Essays on frege’s : basic laws of arithmetic first edition. edition ebert - The ebook with rich cont

Page 1


onFrege’s:basiclawsofarithmeticFirst Edition.EditionEbert

https://ebookmass.com/product/essays-on-freges-basic-lawsof-arithmetic-first-edition-edition-ebert/

Instant digital products (PDF, ePub, MOBI) ready for you

Download now and discover formats that fit your needs...

Frege on Language, Logic, and Psychology: Selected Essays

Eva Picardi

https://ebookmass.com/product/frege-on-language-logic-and-psychologyselected-essays-eva-picardi/ ebookmass.com

Laws of nature First Edition. Edition Ott

https://ebookmass.com/product/laws-of-nature-first-edition-editionott/

ebookmass.com

The Science of Meaning: Essays on the Metatheory of Natural Language Semantics First Edition Derek Ball

https://ebookmass.com/product/the-science-of-meaning-essays-on-themetatheory-of-natural-language-semantics-first-edition-derek-ball/ ebookmass.com

Unlocking dbt : Design and Deploy Transformations in Your Cloud Data Warehouse Cameron Cyr

https://ebookmass.com/product/unlocking-dbt-design-and-deploytransformations-in-your-cloud-data-warehouse-cameron-cyr/

ebookmass.com

Generalist Case Management: A Method of Human Service Delivery 5th Edition Marianne Woodside & Tricia Mcclam

https://ebookmass.com/product/generalist-case-management-a-method-ofhuman-service-delivery-5th-edition-marianne-woodside-tricia-mcclam/

ebookmass.com

Rehabilitation of the Hand and Upper Extremity, 2 Volume Set E Book: Expert Consult 6th Edition, (Ebook PDF)

https://ebookmass.com/product/rehabilitation-of-the-hand-and-upperextremity-2-volume-set-e-book-expert-consult-6th-edition-ebook-pdf/

ebookmass.com

Essential public affairs for journalists Sixth Edition James Morrison

https://ebookmass.com/product/essential-public-affairs-forjournalists-sixth-edition-james-morrison/

ebookmass.com

Biomass, Biofuels, Biochemicals: Green-Economy: Systems Analysis for Sustainability Ganti S. Murthy

https://ebookmass.com/product/biomass-biofuels-biochemicals-greeneconomy-systems-analysis-for-sustainability-ganti-s-murthy/

ebookmass.com

the Secret on Set Hayley Leblanc

https://ebookmass.com/product/the-secret-on-set-hayley-leblanc/

ebookmass.com

Ferri's Clinical Advisor 2021: 5 Books in 1 1st Edition

Fred Ferri

https://ebookmass.com/product/ferris-clinical-advisor-2021-5-booksin-1-1st-edition-fred-ferri/

ebookmass.com

EssaysonFrege’s BasicLawsofArithmetic

EssaysonFrege’s Basic LawsofArithmetic

PHILIPA.EBERTANDMARCUSROSSBERG

GreatClarendonStreet,Oxford,OX26DP, UnitedKingdom

OxfordUniversityPressisadepartmentoftheUniversityofOxford. ItfurtherstheUniversity’sobjectiveofexcellenceinresearch,scholarship, andeducationbypublishingworldwide.Oxfordisaregisteredtrademarkof OxfordUniversityPressintheUKandincertainothercountries

©theseveralcontributors2019

Themoralrightsoftheauthorshavebeenasserted

FirstEditionpublishedin2019

Impression:1

Allrightsreserved.Nopartofthispublicationmaybereproduced,storedin aretrievalsystem,ortransmitted,inanyformorbyanymeans,withoutthe priorpermissioninwritingofOxfordUniversityPress,orasexpresslypermitted bylaw,bylicenceorundertermsagreedwiththeappropriatereprographics rightsorganization.Enquiriesconcerningreproductionoutsidethescopeofthe aboveshouldbesenttotheRightsDepartment,OxfordUniversityPress,atthe addressabove

Youmustnotcirculatethisworkinanyotherform andyoumustimposethissameconditiononanyacquirer

PublishedintheUnitedStatesofAmericabyOxfordUniversityPress 198MadisonAvenue,NewYork,NY10016,UnitedStatesofAmerica

BritishLibraryCataloguinginPublicationData

Dataavailable

LibraryofCongressControlNumber:2019945694

ISBN978–0–19–871208–4

PrintedandboundinGreatBritainby ClaysLtd,ElcografS.p.A.

LinkstothirdpartywebsitesareprovidedbyOxfordingoodfaithand forinformationonly.Oxforddisclaimsanyresponsibilityforthematerials containedinanythirdpartywebsitereferencedinthiswork.

