极限思想外文翻译pdf.docx
《极限思想外文翻译pdf.docx》由会员分享,可在线阅读,更多相关《极限思想外文翻译pdf.docx(10页珍藏版)》请在冰点文库上搜索。
极限思想外文翻译pdf
极限思想外文翻译pdf
BSHMBulletin,2014
DidWeierstrass’sdifferentialcalculushavealimit-avoidingcharacter?
His
,,definitionofalimitinstyle
MICHIYONAKANE
NihonUniversityResearchInstituteofScience&Technology,Japan
Inthe1820s,Cauchyfoundedhiscalculusonhisoriginallimitconceptand
,,developedhisthe-orybyusinginequalities,buthedidnotapplythese
inequalitiesconsistentlytoallpartsofhistheory.Incontrast,Weierstrassconsistentlydevelopedhis1861lecturesondifferentialcalculusintermsofepsilonics.HislectureswerenotbasedonCauchy’slimitandaredistin-guishedby
theirlimit-avoidingcharacter.Dugac’spartialpublicationofthe1861lectures
makesthesedifferencesclear.Butintheunpublishedportionsofthelectures,
,,Weierstrassactu-allydefinedhislimitintermsofinequalities.Weierstrass’s
limitwasaprototypeofthemodernlimitbutdidnotserveasafoundationofhiscalculustheory.Forthisreason,hedidnotprovidethebasicstructureforthemodernedstyleanalysis.ThusitwasDini’s1878text-bookthatintroducedthe
,,definitionofalimitintermsofinequalities.
Introduction
AugustinLouisCauchyandKarlWeierstrassweretwoofthemostimportantmathematiciansassociatedwiththeformalizationofanalysisonthebasisoftheeddoctrine.Inthe1820s,Cauchywasthefirsttogivecomprehensivestatementsofmathematicalanalysisthatwerebasedfromtheoutsetonareasonablycleardefinitionofthelimitconcept(Edwards1979,310).Heintroducedvariousdefinitionsandtheoriesthatinvolvedhislimitconcept.Hisexpressionsweremainlyverbal,buttheycouldbeunderstoodintermsofinequalities:
givenane,findnord(Grabiner1981,7).Asweshowlater,Cauchyactuallyparaphrasedhislimitconceptintermsofe,d,andn0inequalities,inhismorecomplicatedproofs.ButitwasWeierstrass’s1861lectureswhichusedthetechniqueinallproofsandalsoinhisdefi-nition(Lutzen?
2003,185–186).
Weierstrass’sadoptionoffullepsilonicarguments,however,didnotmeanthatheattainedaprototypeofthemoderntheory.Modernanalysistheoryisfoundedonlimitsdefinedintermsofedinequalities.HislectureswerenotfoundedonCauchy’slimitorhisownoriginaldefinitionoflimit(Dugac1973).Therefore,inordertoclarifytheformationofthemoderntheory,itwillbenecessarytoidentifywheretheeddefinitionoflimitwasintroducedandusedasafoundation.
Wedonotfindtheword‘limit’inthepublishedpartofthe1861lectures.
Accord-ingly,Grattan-Guinness(1986,228)characterizesWeierstrass’sanalysisas
limit-avoid-ing.However,Weierstrassactuallydefinedhislimitintermsofepsilonicsintheunpublishedportionofhislectures.Histheoryinvolvedhislimitconcept,althoughtheconceptdidnotfunctionasthefoundationofhistheory.Basedonthisdiscovery,thispaperre-examinestheformationofedcalculustheory,notingmathematicians’treat-mentsoftheirlimits.Werestrictour
attentiontotheprocessofdefiningcontinuityandderivatives.Nonetheless,thisfocusprovidessufficientinformationforourpurposes.
First,weconfirmthatepsilonicsargumentscannotrepresentCauchy’slimit,
thoughtheycandescriberelationshipsthatinvolvedhislimitconcept.Next,weexaminehowWeierstrassconstructedanovelanalysistheorywhichwasnotbased
2013BritishSocietyfortheHistoryofMathematics
52BSHMBulletin
onCauchy’slimitsbutcouldhaveinvolvedCauchy’sresults.Thenweconfirm
Weierstrass’sdefinitionoflimit.Finally,wenotethatDiniorganizedhisanalysistextbookin1878basedonanalysisperformedintheedstyle.
Cauchy’slimitandepsilonicarguments
Cauchy’sseriesoftextbooksoncalculus,Coursd’analyse(1821),Resumedes
lecons?
donneesal’Ecoleroyalepolytechniquesurlecalculinfinitesimaltome
premier(1823),andLecons?
surlecalculdifferentiel(1829),areoftenconsideredasthemainreferen-cesformodernanalysistheory,therigourofwhichisrootedmoreinthenineteenththanthetwentiethcentury.
AtthebeginningofhisCoursd’analyse,Cauchydefinedthelimitconceptasfol-
lows:
‘Whenthesuccessivelyattributedvaluesofthesamevariableindefinitelyapproachafixedvalue,sothatfinallytheydifferfromitbyaslittleasdesired,thelastiscalledthelimitofalltheothers’(1821,19;Englishtranslationfrom
Grabiner1981,80).Startingfromthisconcept,Cauchydevelopedatheoryofcontinuousfunc-tions,infiniteseries,derivatives,andintegrals,constructingananalysisbasedonlim-its(Grabiner1981,77).
