARD

Duringthe yearbeforehis deathin 1716,Leibnizcorresponded


withSamuelClarke(1675-1729),an Englishtheologian, whoseknowl-
edge of philosophyand scienceenabledhim to offera defenseof
Newton'sphilosophical and scientific theoriesagainstthe criticisms
formulated by Leibniz.1 Althoughthe correspondence deals witha
widerangeofsubjects,it is particularly valuableas a sourceofmate-
rialon Leibniz'srelationaltheoryofspaceand time,fornowhereelse
in hiswritingsdoesLeibnizexpoundthistheorywithmorethorough-
nessor consistency.
Besidesproviding a revealingperspective of thesystematic inter-
relationshipbetweenLeibniz'sphysicsand metaphysics, the corre-
spondencerepresents an important documentin the historyof the
conflictbetweenabsoluteand relationaltheoriesof space and time.
ErnstCassirerin factsaid that " the controversy betweenNewton
and Leibnizis one of themostimportant phenomenain the history
ofmodernthought." 2
In orderto understandproperlyLeibniz'srelationaltheoryof
space and timeit is necessary firstto have clearlyin mindtheNew-
tonianabsolutetheory, in oppositionto whichLeibniz'stheorywas
formulated, and againstwhichmostof his arguments weredirected.
Newton'sabsolutetheoryof space and timeis based upon his
beliefthat certaindynamicalphenomenanot onlypermit,but also
require,a distinction betweenabsoluteand relativeacceleration.It
is evident,however, thatthispositionraisestwo separatequestions
whichmustbe critically examined:1) Do the " empiricalfacts" re-
quirea distinctionbetweenabsoluteand relativemotion? 2) Does
thelatterdistinctionpresuppose an absolutetheoryofspaceand time
ofthetypewhichNewtonheld? Thesetwoquestionsareeasilysepa-
rableif we reflect
thattheanswerto thefirstquestionwillbe deter-
minedlargelyby scientific considerations,whereasthe answerto the
secondwill be determined primarily by metaphysical or epistemo-
logicalconsiderations. As is wellknown,Mach and Einstein,among
others,have answeredbothof thesequestionsin a waywhichdiffers
fundamentally fromthe answerswhichNewtongave. Using the
sametypeof empiricalevidencewhichwas availableto Newton,the
moderntheoryof relativity postulatesa purelyrelationaltheoryof
space and time. This factaloneoughtto makeus awarethatNew-
1 The fivelettersby Leibnizand fiverepliesby Clarkeare reprinted
in a
new
volume,editedwithan introductionand notesby H. G. Alexander,
The Leibniz-
ClarkeCorrespondence(Manchester,1956).
" Newtonand Leibniz,"The Philosophical
2 Cassirer, Review(1943), 366.
49

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
50 KAITH EMERSON BALLARD

ton'sinference fromthe " empiricalevidence" to an absolutetheory


of space and timemay legitimately be challenged.Withoutat this
the
timeexamining arguments which Newtonused to supportthis
inference,let us statehis theoryof space and timein enoughdetail
to makeLeibniz'scriticisms ofit intelligible.
Newton'sowntheoryof space and timeis presented in a " Scho-
liumto Definition VIII " of thePrincipia,3but not all of the philo-
sophicalimplications of his positionare developedin detail there.
UnderthestimulusofLeibniz'scriticisms, Clarkeoffered a somewhat
morerigorousand completephilosophical formulation and defense
of thetheory;but it mustnotbe supposedthatClarke,in his argu-
mentsagainstLeibniz,departedin any essentialrespectfromthe
originalNewtonianposition.4Thus,in the following expositionof
the Newtoniantheoryof space and time,I shalldrawmaterialboth
fromNewton'sworksand fromClarke'sletters,assumingthatthere
is no essentialdifference in thetwostatements of thetheory(except
in one or twoinstanceswhereI shallpointout whatseemsto be a
slightvariance).
We maydivideNewton'stheoryofspaceand timeintoa number
of separateitems:
1) The basic claimof Newton'stheoryis that space is logically
priorto matterand has an existenceindependent of matter,and
thattimeis alsopriorto and existsindependently ofeventsorproces-
ses. That is to say,mattercouldnot existwithoutsomespace in
whichto reside,norcouldeventsorprocesses existwithoutsometime
in whichtheyoccurand through whichtheyendure.But spacewould
exist,in exactlythesamesenseas it doesnow,evenif therewereno
matterto occupyit; andtimewouldexisteveniftherewereno events
or processes to occurin it and to endurethrough it. As Newtonsays
in the " Scholium":
Absolute,
true,andmathematical time,ofitself,
andfrom itsownnature,
flows
equablywithoutrelation
toanythingexternal....
space,in its ownnature,
Absolute without relation to anything
external,
remains
alwayssimilar andimmovable.5
2) Newtondrawsa distinction betweenthe volumeor extension
ofa body(i.e.,thesize and shapewhichare semi-permanent proper-
tiesofthebody) and theregionofspacewhichthatbodyhappensto
occupyat a givenmoment orfora givenperiodoftime. It is obvious
3 The relevantpassagesare quotedin Alexander,
op. cit.,152-60.
4 According
to Cassirer(op. cit.,366n.), "the fullauthenticity
of the Clarke
papersis provedby the factthatthe outlinesof Clarke'sreplieshave beenfound
amongNewton'smanuscripts."
5 Quoted fromAlexander,op. cit., 152.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ 'S THEORY OF SPACE AND TIME 51