MichaelHallett

13MathematicalCreationinFrege’s Grundgesetze

PhilipA.EbertandMarcusRossberg

EricSnyderandStewartShapiro

15Frege’sLittleTheoremandFrege’sWayOut

RoyT.Cook

16“HowdidtheserpentofinconsistencyenterFrege’sparadise?”411

CrispinWright

17Second-OrderAbstractionBeforeandAfterRussell’sParadox437

MatthiasSchirn

RichardKimberlyHeck

MichaelKremer

20ABriefHistoryofEnglishTranslationsofFrege’sWritings

MichaelBeaney

21Translating‘Bedeutung’inFrege’sWritings:ACaseStudyand CautionaryTaleintheHistoryandPhilosophyofTranslation588 MichaelBeaney

Foreword

GottlobFregepublishedhis GrundgesetzederArithmetik intwovolumes;the firstappearedin1893,thesecondin1903. Grundgesetze wastofulfillFrege’s ambitiontodemonstratethatarithmeticandanalysisarereducibletologic, andthusestablishaviewthatwenowcall‘logicism’.Inpreviouswork,in particularin Begriffsschrift (1879)and DieGrundlagenderArithmetik (1884), Fregeprovidessomeoftheformalandphilosophicalfoundationsforhislogicism.Morespecifically,hisfirstbookcontainstheinitialformulationofthelogicalsystem—theeponymous Begriffsschrift—whilehissecondbookeschews formulaealtogetherandoffersaphilosophicalfoundationforhislogicistpositioninthephilosophyofmathematics.Bothpublicationsare,intheirown right,groundbreaking.

Begriffsschrift constitutes“perhapsthegreatestsinglecontributiontologicevermadeanditwas,inanyevent,themostimportantadvancesince Aristotle.”1 Ontheotherhand,thephilosophicalmethodologyof Grundlagen,inparticularitsanalyticwritingstyle,ledmanytothinkofFregeasone ofthefoundingfathersofanalyticphilosophy.MichaelDummettconsiders Grundlagen “themostbrilliantpieceofphilosophicalwritingofitslength everpenned”.2 Inthisworkscholarslocatetheoriginsofthelinguisticturn inphilosophy,whichshapedmuchoftwentieth-centuryphilosophy.

However,Frege’sstandinginthehistoryofanalyticphilosophydoesnot simplyderivefromthesetwogroundbreakingbooks,butitisalsobasedon aseriesoflecturesandarticlesthatFregewrotejustbeforethepublication ofthefirstvolumeof Grundgesetze. FunctionundBegriff (1891),‘ÜberSinn undBedeutung’(1892),and‘UeberBegriffundGegenstand’(1892)make foratrilogyofGodfather-likeproportionswithpartIIbeing,ofcourse,one ofthemostwidelyreadarticlesinphilosophy.Itsinfluenceinthephilosophy oflanguageandinlinguisticswouldbehardtoexaggerate.

Howthendoes Grundgesetze,aworkthatFregewithoutdoubtintended tobehis magnumopus,fitintothelistofFrege’sphilosophicalandlogical achievement?Mostscholarshave,untilrecently,viewedFrege’s Grundgesetze

1AlexanderGeorgeandRichardKimberlyHeck(2000),‘Frege,Gottlob’,inEdwardCraig (ed.), ConciseRoutledgeEncyclopediaofPhilosophy,LondonandNewYork:Routledge,page296. (Orig.publ.underthename“RichardG.Heck,Jr”.)

2MichaelDummett(2007),‘IntellectualAutobiography’,inRandallE.AuxierandLewis EdwinHahn,eds. ThePhilosophyofMichaelDummett,vol.XXXIof TheLibraryofLivingPhilosophers,ChicagoandLaSalle,Ill.:OpenCourt,page9.

ascontainingforthemostparts,merelytheformaldetailsofhislogicistprojectwhichwenowknowtohavefailed.Frege’sderivationsoffundamental arithmeticalprinciples,whichcoverabouthalfofeithervolume,arebased onaninconsistentaxiom:theinfamousBasicLawV.Itisthusperhapsno surprisethatlittleattentionwaspaidtoFrege’spredominatelyformalwork.

Interestingly,however,itis Grundgesetze thatcontainsFrege’smostforcefulandmostfamousrejectionofpsychologisminlogic,andtherelevantpassagesfromtheForewordofthefirstvolumewerethefirstpiecesofFrege’scorpustobetranslatedintoEnglish.Moreover,Wittgenstein,Russell,Jourdain, Peano,andotherphilosophersandmathematiciansofthetimewhoseriously engagedwithFrege’sworkpaidcloseattentionto Grundgesetze.Nonetheless, throughouttherestofthetwentiethcenturymoreandmoreemphasiswas placedonFrege’sotherwritings—writingsthatareindependentofthefailureofhislogicalfoundations,orindeedindependentofhislogicismmore generally.

In1983,CrispinWrightpublished Frege’sConceptionofNumbersasObjects,inwhichhereconstructedFrege’sderivationoftheaxiomsofarithmetic fromwhatwouldcometobeknownas“Hume’sPrinciple”,or“HP”.Wright alsoconjecturedtherethatHP,unlikeBasicLawV,isconsistent.3 ThisinspiredGeorgeBoolostolookmorecarefullyatFrege’sinformalderivationsin Grundlagen,andalsoattheformalworkin Begriffsschrift. 4 Boolos’sstudent RichardKimberlyHeckthenbeganinvestigatingFrege’stechnicalworkon arithmeticin Grundgesetze. 5 Aroundthesametime,PeterSimonsandMichael DummettalsopublishedtheirdiscussionsofFrege’sformalworkonthereal numbersinPartIIIof Grundgesetze. 6 TheincreasinginterestinFrege’sproofs wasaccompaniedbyarenewedinterestinFrege’sphilosophyofmathematics, givenalsotheriseofaviewlabelled‘neo-logicism’,‘neo-Fregeanism’,orsimply ‘abstractionism’,aschampionedbyCrispinWrightandBobHale,7 andthe

3ForthefirstpublishedconsistencyproofsofHPseetworeviewsof Frege’sConceptionof NumbersasObjects:onebyJohnP.Burgess(1984, PhilosophicalReview 93:638–40),andtheother byAllenHazen(1985, AustralasianJournalofPhilosophy 63:251–4);andseeBoolos(1987),as citedinnote4below.