Whendiscussingtheevolutionofthelimitconcept,Grabinerwrites:
‘Thiscon-
cept,translatedintothealgebraofinequalities,wasexactlywhatCauchyneededforhiscalculus’(1981,80).Fromthepresent-daypointofview,Cauchydescribedratherthandefinedhiskineticconceptoflimits.Accordingtohis‘definition’—
whichhasthequalityofatranslationordescription—hecoulddevelopanyaspect
ofthetheorybyreducingittothealgebraofinequalities.
Next,Cauchyintroducedinfinitelysmallquantitiesintohistheory.‘Whenthesuc-
cessiveabsolutevaluesofavariabledecreaseindefinitely,insuchawayastobecomelessthananygivenquantity,thatvariablebecomeswhatiscalledaninfinitesimal.Suchavariablehaszeroforitslimit’(1821,19;Englishtranslation
fromBirkhoffandMerzbach1973,2).Thatistosay,inCauchy’sframework‘the
limitofvariablexisc’isintuitivelyunderstoodas‘xindefinitelyapproachesc’,
andisrepresentedas‘jxcjisaslittleasdesired’or‘jxcjisinfinitesimal’.
Cauchy’sideaofdefininginfinitesimalsasvariablesofaspecialkindwasoriginal,becauseLeibnizandEuler,forexample,hadtreatedthemasconstants(Boyer1989,575;Lutzen?
2003,164).
InCoursd’analyseCauchyatfirstgaveaverbaldefinitionofacontinuousfunc-tion.Then,herewroteitintermsofinfinitesimals:
[Inotherwords,]thefunctionfðxÞwillremaincontinuousrelativetoxinagivenintervalif(inthisinterval)aninfinitesimalincrementinthevariablealwayspro-ducesaninfinitesimalincrementinthefunctionitself.(1821,43;Englishtransla-tionfromBirkhoffandMerzbach1973,2).
Heintroducedtheinfinitesimal-involvingdefinitionandadoptedamodifiedversionofitinResume(1823,19–20)andLecons?
(1829,278).
FollowingCauchy’sdefinitionofinfinitesimals,acontinuousfunctioncanbedefinedasafunctionfðxÞinwhich‘thevariablefðxþaÞfðxÞisaninfinitely
smallquantity(aspreviouslydefined)wheneverthevariableais,thatis,thatfðxþaÞfðxÞapproachestozeroasadoes’,asnotedbyEdwards(1979,311).Thus,
thedefinitioncanbetranslatedintothelanguageofedinequalitiesfromamodernviewpoint.Cauchy’sinfinitesimalsarevariables,andwecanalsotake
suchaninterpretation.
Volume29(2014)53
Cauchyhimselftranslatedhislimitconceptintermsofedinequalities.Hechanged‘Ifthedifferencefðxþ1ÞfðxÞconvergestowardsacertainlimitk,forincreasingvaluesofx,(...)’to‘Firstsupposethatthequantitykhasafinite
value,anddenotebyeanumberassmallaswewish....wecangivethenumberhavaluelargeenoughthat,whenxisequaltoorgreaterthanh,thedifferenceinquestionisalwayscontainedbetweenthelimitske;kþe’(1821,54;English
translationfromBradleyandSandifer2009,35).
InResume,Cauchygaveadefinitionofaderivative:
‘iffðxÞiscontinuous,then
itsderivativeisthelimitofthedifferencequotient,
yf(x,i),f(x)
,xi
asitendsto0’(1823,22–23).Healsotranslatedtheconceptofderivativeas
follows:
‘Designatebydandetwoverysmallnumbers;thefirstbeingchoseninsuchawaythat,fornumericalvaluesofilessthand,[...],theratiofðxþiÞf
ðxÞ=ialwaysremainsgreaterthanf’ðxÞeandlessthanf’ðxÞþe’(1823,
44–45;Englishtransla-tionfromGrabiner1981,115).
TheseexamplesshowthatCauchynotedthatrelationshipsinvolvinglimitsorinfinitesimalscouldberewrittenintermofinequalities.Cauchy’sarguments
aboutinfiniteseriesinCoursd’analyse,whichdealtwiththerelationshipbetween
increasingnumbersandinfinitesimals,hadsuchacharacter.Laugwitz(1987,264;1999,58)andLutzen?
(2003,167)havenotedCauchy’sstrictuseoftheeN
characterizationofconvergenceinseveralofhisproofs.BorovickandKatz(2012)indicatethatthereisroomtoquestionwhetherornotourrepresentationusingedinequalitiesconveysmessagesdifferentfromCauchy’soriginalintention.But
thispaperacceptstheinter-pretationsofEdwards,Laugwitz,andLutzen?
.
Cauchy’slecturesmainlydiscussedpropertiesofseriesandfunctionsinthelimitprocess,whichwererepresentedasrelationshipsbetweenhislimitsorhisinfinitesi-mals,orbetweenincreasingnumbersandinfinitesimals.Hiscontemporariespresum-ablyrecognizedthepossibilityofdevelopinganalysistheoryintermsofonlye,d,andn0inequalities.Withafewnotableexceptions,allofCauchy’s
lecturescouldberewrit-tenintermsofedinequalities.Cauchy’slimitsandhis
infinitesimalswerenotfunc-tionalrelationships,1sotheywerenotrepresentableintermsofedinequalities.
Cauchy’slimitconceptwasthefoundationofhistheory.Thus,Weierstrass’sfull
epsilonicanalysistheoryhasadifferentfoundationfromthatofCauchy.
Weierstrass’s1861lectures
Weierstrass’sconsistentuseofedarguments
Weierstrassdeliveredhislectures‘Onthedifferentialcalculus’attheGewerbe
Insti-tutBerlin2inthesummersemesterof1861.Notesoftheselecturesweretakenby
1Edwards(1979,310),L