thatthesamebodycouldoccupydifferent regionsofspaceat different


timeswithoutany alterationin its volumeor extension.Thus,the
extensionof a givenbodyis a property of thatbodyin a sensein
whichthespacethatit occupiesat anygiventimeis not. A finitely-
extendedbodyis in a limitedregionof space,but it doesnotpossess
thatregionas a property.The bodymerely has theproperty ofbeing
in theregion. Regionsofspacearenotproperties ofbodiesbutparts
of the one infinite and absolutespace in whichall finitely-extended
bodiesexist. And even if it werethe case (as Newtonand Clarke
denied) that the whole extentof infinitespace were filledby a
" plenum" of matter,space wouldstillbe something logicallyprior
to and existing independently ofthatinfinitely-extended plenumand
nota property ofit. The infinitely-extendedplenumwouldstillbe in
spacein thesamesensein whichfinitely-extended bodiesarein it."
Clarkeholdsa similarposition, mutatismutandis, withrespectto
the relationship betweeneventsor processesand time,althoughthe
detailsofthispositionarenotfullystated. In general, it willsuffice
to say that on the Newtoniantheorywe mustdistinguish between
thedurationofa processand theparticular intervalof absolutetime
whichthatprocessin factoccupies.
3) Space is continuous,i.e.,thepartsof spaceare notseparable.
4) Space is actuallyinfinite.This is a directconsequence of the
factthatEuclideanspaceis homogeneous and isotropic, forif space
werefiniteit wouldnotbe homogeneous, in thesensethateachpoint
of space wouldhave its ownuniqueand unduplicatable relationship
to theboundaries ofspace. Thereis no evidencethateitherNewton
or Clarke(or Leibniz,forthatmatter)everconsidered thepossibility
that the geometry of space mightbe non-Euclideanand therefore
possiblyfinite.
5) The bodiesand eventswhichoccupyspace and timeare per-
ceptible,but (on Newton'stheory)theindividualpointsofspaceand
momentsof time are not themselvesindependently perceptible.
Nevertheless, Newtonmaintained that,sinceabsolutetime" ofitself,
and fromits ownnature,flowsequablywithoutrelationto anything
external," it is a meaningful and significant possibilitythatthe uni-
versemighthave been createdat a different momentthan that at
whichit was in factcreated. The samepossibility appliesto space.
If we assumethatthematerialuniverseis finitely-extended, thenit
is meaningful to assertthattheuniverseas a wholemighthavebeen
createdin a different regionof space,and also that at any timeit
mightbe moved(presumably by God) fromone regionof absolute
6 I take it thatthisis an implicit
denialof Aristotle's
viewthatprimematter
embraceseverything and thatspaceis theextensionofthisprimematter.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
52 KAITH EMERSON BALLARD

space to another. If thismotionwerenon-uniform (i.e., accelerated


withrespectto absolutespace) or if the universeweregivenan ab-
soluterotationabout an axis fixedin absolutespace,such absolute
motionwouldbe reflected by observableforceswithinthe universe,
even thoughthe internalstructure of the materialuniverseshould
remainthesame.
6) Newtondrawsa further distinction " betweenthesensedspace
and timein whichthesensedsecondary qualitiesappearandthequite
differentpostulatedmathematically definedspace and timein which
the unseenmassesand atomsof physicsare located."7 Newtonre-
postulatedspaceand timeas " absolute,true,
fersto thescientifically
and mathematical," and to the sensedspace and timeas "relative,
apparent,and common." 8
Whathas beenstatedin the above sectionsis Newton'sview of
"absolute,true,and mathematical " space and time. Now that we
have givena fairlydetailedaccountofNewton'stheory, we areready
to considerLeibniz'sarguments againstthistheory,as theyemerge
in thedevelopment ofLeibniz'sownrelationaltheory.
The principleof sufficientreasonis one of the fundamental as-
sumptionsfromwhichLeibniz'stheoryof space and timeis devel-
oped. Statedbriefly, the principleassertsthattheremustbe a suf-
ficientreasonwhyeverything is as it is andnototherwise (L. II. 1.).'1
Appliedto thedivinecreationand providence, theprinciple has been
statedin twodifferent ways: 1) God mustalwayshave a motivefor
acting;2) God mustalwaysact forthe best. The secondof these
statements is the so-called" principleof perfection," interpreted in
termsof God's actions. The important differences betweenthe two
statements becomeapparentin a detailedexamination of the use
whichLeibnizmakesof the principleof sufficient reasonand the
principleofperfection in thedevelopment ofhis theoryof spaceand
time.
Leibnizfrequently accusedthe Newtoniansof not understanding
theprinciple ofsufficientreasoncorrectly (L. III. 7-8). Clarkehim-
selfpersistedin interpreting the principleas beingmerelyidentical
withthecausalprinciple, viz.,thateveryeventhas a cause. Clarke,
accordingly,held that theremay be an event,the onlyreasonfor
7F. S. C. Northrop, "Leibniz'sTheoryofSpace,"thisJournalVII (1946), 426.
This wholearticleis devotedto an analysisof the consequences of thisdistinction
withrespectto Newton,Leibniz,and Einstein.
8 Quotedfrom Alexander, op. cit.,152.
9 CL. II. 1." refersto the firstparagraphof Leibniz'ssecondletterto Clarke.
Similarly," C. IV. 46." wouldreferto theforty-sixthparagraphof Clarke'sfourth
reply. I shall use this type of reference throughoutto cite passagesfromthe
correspondence.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ' S THEORY OF SPACE AND TIME 53