4GeorgeBoolos(1987),‘TheConsistencyofFrege’sFoundationsofArithmetic’,inJudith JarvisThomson(ed.), OnBeingandSaying:EssaysforRichardCartwright,Cambridge,Mass.: MITPress,pages3–20;GeorgeBoolos(1985),‘Readingthe Begriffsschrift’, Mind 94:331–44.

5RichardKimberlyHeck(1993),‘TheDevelopmentofArithmeticinFrege’s Grundgesetzeder Arithmetik’, JournalofSymbolicLogic 58:579–601(originallypublishedunderthename“Richard G.Heck,Jr”).Seealsotheirmorerecentbooks Frege’sTheorem (2011)and ReadingFrege’s Grundgesetze(2012),bothOxford:ClarendonPress.

6PeterSimons(1987),‘Frege’sTheoryoftheRealNumbers’, HistoryandPhilosophyofLogic 8:25–44;MichaelDummett(1991), Frege:PhilosophyofMathematics.London:Duckworth.

7Comparehere,e.g.,CrispinWright(1983), Frege’sConceptionofNumbersasObjects,Aberdeen:AberdeenUniversityPress;BobHale(1987), AbstractObjects,Oxford:BasilBlackwell; BobHaleandCrispinWright(2001), TheReasons’ProperStudy:EssaystowardsaNeo-Fregean PhilosophyofMathematics,Oxford:ClarendonPress;PhilipA.EbertandMarcusRossberg,eds. (2016) Abstractionism:EssaysinPhilosophyofMathematics,Oxford:OxfordUniversityPress.

workbyMichaelDummett.YetabroaderengagementwithFrege’s magnum opus facedtheobstaclethatonlypartsofthetwovolumeshadbeentranslated. TogetherwithCrispinWright,westartedworkonanewtranslationof Grundgesetze in2003,and,withthehelpofmanydistinguishedFregescholars,the firstcompleteEnglishtranslation, BasicLawsofArithmetic,volumesIandII, appearedin2013.8

ThepresentvolumeisthefirstcollectionofessaysthatfocusesonFrege’s Grundgesetze andaimstohighlightthetechnicalaswellasphilosophicalrichnessofFrege’smajorwork.Thecompanionbringstogethertwenty-twoFrege scholars,whosecontributionsdiscussawiderangeoftopicsarisingfromboth volumesof Grundgesetze.Andsomaythiscollection,evenifbelatedly,contributetoarenaissanceof Grundgesetze andhelptoestablishthisworkasone ofFrege’smanymasterpieces.

Wewouldliketothankoureditor,PeterMomtchiloffatOxfordUniversityPress,forhissupportandhispatience.ThankstoDeniseBannerman formeticulousproof-reading.Typesettinganearly700-pagevolumeinLATEX provedtobeyetanotherbigchallenge,andwewouldliketothankColin McCullough-BennerandAndrewParisiwhoprovidedinvaluableassistance withthisarduoustask.WewouldalsoliketothankColinforhisworkonthe Indexandhisadditionalhelpwithcopyeditingandproof-reading.

AndthankstoCrispin,forgettingitallstarted,andforeverything.

PhilipEbert,Stirling,Scotland MarcusRossberg,Storrs,Conn.,USA

8Formoredetailsonearliertranslationsaswellasthefirstcompletetranslationof Grundgesetze, andhowthelattercameabout,consultCrispinWright’sForewordandourTranslators’IntroductioninGottlobFrege(2013), BasicLawsofArithmetic.DerivedUsingConcept-Script,volumesI andII,trans.anded.PhilipA.EbertandMarcusRossbergwithCrispinWright.Oxford:Oxford UniversityPress.

Contributors

MichaelBeaney isProfessorofHistoryofAnalyticPhilosophyattheHumboldt-UniversitätzuBerlinandProfessorofPhilosophyatKing’sCollegeLondon.

PatriciaA.Blanchette isGlynnFamilyHonorsChairofPhilosophyatthe UniversityofNotreDame.

RoyT.Cook isCLAScholaroftheCollegeandJohnM.DolanProfessorof PhilosophyattheUniversityofMinnesota,TwinCities.

PhilipA.Ebert isSeniorLecturerinPhilosophyattheUniversityofStirling.

MichaelHallett isJohnFrothinghamProfessorofLogicandMetaphysicsat McGillUniversity,Montreal.

RichardKimberlyHeck isProfessorofPhilosophyatBrownUniversity,in Providence,RhodeIsland.

KevinC.Klement isProfessorofPhilosophyattheUniversityofMassachusettsAmherst.

MichaelKremer istheMaryR.MortonProfessorofPhilosophyandinthe CollegeattheUniversityofChicago.