which is the mere will of God. But this contradictsLeibniz's own


statementof the principle,as Leibniz makes clear in the following
quotation,wherehe is referringto Clarke (L. III. 7.):
'Tis true,says he,thatthereis nothing reason. . . why
withouta sufficient
it is thus,ratherthanotherwise: but he adds,thatthissufficient
reason,is
oftenthesimpleor merewillofGod: as, whenit is askedwhymatterwas
notplacedotherwhere in space; thesamesituations ofbodiesamongthem-
selvesbeingpreserved.Butthisis plainlymaintaining, thatGodwillssome-
thing, withoutany sufficient
reason for hiswill.
In his nextletter,Leibnizmakeshis positionevenmoreexplicit(L.
IV. 19.):
Whentwo thingswhichcannotbothbe together, are equally good; and
neitherin themselves, withotherthings,has the
norby theircombination
one anyadvantageovertheother;God willproduceneither
ofthem.
In otherwords,God mustalwayshave a sufficient reasonforanyac-
tion,andHis merewillcannotitselfbe thatreason.
Fromthisstrongstatement of the principleof sufficient
reason,
Leibnizis able to deducetheprinciple oftheidentity ofindiscernibles
in a formwhichwillbe powerful enoughto refuteNewton'stheoryof
absolutespaceand time,as wellas theexistence ofidenticalparticles.
The mostcommonformulation oftheprinciple oftheidentity ofin-
discernibles is thattherecannotexisttwo indistinguishable entities
in nature. Leibnizdeducesthis(L. IV. 3.) fromtheassumption that
God couldhave had no sufficient reasonforarranging two identical
entitiesin one wayratherthanin thereverseway. Therefore, God
will not have createdtwo identicalentities. Leibnizconsidersthis
argumentas a conclusiverefutation of the existenceof Newtonian
particles.
WhenLeibnizuses the principleof the identityof indiscernibles
as an argument againstNewton'sabsolutetheoryof space and time,
he statestheprinciple in a slightlydifferent
way. Leibnizholdsthat
if it is impossibleto distinguishbetweentwo statesof affairs which
are allegedto be distinct, thenin factthereis onlyone and thesame
stateof affairs(L. IV. 6.). In otherwords,it is meaningless to talk
oftwoseparatestatesof affairs unlessone can specifysomediscerni-
ble difference betweenthem. Indiscernible (indistinguishable) en-
titiesor statesof affairs are identical. Leibnizuses thisinterpreta-
tion of the principleto showthat thereis no meaningin Clarke's
contention thatthe universeas a wholemighthave been createdin
a different regionof absolutespace or time,sincethe so-calleddif-
ference wouldin facthave no discernible effects.
In general,Leibniz'sargumentmay be summedup as follows:
Since Newton'sspace is homogeneous, thereare no qualitativedif-

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
54 KAITH EMERSON BALLARD

ferences bywhichto distinguish one partofit fromanother.Hence,


the parts,considered by themselves, independent of matter,violate
the Leibnizeanprincipleof the identityof indiscernibles (cf. L. III.
5.).
Clarkehad maintained 10that the materialuniverse(if finite)
mighthave been createdso as to occupya different regionof abso-
lutespaceand timethanthatwhichit in factdoesoccupy. In criti-
cizingthisposition,Leibnizattemptsto provethatthe assumption
of an absolutespace and timeleads to consequences whichdirectly
contradict theprinciple ofsufficient
reason. Leibniz'sargument may
be summarized as follows:
If the absolutetheoryof space is trueand if the materialuni-
verseis offiniteextent,thenit is possiblethatthematerialuniverse
as a whole (i.e., withoutany changein the mutualrelationsof its
parts) mighthave been createdby God in any one of an infinite
numberofdifferent regionsofspacefromthatin whichit has in fact
beencreated. But sincespaceis homogeneous, therecan be nothing
to distinguish thesevariousregionsof space,and therefore God can
have had no sufficient reasonto createthe materialuniversein one
regionratherthanin another. But on the absolutetheoryof space,
God musthave createdthe universein one regionratherthan an-
other,and thisconclusioncontradicts the principlethat God never
acts withoutsufficient reason. Therefore,the premissupon which
this conclusionis based (the existenceof absolutespace) mustbe
false.
On the otherhand,assumingthe relationaltheoryof space and
time,thisproblemneverarises,forthereis no antecedently existent
spacein whichtheworldcouldbe placedin a numberof alternative
ways. Space,forLeibniz,is merelythe orderof coexistence of mat-
ter,and is therefore establishedas a consequenceof God's creating
and arranging matter. On the relationaltheoryit is literallynon-
sensicalto suggestthat,withoutany difference in the mutualrela-
tionsof the variousparts,the materialuniverseas a wholecould
have beencreatedin a different regionof space,forspaceis nothing
but the orderor relationof the variouscoexistent partsof the uni-
verse,and comesinto existenceonlywhenthosepartsare created.
Therefore, Leibnizconcludedthat we oughtto replacethe rejected
absolutetheoryby a relationaltheoryofspace.
The samearguments apply,mutatismutandis, to the problemof
time(L. IV. 15.). If we assumean absolutetheoryof time,thenwe
are confronted withthepossibility thatGod mighthave createdthe
materialuniverseat a different momentof absolutetime. And we
10 Supra, 51 f.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ 'S THEORY OF SPACE AND TIME 55

must acknowledgethat He can have had no sufficient reasonto create


it at the momentHe did, ratherthan at some other time, since all
momentsof time are alike. But this conclusioncontradictsthe prin-
ciple of sufficientreason,and thereforethe assumptionof an absolute
timemustbe rejected. Once again, a relationaltheoryof timeavoids
the difficulties,for time (being merelythe order of events) begins
with the creationof the world,and God is not forcedto choose be-
tweenindifferent alternativesin the creationof the world,as He was
on the absolute theory. On the relationaltheory,the possibilityof
the materialuniverseas a wholebeing moved eitherin space or time
is strictlymeaningless,since thereis no space or time existinginde-
pendentlyof the orderof bodies and eventswithinthe universe.
Leibniz states these argumentsclearlyin his thirdletter (L. III.
5.):
I say then,thatifspacewas an absolutebeing,therewouldsomething hap-
pen forwhichit wouldbe impossiblethereshouldbe a sufficient reason.
Whichis againstmyaxiom. AndI proveit thus. Space is something abso-
lutelyuniform; and,without thethingsplacedin it,onepointofspacedoes
notabsolutely differ
in anyrespectwhatsoever fromanotherpointofspace.
Now fromhenceit follows,(supposingspace to be something in itself,be-
sidestheorderofbodiesamongthemselves,) that'tisimpossiblethereshould
be a reason,whyGod,preserving thesamesituations ofbodiesamongthem-
selves,shouldhave placedthemin space afterone certainparticularman-
ner,and nototherwise; whyeverythingwas notplacedthequitecontrary
way,forinstance, by changingEast intoWest. But ifspaceis nothing else,
but that orderor relation;and is nothingat all withoutbodies,but the
ofplacingthem;thenthosetwostates,theonesuchas it nowis,
possibility
theothersupposedto be thequitecontrary way,wouldnotat all differ from
one another.Theirdifference thereforeis onlyto be foundin ourchimeri-
cal supposition
oftherealityof space in itself.
To this argumentClarke replies,as he had before,holdingthat
God's will alone may be a sufficient
reason (C. III. 5.):
The uniformityofspace,doesindeedprove,thattherecouldbe no (external)
reasonwhyGod shouldcreatethingsin one place ratherthanin another:
butdoesthathinderhis ownwill,frombeingto itselfa sufficientreasonof
actingin any place,whenall places are indifferent
or alike,and therebe
goodreasonto act in someplace?
But Leibniz explicitlydenies that this is compatiblewiththe cor-
rect interpretation
of the principleof sufficient reason (L. IV. 19.):
Whentwo thingswhichcannotbothbe together, are equally good; and
neitherin themselves,
norby theircombinationwithotherthings,has the
one any advantageovertheother;God willproduceneither
ofthem.
In otherwords,God's will cannotbe a sufficient
reasonforany of His

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
56 KAITH EMERSON BALLARD

actions,according to Leibniz. God mustfindHis sufficient reasonin


something liketheprinciple ofperfection.
Leibnizalso formulated anotherargument(againstthe absolute
theoryof space and time) based upon the principleof the identity
ofindiscernibles. In his fourthletterLeibnizwrites(L. IV. 6.):
To suppose twothings is to suppose
indiscernible, thesamething under two
names.Andtherefore to suppose
thattheuniverse couldhavehadat first
another oftimeandplace,thanthatwhich
position itactually
had;andyet
thatall thepartsoftheuniverseshould havehadthesamesituation among
themselves,as thatwhichtheyactually had;sucha supposition,
I say,is an
impossiblefiction.
Leibnizseemsto be holdingherethatthe displacement of the uni-
verseas a wholethrough absolutespaceor timeis a meaningless no-
tion,becauseit has no discernible effects.If spaceis an orderofco-
existentthings,thenit is meaningless to speakof space as if it were
something independent of the totalityof coexistent things,through
whichthesethingscouldbe movedas a whole,withoutanyalteration
in the mutualrelationsof the partsof this whole. The difficulty
vanishesif we considerspace merelyas a relationbetweenbodies,
ratherthanas an absolutereality, independent ofbodies.
Leibnizwas surelycorrectin holding(L. IV. 13.) thata uniform
translation of the universeas a wholethroughabsolutespace would
have no discernible effectuponthe mutualrelationsof the partsof
theuniverse, and thattherefore sucha notionis scientificallyuseless,
ifnotmeaningless.But whenwe cometo consider non-uniform mo-
tion,and especiallyrotation,Leibniz'spositionbecomesmuchmore
difficultto defend.Newtonhad claimedthat the effects of " abso-
luterotation " wereobservableand if thisclaimis justifiable, Leib-
niz'sargument appearsto be weakened.Clarkeintroduces thisargu-
mentforthefirsttimein hisfourth letter(C. IV. 13.):
... themotion or restoftheuniverse [is not]thesamestate;anymore
thanthemotion orrestofa ship,is thesamestate,becausea manshutup
in thecabincannotperceive whether theshipsailsor not,so longas it
movesuniformly. Themotion oftheship,though themanperceives itnot,
is a realdifferentstate,andhas realdifferenteffects;
and,upona sudden
stop,itwouldhaveother realeffects;
andso likewisewouldan indiscernible
motion oftheuniverse.
If Clarkeis goingto admit(as he seemsto in the firstpartof this
quotation)thatthe man shutup in the cabincan perceiveno ob-
servableeffects of a uniform stateof motionof the ship,he has no
rightto assert that " the motion of the ship . . . is a real different
state,and has real different effects."At least,Leibnizwouldnot
yieldto sucha claim. But evenif Clarkeadmitsthat the uniform

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ 'S THEORY OF SPACE AND TIME 57

motionof theshiphas no observableeffects fortheman shutup in


the cabin,"he can stilllegitimately
assertthata suddenstopof the
shipwouldhave observableeffects, i.e., the man wouldfeela force
actingon him. And suchan observableeffect seemsto raisea real
problemforLeibniz'srelationaltheoryof space. WhatClarkeover-
looks,however, is thatthe man shutup in the cabinwouldnot be
able to determine whether thesuddenforceon himresultedfromthe
shipslowingdownfroma stateof uniform forward motion,or from
theshipaccelerating backwardsfroma stateofrest.'2In otherwords,
the " observableeffects" giveno definite information as to the state
"
ofmotionof themanrelativeto absolutespace."
Leibnizhimselfgave a slightly different answerto Clarke'sargu-
ment. He heldthatanydiscernible effect of a so-calledaccelerated-
wouldbe reflected
motion-relative-to-absolute-space in alterationsin
themutualrelationsofthevariouspartsof theuniverse, and thata
consideration of thesealterationswouldbe sufficient to explainthe
existenceof the new force,whichwouldotherwise have to be ex-
plainedby postulating an absoluteacceleration relativeto absolute
space (cf. L. V. 52.). This answeris certainlyplausibleenough.
Whetheror notit is scientifically
acceptablecannotbe decidedwith-
out a detailedexamination of Newton'sexperiment of the rotating
vessel of water; and such an examinationcannotbe undertaken
here.'3
In orderto provethat space is not a property
or attribute(as
Clarke maintained),'4Leibniz uses anotherreductioad absurdum
argument(L. IV. 8.):
If spaceis a property
or attribute,
it mustbe theproperty
ofsomesub-
stance.Butwhatsubstance willthatboundedemptyspacebe an affection
1Newton certainly
admitted
thismuch.
12 Cf. the examples
whichEinsteinusesin thisregard. As Einstein pointsout,
the forcecouldequallywellbe accountedforby a gravitating body,withoutany
" motionrelativeto absolutespace."
13 For relevantscientific
arguments see E. Mach, Scienceof Mechanics,trans-
latedfromtheoriginalGermanedition(Die Mechanik, 1883),by T. J. McCormack
(Chicago,1942),especially271-281.
14 Clarkelateradmitted(see Des Maiseaux'sprefaceto his 1720editionof the
Correspondence) thatin referringto space and timeas qualitiesor properties,he
maybe usingthe wordsin an unusualsense. See Alexander, op. cit.,xxix. R. S.
Wells,in an excellentarticlereviewingthe recentLeibnizliterature,also mentions
thispoint(" LeibnizToday,"ReviewofMetaphysics, X [1957],521). In pointof
fact,thereis no evidencethatClarke'stheory(thatspaceand timeare neither sub-
stancesnorrelations,butattributes)was everheldbyanyoneelseon theNewtonian
side. Newtonhimselfclearlyheld that space and timeare substances.Clarke's
theoryof space and timeas attributes is expoundedbrieflyin his BoyleLectures
(Works,II, 527-530).