ØysteinLinnebo isProfessorofPhilosophyattheUniversityofOslo.

RobertC.May isDistinguishedProfessorofPhilosophyandLinguisticsat theUniversityofCalifornia,Davis.

WalterB.Pedriali isAssociateLecturerinPhilosophyattheUniversityofSt Andrews.

ErichH.Reck isProfessorofPhilosophyattheUniversityofCaliforniaat Riverside.

MarcusRossberg isAssociateProfessorofPhilosophyattheUniversityof Connecticut.

MatthiasSchirn isProfessorEmeritusofPhilosophyattheUniversityofMunichandamemberoftheMunichCenterforMathematicalPhilosophy.

Contributors xi

StewartShapiro isO’DonnellProfessorofPhilosophyatTheOhioStateUniversity,DistinguishedVisitingProfessorattheUniversityofConnecticut,andDistinguishedPresidentialFellowattheHebrewUniversityof Jerusalem

PeterSimons, FBA,MRIA,isProfessorEmeritusofPhilosophyatTrinity CollegeDublin.

EricSnyder isPostdoctoralFellowattheMunichCenterforMathematical Philosophy,LMUMunich,andAssociateProfessorofPhilosophyat AshokaUniversity.

WilliamStirton worksasanadministrativeassistantforEdinburghLeisure Ltd.

JamieTappenden isProfessorofPhilosophyattheUniversityofMichigan, AnnArbor.

KaiF.Wehmeier isDean’sProfessorofLogicandPhilosophyofScienceand ofLanguageScienceattheUniversityofCalifornia,Irvine.

JoanWeiner isProfessorofPhilosophyatIndianaUniversity.

CrispinWright, FBA,FRSE,FAAAS,isProfessorofPhilosophyatNewYork University,ProfessorofPhilosophicalResearchattheUniversityofStirling,andRegiusProfessorofLogicEmeritusatAberdeenUniversity.

TheBasicLawsofCardinalNumber

RichardKimberlyHeck

Fregebeginshis GrundgesetzederArithmetik asfollows:

Inmy GrundlagenderArithmetik,Iaimedtomakeitplausiblethatarithmeticisa branchoflogicandneedstorelyneitheronexperiencenorintuitionasabasisforits proofs.Inthepresentbookthisisnowtobeestablishedbydeductionofthesimplest lawsofcardinalnumberbylogicalmeansalone. (Grundgesetze I,1)

Theplausibilityofwhatisnowcalled‘logicism’wassupposedtohavebeen establishednotonlybythephilosophicalargumentsin DieGrundlagen but, moreimportantly,bytheproofsofbasicarithmeticalprinciplesthatFrege sketchesin§§70–83.Butthecharacterofthoseargumentsleftalargelacuna: Idonotclaimtohavemadetheanalyticcharacterofarithmeticalpropositionsmore thanplausible,1 becauseitcanalwaysstillbedoubtedwhethertheyarededuciblesolely frompurelylogicallaws,orwhethersomeothertypeofpremissisnotinvolvedatsome pointintheirproofwithoutournoticingit.Thismisgivingwillnotbecompletely allayedevenbytheindicationsIhavegivenoftheproofofsomeofthepropositions; itcanonlyberemovedbyproducingachainofdeductionswithnolinkmissing,such thatnostepinitistakenwhichdoesnotconformtosomeoneofasmallnumberof principlesofinferencerecognizedaspurelylogical. (Frege,1884,§90)

Buthowcanwebesurethatnolinkismissing?Thatproblemwastheone thathadledtoFrege’sinterestinlogic,ashemakesexplicitin Begriffsschrift: [W]edividealltruthsthatrequirejustificationintotwokinds,thoseforwhichthe proofcanbecarriedoutpurelybymeansoflogicandthoseforwhichitmustbesupportedbyfactsofexperience.…[W]henIcametoconsiderthequestiontowhichof thesetwokindsthejudgmentsofarithmeticbelong,Ifirsthadtoascertainhowfar onecouldproceedinarithmeticbymeansofinferencesalone,withthesolesupportof thoselawsofthoughtthattranscendallparticulars.…Topreventanythingintuitive frompenetratinghereunnoticed,Ihadtobendeveryefforttokeepthechainofinferencesfreeofgaps.Inattemptingtocomplywiththisrequirementinthestrictestway possibleIfoundtheinadequacyoflanguagetobeanobstacle;nomatterhowunwieldy

1Austintranslatesthisas‘probable’,butIhavealteredthetranslation,sinceFregeusesthe samewordhereasin Grundgesetze

theexpressionsIwasreadytoaccept,Iwaslessandlessable,astherelationsbecame moreandmorecomplex,toattaintheprecisionthatmypurposerequired.