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
58 KAITH EMERSON BALLARD

orproperty
of,which
thepersons
I amarguing
with, to be between
suppose
twobodies?
Clarkecannotconsistently defendthe thesisthat space is an at-
tributeand also affirmtheexistenceof vacua in nature,foras Leib-
niz pointsout (L. V. 38.): " if space is a property, whatthingwill
an emptylimitedspace (such as thatwhichmy adversary imagines
in an exhaustedreceiver)be theproperty of?"
This referenceto vacua in naturebringsus to a problemwhich
was sharplydebatedby Leibnizand Clarkein the correspondence.
The Newtoniantheories in generaldealtwithsolidmaterialparticles
in
moving absolutespace and time. Hence,betweensuchparticles
theremustbe vacua. In denyingthe existenceof vacua in nature,
Leibniz had difficulty in explainingaway the empiricalevidence
whichClarkehad citedin supportof the Newtoniantheoryof solid
particlesmovingin absolutespaceand time. Guericke's experiments
withthe air pumpand Torricelli's development of the mercury ba-
rometerseemedto offeralmostconclusiveproofof the existenceof
vacua. Andfromthefactthatdifferent liquidshavedifferent specific
gravitiesand offerdifferent degreesof resistanceto movingbodies
passingthroughthem,Clarkeconcludedthat certainliquidsmust
have morematterper unitvolumethanothers,and that therefore,
sincetheultimateparticles ofmatterareuniform, theremustbe some
emptyspacebetweenthem.
AgainsttheNewton-Clarke postulation of vacua,whichpurports
to be foundedon physicalevidence,Leibnizoffers threearguments,
all dependentuponmetaphysical considerations. First,he holdsthat
"the morematterthereis, themoreGod has occasionto exercisehis
wisdomandpower" (L. II. 2.). Therefore, thereis no vacuumat all,
sinceGod will have filledthe wholeuniversewithmatter. Leibniz
statesthissameargument in moredetailin his fourthletter(L. IV.
PS.):
I layit downas a principle,
thateveryperfection,
whichGodcouldimpart
tothings withoutderogatingfromtheirotherperfections, hasactually
been
imparted to them.Nowletus fancya spacewhollyempty.God could
haveplacedsomematter in it,withoutderogating
in anyrespect from all
otherthings:thereforehe has actuallyplacedsomematter in thatspace:
therefore,
thereis nospacewholly empty: therefore
all is full.
This argument assumesthatmatter,as such,is moreperfectthana
vacuum,and Leibniz,in fact,makesthisassumptionexplicitwhen
he says (L. IV. PS.):
becausematteris moreperfect
thana vacuum. .. thereshouldbe as much
morematter thanvacuum, as theformer deserves
to havethepreference
beforethelatter.Butthentheremustbenovacuumat all; fortheperfec-
tionofmatteris tothatofa vacuum,as something
tonothing.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ' S THEORY OF SPACE AND TIME 59

This notion of perfectionis certainlypuzzling. In many passages


Leibnizdefinesdegreeofperfection in termsof amountof existence.
But whyis it thatmatter,as such,shouldbe moreperfect?Perhaps,
as Leibnizsays," themorematterthereis,themoreGodhas occasion
to exercisehiswisdomand power" (L. II. 2.). But Clarkeraisesthe
obviousobjectionto thiscontention (C. III. 9.):
thatdeterminate
I suppose, quantity ofmatter, whichis nowin theworld,
is themostconvenientforthepresent frame ofnature,orthepresentstate
ofthings:andthata greater (as wellas a less) quantityofmatter,
would
havemadethepresent frameoftheworldlessconvenient; andconsequently
wouldnothavebeena greater objectforGodtohaveexercised hisgoodness
upon.
In orderto meet this objection,Leibnizintroduces his second
argument-that therecanbe no sufficient reasonto limitthequantity
ofmatter(L. IV. 21-23.). If perfection consistsin the existenceof
the greatestnumberof compossibles, Leibniz'spositionseemsjusti-
fiable,forthentherecouldbe no sufficient reasonforGod to refrain
fromfillinganyvacuumwithmatter(providedthatGod was acting
on the principleof perfection).But suppose(contrary to Leibniz's
notionof perfection) thatperfection consistednotmerelyin the ex-
istenceofthegreatest numberofcompossibles, butin theexistence of
thatset of compossibles whichmostnearlyembodiedsomedesirable
property.In thiscase,it wouldbe possibleforGod's notionof per-
fectionto lead Him to createa limitedquantityof matter,forthe
sufficientreasonthat any greaterquantitywouldrealizeless fully
thatproperty whoseexistenceconstitutes perfection.In thissense,
theprinciple ofsufficientreasonaloneis notstrongenoughto justify
Leibniz'srejectionoftheconceptofa vacuum. He mustalso assume
someparticularconception of perfection,
and mustbe able to show
thatthisperfection can be realizedonlybytheexistence ofa plenum.
To equate perfection witha plenumby definition would,of course,
giveLeibnizan immediate victoryoverClarkein the argument over
the existenceof a vacuum,but to assumethat perfection can be
realizedonlythroughthe existenceof a plenumis to beg the very
pointat issue. In otherwords,Leibniz'spositionis self-consistent
and impregnable to attack,but forthisveryreasonhe has not yet
offered anyconclusive arguments againsttheNewton-Clarke position.
As faras thesetwometaphysical arguments are concerned, the con-
troversy overthe existenceof vacua in naturemustend in a stale-
mate.
Thereis,however, a further argument againstvacua basedon the
principleof the identity of indiscernibles.
In short,howis one part
ofa vacuumto be distinguished fromanotherpartofidenticalshape?