(Frege,1879,5–6)

WhatFregeneededtodo,then,tofillthelacuna,wastoprovideformalproofs ofthevariouspropositionshehadonlyproveninformallyin DieGrundlagen. Fregeseemsalreadytohaveachievedsomethingalongtheselineseven beforehewrote DieGrundlagen.InaletterwritteninAugust1882,hesays: IhavenownearlycompletedabookinwhichItreattheconceptofnumberand demonstratethatthefirstprinciplesofcomputation,whichuptonowhavegenerally beenregardedasunprovableaxioms,canbeprovedfromdefinitionsbymeansoflogical lawsalone,sothattheymayhavetoberegardedasanalyticjudgementsinKant’ssense. ItwillnotsurprisemeandIevenexpectthatyouwillraisesomedoubtsaboutthis andimaginethatthereisamistakeinthedefinitions,inthat,tobepossible,they presupposejudgementswhichIhavefailedtonotice,orinthatsomeotheressential contentfromanothersourceofknowledgehascreptinunawares.Myconfidencethat thishasnothappenedisbasedontheapplicationofmyconcept-script,whichwillnot letthroughanythingthatwasnotexpresslypresupposed…(Frege,1980,99–100) Nonetheless,itwouldbemorethanadecadeafterFregewrotethosewords, andnineyearsafterthepublicationof DieGrundlagen,beforeheactually wouldprovidethegap-freeproofshehadpromised.2

ThoseproofsarecontainedinPartIIof Grundgesetze.PartIofthebookis devotedtothe‘ExpositionoftheConcept-Script’,thatis,totheexplanation oftheformallanguageinwhichFrege’sproofswillbestatedandoftheformal theoryinwhichtheywillbedeveloped,thatis,ofthebasiclawsandrulesof inferenceofhissystem(seeHeck,2012,PartI).PartIIcontainsthe‘Proofs oftheBasicLawsofCardinalNumber’.AlloftheDedekind–Peanoaxioms forarithmeticareproventhere,includingthestatementthateverynatural numberhasasuccessor,whoseproofFregehadsketchedin§§82–3of Die Grundlagen.

ItisthereforeclearthatPartIIof Grundgesetze playsanimportantrole inFrege’sphilosophyofmathematics.Thatmakesitreallyquiteastonishing thatithasonlyrecentlybeenpublishedinEnglishtranslation(Frege,2013).3 Thereare,ofcourse,severalreasonswhythatis.Frege’sformalsystemis,asis well-known,inconsistent,sinceRussell’sParadoxisderivableintheconceptscriptfromFrege’sBasicLawV.OnemightthereforesupposethatFrege’s proofscanbeoflittleinterest,sinceanythingcanbeproveninaninconsistent system.ThereisalsotheproblemofFrege’snotation,whichisutterlyunlike thatusedbyanyotherauthorandwhichhasareputationforbeingdifficultto

2Idiscusssomeofthereasonsforthedelayinmyothercontributiontothisvolume(Heck, 2019).

3TherewerenotranslationsatallavailablewhenIstartedworkingon Grundgesetze inthe early1990s,untilJasonStanleyandIdidavery(very)roughoneinthesummerof1992.That wasusedinaseminarGeorgeBoolosandItaughttogetherin1993.Itwasoneofthefirstthings IputonmywebsitewhenIgotone,around1996,andatleastafewotherpeopleuseditin seminarsoftheirown.

read.4 Infact,however,thoseofuswhohavelearnedtoreaditknowthatitis notdifficulttoread.Rather,itsunfamiliaritymakesitsomethingofachallenge to learn toread.Andwehaveknownsincethemid-1980sthatFrege’ssystem, thoughinconsistent,isnotirremediablyinconsistent.Aswasfirstobserved byPeterGeach(1955,570),andemphasizedshortlythereafterbyCharles Parsons(1995,198),Frege’sownargumentsin DieGrundlagen makevery limitedappealtoBasicLawV,whichisthesourceoftheinconsistency.Law Visusedonlyintheproofofwhatisnowknownas“Hume’sPrinciple”,or HP:Thenumberof F sisthesameasthenumberof Gsif,andonlyif,the F s areinone–onecorrespondencewiththe Gs.Theremainderoftheargument appealsonlytoHP.And,asCrispinWright(1983,154–8)conjecturedand severalpeoplethenproved(Burgess,1984;Hazen,1985;Boolos,1998a),HP isconsistent.So,asWright(1983,§xix)showedindetail,Frege’sproofsin Die Grundlagen canbereconstructedinaconsistentsub-theoryoftheinconsistent theoryheimplicitlyassumes.

Theobviousquestion,whichGeorgeBoolosdirectedtomeinthesummer of1991,iswhethersomethingsimilarbutstrongeristrueof Grundgesetze.In Frege:PhilosophyofMathematics,whichwaspublishedthatyear,SirMichael Dummettseemstoassertthatthereis:

CrispinWrightdevotesawholesectionofhisbook…todemonstratingthat,ifwe weretotake[HP]asanimplicitorcontextualdefinitionofthecardinalityoperator, wecouldstillderiveallthesametheoremsasFregedoes.Hecouldhaveachievedthe sameresultwithlesstroublebyobservingthatFregehimselfgivesjustsuchaderivation ofthosetheorems.Hederivesthemfrom[HP],withnofurtherappealtohisexplicit definition. (Dummett,1991,123)

WhatBoolosaskedmewassimplywhetherthisistrue.Isettoreading Grundgesetze andsoondiscoveredthat,ifitwas,itwasgoingtotakeworktoshow it.Itiseasyenoughtoverifythat,afterprovingHP,Fregemakes“nofurther appealtohisexplicitdefinition”.Butthatisnotenough.ThecrucialquestioniswhetherFregemakesnofurtherappeal toBasicLawV,andhemost certainlydoes.HardlyapageofPartIIlackstermsforvalue-ranges,ofwhich extensionsofconceptsareaspecialcase,andthelogicallawgoverningvaluerangenamesis,ofcourse,BasicLawV.Moreprecisely,duetothedetailsof howFregeformalizesvariousnotions,almosteveryresultheprovesdepends uponhisTheorem1,whichisageneralizationoftheprincipleknownasnaïve comprehension: a ∈{x : Fx}≡ Fa

AndthatprincipleleadsdirectlytoRussell’sParadox,oncewetake Fξ tobe: ξ/ ∈ ξ and a tobe: {x : x/ ∈ x}. 5

4Nottomentiontypeset.IscannedtheformulasforthetranslationJasonandIdid.How theywerehandledinthenewtranslationiswell-relatedbyEbertandRossberg(Frege,2013, xxx–xxxii).

5Theorem1itselfisprovenfromBasicLawVandFrege’sdefinitionoftheanalogue,for value-ranges,ofmembership(Heck,2012,§1.2).

ButitalsoquicklybecamecleartomethatmanyoftheusesFregemakes ofvalue-rangescaneasilybeeliminated.Forexample,Fregealmostalways quantifiesovertheextensionsofconceptsinsteadofoverconcepts,sothatwe findthingslike:

But,asjustillustrated,thisiseasilyremedied.AndFrege’sotherusesofvaluerangesprovedtobeeliminableaswell.SoDummettturnedouttoberight,in spiritifnotindetail:Modulousesofvalue-rangesthatareessentiallyjustfor convenience,PartIIof Grundgesetze reallydoescontainaformalderivation ofaxiomsforarithmeticfromHP.

AndthereismuchmoreinPartII.Frege’sproofofaxiomsforarithmetic comprisesonlyaboutathirdofit.Intheremainder,Fregeprovesanumberof resultsconcerningfinitude,infinity,andtherelationshipbetweenthesetwo notions.WhenIexaminedthoseproofsclosely,itturnedoutthatFregeused LawVinthem,too,onlyforconvenience.AndtherewasmuchofphilosophicalinterestbothinFrege’sformalargumentsandintheinformaldiscussion ofthemcontainedinthe“Analysis”sections.

Mygoalinthischapter,then,istoprovideabriefoverviewofwhatFrege accomplishesinPartIIandtogivesomeindicationofthephilosophicaland historicalinterestthismaterialhas.Furtherdetails,andactualargumentsfor theinterpretiveclaimstobemadebelow,canbefoundinPartIIofmybook ReadingFrege’s Grundgesetze,ofwhichthischapterisakindofprécis.

AndsincethischapterismeanttoprovideanintroductiontoFrege’s formalworkonarithmetic,Iwillpresenthisresultsusingmodernnotation, soastomakethediscussionmoreaccessible.Iwillalsosilentlytranslateaway Frege’srelianceuponvalue-ranges,sincethatservesonlytoobscurehisaccomplishments.6

1.1THEPROOFOFHP

Frege’sfirsttaskin Grundgesetze istoproveHP,whichmaybestated,inmodernnotation,as:

Here,‘Nx : Fx’istoberead:thenumberof F s.

6IshallalsosilentlyaltersomeofthetranslationsfromwhichIquote,tomakethemuniform intheirterminology.

Frege’sformulationofHPmightinitiallyseemverydifferent,andnotjust becausehisnotationissodifferent.7 TranslatingFrege’snotationintoours,of course,hewouldwriteHPas:

Nx : Fx = Nx : Gx ≡∃R[Map(R)(F,G) ∧ Map(Conv(R))(G,F )]

Here,‘Map(R)(F,G)’,whichFregewouldwriteas‘f SgS⟩r’,8 mayberead: R mapsthe F sintothe Gs.Conv(R),whichFregewouldwriteas‘Ur’,isthe converseof R,definedtheobviousway:

Conv(R)(a,b) df ≡ Rba

SoHPitself,asFregewouldformulateit,saysthatthenumberof F sisthe sameasthenumberof Gsif,andonlyif,thereisarelationthatmapsthe F s intothe Gsandwhoseconversemapsthe Gsintothe F s. Themappingrelationitselfisdefinedasfollows(Grundgesetze I,§38):

Map(R)(F,G) df ≡ Func(R) ∧∀x(Fx →∃y(Rxy ∧ Gy))

Here,‘Func(R)’,whichFregewouldwriteas‘Ir’,meansthat R is“singlevalued”or“functional”.Ittooisdefinedtheobviousway(Grundgesetze I, §37):

Func(R) df ≡∀x∀y(Rxy →∀z(Rxz → y = z))

So R mapsthe F sintothe Gsjustincase R issingle-valuedandeach F is relatedby Rξη tosome G.Notecarefully: into,not onto.That R mapsthe F s intothe Gssays,ofitself,nothingwhatsoeverabouttherelativecardinalities ofthe F sandthe Gs:Aslongasthereisatleastone G,therewillalwaysbe arelationwhichmapsthe F sintothe Gs,inFrege’ssense,whateverconcept Fξ maybe.