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
60 KAITH EMERSON BALLARD

In fact,howareanytworegionsofabsolutespacetobe distinguished?
If thereare no discernible differences,two equallyshapedpartsof
a vacuummustbe identical. But thisis absurd. Therefore, theas-
sumption ofvacua in naturemustbe abandoned, becauseit is incom-
patiblewiththeprinciple oftheidentity ofindiscernibles.
In theattemptto explainawaythescientific evidenceforthe ex-
istenceof vacua,Leibniz'spositionseemsto lose someof its initial
plausibility.In his fourthletter,Clarkecites the evidenceof an
"exhaustedreceiver " in which,thoughtheremaybe lightraysand
perhapsan exceedingly smallquantityofothermatter, " yetthewant
of resistanceplainlyshows,that the greatestpart of that space is
void of matter"15 (C. IV. 7.). Leibnizattemptedto answerthis
argumentby holding,with the Aristotelians and Cartesians,that
" thereis no vacuumat all in the tube or receiver;sinceglasshas
smallpores,whichthebeamsof light,the effluvia of the load-stone,
and otherverythinfluidsmaygo through " (L. V. 34.). Leibnizis
certainly correctin holdingthatan electro-magnetic or gravitational
fieldmay be presentin an apparentvacuum,but his reference to
" otherverythinfluids" is neverclarified or developed. He merely
assumesthatthereis a fluid(something like the 19th-century con-
ception of the aether) which has thesame " quantity"as othermat-
ter but not the same " grossness " (i.e., it is more "subtile" and
offerslittleor no resistance).Andthismatter," voidofheaviness,"
replacestheairwhichis removed froma receiver in thecreationofan
apparentvacuum. But Leibnizhad arguedat somelengthagainst
Descartesthatmatterwas to be definednotby extension but by its
inertialproperties, so thatmatterwithoutinertia,as is clearlyim-
pliedhere,seemsto be a self-contradictory or elsemeaningless notion
forLeibniz.
AgainstClarke'sotherargument forvacua, Leibnizis somewhat
moresuccessful.In his fourthletter(C. IV. 7.) Clarkehad main-
tainedthatthedifferent resistances
ofmercury and watercan be ac-
countedforonlyon the basis of theirdifferent densities,and con-
sequentlythatthereis of necessity morevacuumwherethereis less
resistance. To this Leibniz repliesthat "it is not so much the
quantityof matter,as its difficultyof givingplace [viscosity],that
makesresistance"(L. V. 34.). Leibnizis perfectly correctin hold-
Torricelli(1608-47)ofFlorence,
15 a pupilofGalileoandinventorofthebarom-
eter,had producedan apparentvacuumin the top of the tube of his mercury
barometer.Guericke(1602-86)of Magdeburg, whoinvented theair pump,is said
to have performed forEmperorFerdinandIII an experiment in whichtwohollow
copperhemispheres, fromwhichtheair had beenexhausted, couldnotbe separated
by thirtyhorses,fifteenpullingon each hemisphere.Leibniz,who corresponded
withGuericke abouttheairpumpin 1671-2,was wellawareofthedifficulties which
theseexperiments createdforhis theoryof space and theplenum.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ' S THEORY OF SPACE AND TIME 61

ing that viscosityin liquids is not a functionof densityalone, and


thus (from a modern,scientificpoint of view) he seems to emerge
the victorin this part of the controversy over the vacuum.
But, in fact,Leibniz neversuccessfullyansweredClarke's conten-
tion that the different specificgravitiesof mercuryand waternecessi-
tate vacua in water. Only by postulatingmattervoid of all inertial
propertiescould Leibniz protect his plenum against the empirical
evidence of Guericke,Torricelli,and Newton. And such a postula-
tion runs seriouslycounterto his originaldefinitionof matter.
In one of his earliestformulations of the relationaltheoryof space
and time, Leibniz says that " space is nothingbut the orderof ex-
istenceof thingspossibleat the same time,while timeis the orderof
existenceof things possible successively."16 By " things" Leibniz
means here " phenomena." Space and time are definitelynot rela-
tions among individual monads, althoughthe precise nature of the
relationshipbetweenspace and time and the " point of view " of the
monads is a question to which we shall returnlater.
So far we know merelythat space is the orderof coexistentphe-
nomena,while time is the orderof successivephenomena. Leibniz's
cleareststatementof this definitionis in his thirdletterto Clarke (L.
111.4.):
As formyownopinion, I havesaid morethanonce,thatI holdspaceto be
something merely[purement] relative,as timeis; thatI holdit to be an
orderofcoexistences,
as timeis an orderofsuccessions.For space denotes,
in termsofpossibility,
an orderofthingswhichexistat thesametime,con-
sideredas existing
together;
without enquiring intotheirmannerofexisting.
In other words,space is dependentupon the existenceof " things."
Space by itselfhas no independentreality. It is merelya systemof
relationsamong things. From this,Leibniz concluded (in directop-
positionto the Newtonianabsolute theoryof space and time as real,
independententities) that space and time are both ideal. What
Leibniz means by " ideal " is made clear in a passage fromhis fourth
letter(L. IV. 41.):
The authorcontends, that space does not dependupon the situationof
bodies. I answer:'tistrue,it doesnotdependuponsuchor sucha situation
of bodies;but it is thatorder,whichrendersbodiescapableof beingsitu-
ated,and bywhichtheyhavea situationamongthemselves whentheyexist
together;as timeis thatorder,withrespectto theirsuccessive
position.But
iftherewereno creatures, spaceandtimewouldbe onlyintheideasofGod.
Space and time are relationsamong bodies; they have no inde-
pendentrealityof theirown. They are purelyideal. " If therewere
16Leibniz:Philosophical
PapersandLetters,2 vols.,ed. LeroyE. Loemker(Chi-
cago,1956),874.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
62 KAITH EMERSON BALLARD