ToseetherelationofFrege’sformulationofHPtotheusualone,unpack theright-handsideofhisversionHPusingthedefinitions:

∃R[Map(R)(F,G) ∧

Map(Conv(R))(G,F )]

7Theorem32,whichistheright-to-leftdirection,reads:

”u = ”v uS(vS⟩q) vS(uS⟩Uq)

Talkaboutdifferent!Green,Rossberg,andEbert(2015)discussFrege’snotationindetail.

8Hereandbelow,Ishalluseuppercaselettersforconceptsandrelationsandthecorresponding lowercaselettersfortheextensionsofthoseconceptsandrelations.

∃R[Func(R) ∧

∀x(Fx →∃y(Rxy ∧ Gy)) ∧ Func(Conv(R)) ∧

∀x(Gx →∃y(Conv(R)(x,y) ∧ Fy))]

∃R[∀x∀y(Rxy →∀z(Rxz → y = z)) ∧

∀x(Fx →∃y(Rxy ∧ Gy)) ∧ ∀x∀y(Conv(R)(x,y) →∀z(Conv(R)(x,z) → y = z)) ∧

∀x(Gx →∃y(Conv(R)(x,y) ∧ Fy))]

∃R[∀x∀y(Rxy →∀z(Rxz → y = z)) ∧

∀x(Fx →∃y(Rxy ∧ Gy)) ∧ ∀x∀y(Ryx →∀z(Rzx → y = z)) ∧ ∀x(Gx →∃y(Ryx ∧ Fy))]

WhatFregehasdoneisgroup‘∀x∀y(Rxy →∀z(Rxz → y = z))’and ‘∀x(Fx →∃y(Gy∧Rxy)’inthefirstconjunct,‘Map(R)(F,G)’,andtogroup ‘∀x∀y(Rxz →∀z(Ryz → x = y))’and‘∀y(Gy →∃x(Fx ∧ Rxy)’inthe secondconjunct,‘Map(Conv(R))(G,F )’.Wearemoreinclinednowadaysto grouptheconjuncts‘∀x∀y(Rxy →∀z(Rxz → y = z))’and‘∀x∀y(Rxz → ∀z(Ryz → x = y))’(R isaone–onefunction…)and‘∀x(Fx →∃y(Gy ∧ Rxy)’and‘∀y(Gy →∃x(Fx ∧ Rxy)’(…fromthe F sontothe Gs).So,in theend,thedifferencebetweenFrege’sformulationandoursismostlyoneof emphasis,thoughFrege’sformulationhassometechnicaladvantagesoverthe usualone(Heck,2012,§6.3).

Inmodernpresentations,‘Nx : Fx’istreatedasaprimitivenotiongovernedbyHP,whichisitselftreatedasanaxiom.Frege,byconstrast,means toproveHPandsodefines‘Nx : Fx’intermsofextensions.9 Now,in Grundgesetze,Fregetreatsextensionsasakindofvalue-range,buthisdefinitionof numberin Grundgesetze isotherwisethesameastheonegivenin§68of Die Grundlagen:10

Thenumberof F sistheextensionoftheconcept:is[theextensionof] aconceptthatisequinumerouswith F . Wecanformalizethisas:

Nx : Fx df ≡ ˆ x{∃G[(x =ˆ y(Gy) ∧ Eq(F,G)]}

9Whatmostobviouslycorrespondstoour‘Nx : Fx’isFrege’s‘”f’.Infact,however,as GregoryLandinipointedouttome,‘Nx : Fx’isdefinableinFrege’ssystemas‘” –εFε’.

10Frege’sdefinitiondoesnotcontainthebracketedoccurrenceofthephrase‘theextensionof’. Iargueelsewhere(Heck,2019,§18.1)thatitisnonethelesswhathemeans.

where‘Eq(F,G)’abbreviates: ∃R[Map(R)(F,G) ∧ Map(Conv(R))(G,F )]. Here‘ˆ x(Fx)’means:theextensionoftheconcept F ,andthenotionofextensionistobegovernedbyaversionofBasicLawV:11

TheproofofHPthenneedslittlemorethantheobservationthatEqisan equivalencerelation.

Infact,however,asMayandWehmeier(2019)pointoutintheircontributiontothisvolume,FregeneveractuallyprovesHPasabiconditional:He provesitstwodirections,butneverbotherstoputthemtogether.Therightto-leftdirectionisTheorem32,whichisthegoalofthefirstchapterofPartII, ChapterAlpha.12 Theleft-to-rightdirectionisTheorem49,whichisprovenin ChapterBeta.Theproofoftheformerisquitestraightforward,anditfollows theoutlinein§73of DieGrundlagen closely.Theproofneedsonlythetransitivityandsymmetryofequinumerosity.Theproofof(49)thatFregegivesis somewhatpeculiar,becauseitusesthedefinitionofnumberinamoreessentialwaythanitreallyshould.Thereis,however,asimplerproof,whichFrege musthaveknown,thatneedsonlythereflexivityofequinumerosity(Heck, 2012,§6.8).