therewouldbe neithertimenorplace,and consequently


no creatures,
no actualspace" (L. V. 106.). But ifthereis no real and independ-
ent entity,distinctfromthe mind,to whichour idea of space cor-
responds,thenhowdoesthemindcometo formforitselftheidea of
space? Leibniztriesto answerthisquestionin hisfifth letter(L. V.
47.):
I willhereshow,howmencometo form to themselves thenotionofspace.
Theyconsider thatmanythings existat onceandtheyobserve in thema
certainorder ofco-existence,
accordingtowhich therelationofonething to
another is moreor lesssimple.Thisorder, is theirsituation
or distance.
Whenit happens thatoneofthoseco-existent thingschanges its relation
to a multitude whichdo notchangetheirrelation
ofothers, amongthem-
selves;andthatanother newlycome,acquiresthesamerelation
thing, to
theothers, as theformerhad;wethensay,it is comeintotheplaceofthe
former; and thischange, we call a motionin thatbody,wherein is the
immediate causeofthechange.
Let us state thisin a somewhatmorepreciseway. We are given
twoassumptions:1) themutualspatialrelationsbetweenthebodies
A, B, C, and D do notchangethroughout a certainintervaloftime;
2) if,duringthisinterval, thereis a changein themutualspatialre-
lationsof A-B-C-D,consideredas a rigidsystem,to certainother
bodies,thenthecauseofthischangeis notin therigidsystem.
From thesetwo assumptions, we can concludethat the bodies
A-B-C-D have constituted a fixedrigidsystemthroughout the in-
tervaloftime. Andfromthisnotionofa fixedrigidsystemit is pos-
sibleto definewhatwe meanby " spatiallocation" or " place." If X
has a certainrelationto a fixedrigidsystem(as defined above) and if
subsequently X changesthisrelation(the cause of the motion,or
change,beingin X), and if Y acquiresthe precisely similarrelation
to thefixedrigidsystemwhichX formerly had,thenwe say thatY
nowoccupiesthesameplace as X formerly occupied.
This definition of place seemsto me to be perfectly intelligible,
and adequate. Thereis, however,one seriousdiffi-
self-consistent,
cultywhichcriticsof Leibnizwerequick to pointout. By postu-
latingthatthecauseoftherelativemotionofX to A-B-C-Dmustbe
in X (if we are to define" place"), Leibnizwas in effect
admitting a
kindof absolutemotion. This chargeseemsto be evenmorejusti-
fiedin reference to anotherpassagein thefifth letter(L. V. 53.):
between
thereis a difference an absolute
truemotion
ofa bodyanda mere
withrespectto anotherbody. For whenthe
relativechangeofits situation
immediate cause ofthechangeis in thebody,thatbodyis trulyin motion.
There seems to be little doubt that this admissionof the distinction
betweenabsoluteand relativemotionraises grave difficulties
for

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ S THEORY OF SPACE AND TIME 63

Leibniz's relational theoryof space.'7 In order to define " place "


Leibniz was forced to introduce an assumption which in effect
amountedto the conditionthat the rigidsystemof referencedid not
move duringthe intervalin question. And the criterionby whichone
determinesif a body has moved is whetheror not the cause of the
change of relative position was in the body itself. But, as C. D.
Broad has pointed out, this criterionhardlyseems acceptable.
It seemsto methata primafacieobjectionto thiscriterion is that,accord-
ingto theFirstLaw ofMotion,thecauseofan acceleratedor a curvilinear
motionof a bodyneveris in that body itself. Leibnizwouldnot have
acceptedthisobjectionbecausehe had a generalmetaphysical principle that
all thechangesin any substanceare causedby its ownpreviousstates,and
thattheappearanceofinteraction between differentsubstancesis delusive.18
Whetherit is possible to constructa systemof dynamicsin accord-
ance with Leibniz's principleis a questionwhichwill probablynever
be answered. The factremainsthat in allowinga distinctionbetween
so-called absolute and merelyrelativemotion,Leibniz was tryingto
defenda positionforwhichtherewas, at his time, almost no scien-
tificsupport. The fact that the Special Theory of Relativity now
refersto preferredinertialframesand absolute rotations,withinthe
contextof a purelyrelationaltheoryof space, may tend to make us
a little more sympathetictoward Leibniz's attempt to formulatea
relationaltheoryof space, but it ought not to obscurethe fact that
thereare seriousweaknessesin many of the argumentswhichLeibniz
used to establishand defendsuch a theory.
However,if we grantLeibniz that " place ' can be definedsolely
in termsof relations,then the rest of his theorycan be developed
quite readily. " Space'" is definedas the " collectionof all places."
Leibniz sums up his definitionas follows(L. V. 47.):
in orderto have an idea ofplace,and consequently of space,it is sufficient
to considertheserelations,and the rulesof theirchanges,withoutneeding
to fancyany absoluterealityoutofthethingswhosesituationwe consider.
We have so far discussedLeibniz's relationaltheoryof space and
time primarilywithinthe contextof his lettersto Clarke. This em-
phasis is requiredby the fact that the Leibniz-Clarkecorrespondence
containsthe most thoroughand consistent(as well as mature) state-
mentof Leibniz's theoryof space and time. It also representsone of
the few opportunitieswhich Leibniz had to reply in detail to criti-
cismsmade of his theoryfroma Newtonianpoint of view. It there-
17 AIexandersuggests(op. cit.,pp. xxvi-xxvii)
thatthedistinction
betweenabso-
luteand relativemotionmaybe metaphysical andnotscientific,
butthisexplanation
completely ignoresthe intentof Leibniz'sargumentsin passagessuchas L. V. 47.
18 C. D. Broad,
Ethicsand theHistoryofPhilosophy (London,1952),184.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
64 KAITH EMERSON BALLARD

foreraisesproblemswhichare stillrelevantin contemporary discus-


sionsof absoluteand relationaltheoriesof space and time. But the
arguments whichLeibnizusedin hislettersto Clarke,ifconsidered by
themselves, do notprovidea whollyadequateunderstanding of Leib-
niz'stheory.The underlying metaphysical foundation forhis theory
of space and timeis neverexplicitly formulated or examinedin the
correspondence. In fact,thewholeLeibniz-Clarke controversy is con-
ductedon whatLeibniz,at least,wouldregardas a levelof limited
philosophical rigorand thoroughness. It is truethatthe discussion
is philosophicalas wellas scientific.But Leibnizneverhas theoppor-
tunityto statehismetaphysical positionfullyor accurately, and this
no doubtaffects theexposition ofhis theoryof spaceand time. For
example,Leibnizgrants, solelyforthesakeofargument, a numberof
basic assumptions whichhe explicitly rejectedin his morecarefully
developedmetaphysical writings.For instance,he admitsimplicitly
the realityof extendedsubstances, whereas,according to his formal
metaphysical doctrine, thenotionof an extendedsubstanceinvolves
a contradiction. Moreover, in his controversy withClarke,he seems
to granttherealityof (spatial) relationsamongextendedsubstances,
whereasin hismetaphysical writingsheholdsthattherecanbe noreal
relationsbetweensubstances, thateachmonadis essentially window-
less. Hence,spaceandtimecannotbe realrelations betweenmonads.
This latterdistinction meansthat Leibniz'srelationaltheoryof
spaceand timemustbe reinterpreted, ifit is to be considered within
thetotalcontextofhismetaphysical position.According to thisposi-
tion,whatwe misperceive as an extendedmaterialsubstanceis really
ourownconfused mentalstate,whichreflects froma particular " point
ofview" an infinite numberofunextended mentalsubstances.Like-
wise,whatwe misperceive as a relationbetweentwoso-calledmate-
rialsubstances is reallyto be accountedforby certaininternalquali-
ties in the mentalsubstances, whichwe misperceive as interrelated.
This interpretation of the windowlessness of themonadshas an iin-
portantbearingon Leibniz'stheoryof spaceand time.
Relationsbetweensubstancesare notreal,but purelyideal; that
is,theyarebuiltup outofqualitieswhichareinternal to themonads.
Thismeans,in particular, thattherelationofspatiallocationmustbe
accountedforbysomequalitywhichis internal to eachmonad.19The
mostobvioussolutionto thisproblemis to identify the quality of
spatiallocation(out of whichthe systemof spatialrelationsis con-
structed)withthe" pointofview" ofeachmonad. But thissolution
willnotdo as it stands,forthe " pointof view" represents a deter-
minablequalitywhichcharacterizes each mentalsubstance.There-
19Northropemphasizesthe importanceof this theoryof relationsfor Leibniz's
viewofspaceand time(op. cit.,434-5).

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions
LEIBNIZ S THEORY OF SPACE AND TIME 65

fore,it cannotaccountdirectly forspaceand time,whicharerelations


amongmaterialsubstances(so-called).
The following modificationseemstobe required.Justas weregard
ournotionsof materialsubstancesas resulting froma confused state
of perception of mentalsubstances, so likewisewe may regardthe
relations ofspaceand timebetweenthesematenral bodiesas resulting
froma confusedawarenessof the " pointof view" of mentalsub-
stances. In otherwords,the" pointofview" is a determinable qual-
ity of monads-eachmonadhavring its own particulardeterminate
value of thisquality. But in the state of confusedperception, in
are
whichmentalsubstances misperceived as material bodies, the
"pointofview" (whichnormally represents a relationamongmental
substances)is also confusedlyinterpreted as therelationofspaceand
timeamongmaterialbodies. Thisseemsto be thesimplestandmost
adequatewayto harmonize Leibniz'smetaphysical doctrineswithhis
relationaltheoryofspaceand time. The advantagesof thissolution
are obvious:
1) It givesto spaceand time(whichare relationsbetweenmate-
rial bodies) the samestatusas materialbodies,i.e.,theyare all less
real thanthemonads. In particular, space and timeare said to be
merely"'ideal."
2) It putsourknowledge of space and timeon thesamelevelas
that of materialbodies,i.e., our ideas of space and timeare to be
accountedforin the same way as ourideas of materialbodies(viz.
through theconfusion ofourperceptions).
3) It accordsadmirably withLeibniz'stheoryofrelations, accord-
ingto whichrelationsare constructed out of qualitiesin theentities
whichare related. In thiscase,the relationsof space and timeare
constructed out of the quality" pointof view" (confusedly inter-
preted).
4) It justifiesLeibniz'sarguments againstNewtonand Clarke,for
ifspaceis a relationwhichis constructed out ofa quality,thenspace
cannotbe an independent substance, as Newtonheld. Noris spacea
quality,forrelations,eventhoughconstructed outofqualities,arenot
themselves identicalwiththosequalities.
The attemptto interpret Leibniz'srelationaltheoryof space and
timewithinthetotalcontextof his metaphysics is a difficult
under-
taking,whichwouldrequireforadequatetreatment a paperoflength
comparable to thepresentone. The ideasoffered aboveareonlyten-
tativesuggestions forthesolutionofthisproblem.
Yale University.

This content downloaded from 66.77.17.54 on Sat, 4 Jan 2014 13:14:35 PM


All use subject to JSTOR Terms and Conditions

You might also like