Anotherpointworthnotingabouttheseproofsisthattheproofof(32) needsonlytheright-to-leftdirectionofLawV,whichFregecallsLawVaand whichisthe“safe”direction,whereas(49)needstheleft-to-rightdirection, LawVb,whichisthe“unsafe”direction,theonethatgivesrisetoRussell’s Paradox.Onreflection,thisshouldnotbesurprising,since(32)isthe“safe” directionofHP,whichbyitselfhasnosignificantontologicalconsequences, sinceitiscompatiblewiththerebeingonlyonenumber,sharedbyalltheconceptsthereare.Theorem49,ontheotherhand,istheontologicallyprofligate directionofHP,whichentailstheexistenceofinfinitelymanynumbers.

1.2THEAXIOMSOFARITHMETIC

AfterhavingprovenHP,Fregeturnshisattentiontotheproofsofvariousfundamentalprinciplesconcerningcardinalnumbers,includingwhatwenowcall theDedekind–Peanoaxioms,forwhichseeTable1.1.Here‘Nξ’isapredicate toberead‘ξ isanaturalnumber’,and‘Pξη’isapredicatetobereadas‘ξ immediatelyprecedes η inthenumber-series’.Toprovetheseaxioms,Frege mustofcoursedefinethearithmeticalnotionsthatoccurinthem.

11Assaid,Fregeactuallyworkswiththemoregeneralnotionofavalue-range,but,surprisingly,henevermakesuseofthemoregeneralnotion.Allthevalue-rangesinwhichheisactually interestedin Grundgesetze areextensionsofconcepts.

12Fregedoesnotcallthesedivisionschapters,butitseemstheobviousnameforthem.

1. N0

Table1.1.OneVersionoftheDedekind–PeanoAxioms

2. ∀x∀y(Nx ∧ Pxy → Ny)

3. ∀x(Nx →∃y(Pxy))

4. ¬∃x(Nx ∧ Px0)

5. ∀x(Nx →∀y∀z(Pxy ∧ Pxz → y = z))

6. ∀x∀y∀z(Nx ∧ Ny ∧ Pxz ∧ Pyz → x = y)

7. ∀F [F 0 ∧∀x(Nx ∧ Fx →∀y(Pxy → Fy)) →∀x(Nx → Fx)]

ThedefinitionsFregegivesin Grundgesetze arethesameastheonesgiven in DieGrundlagen.Fregedefineszero,whichhewritesas‘0’,asthenumberof objectsthatarenotself-identical(Grundgesetze I,§41;seeFrege,1884,§74): 0 df ≡ Nx : x = x

Hisdefinitionofpredecession,whichhewritesas‘mS(nSs)’,isasfollows (Grundgesetze I,§43):

mn df ≡∃F ∃x[Fx ∧ n = Nz : Fz ∧ m = Nz : (Fz ∧

Thatis, m precedes n if,asFregeputsitin DieGrundlagen,‘thereexistsa concept F ,andanobjectfallingunderit x,suchthattheNumberwhich belongstotheconcept F is n andtheNumberwhichbelongstotheconcept “fallingunder F butnotidenticalwith x”is m’(Frege,1884,§76).Weshall returntothedefinitionof‘Nξ’.

FregeprovesAxiom5inChapterBetaasTheorem71;Axiom6inChapter Γ asTheorem89;andAxiom4inChapterEpsilonasTheorem108.The proofsarestraightforward,butthereisaphilosophicaldiscussionthatoccurs duringFrege’sinformalexpositionoftheproofofTheorem71thatisofsubstantialinterest.ItconcernstheproofofTheorem66:

Fc ∧ Gb ∧ Nz : (Gz ∧ z = b)= Nz : (Fz ∧ z = c) → Nz : Fz = Nz : Gz (66) whichisthekeylemmaintheproofof(71).13 Toprove(66),whatwewant toshowisthat,ifthereisaone–onecorrelationbetweenthe Gsotherthan

13SupposethatPxy andPxw.Then,bythedefinitionof‘P’,thereare F and c suchthat:

Fc ∧ Nz : Fz = y ∧ Nz : (Fz ∧ z = c)= x andthereare G and b suchthat:

Gb ∧ Nz : Gz = w ∧ Nz : (Gz ∧ z = b)= x

SoNz : (Gz ∧ z = b)= x = Nz : (Fz ∧ z = c),and(66)nowimpliesthatNz : Gz = Nz : Fz, so w = y,andwearedone.

b andthe F sotherthan c,andif b isa G and c isan F ,thenthereisalsoa one–onecorrelationbetweenthe F sandthe Gs.Fregewrites:14

Ifoneweretofollowtheusualpracticeofmathematicians,onemightsaysomething likethis:wecorrelatetheobjects,otherthan b,fallingundertheconcept G,withthe objects,otherthan c,fallingundertheconcept F bymeansoftheknownrelation, andwecorrelate b with c.Inthisway,wehavemappedtheconcept G intotheconcept F and,conversely,thelatterintotheformer.So…thecardinalnumbersthatbelong tothemareequal.Thisisindeedmuchbrieferthantheprooftofollowwhichsome, misunderstandingmyproject,willdeploreonaccountofitslength.Whatisitthatwe aredoingwhenwecorrelateobjectsforthepurposeofaproof?(Grundgesetze I,§66) Thereis,ofcourse,nothingunusualaboutthesortofreasoningFregerehearses,buthehasaquestiontoraiseaboutit.Itisnot,ofcourse,thathe thinkssuchreasoningmightbeinvalid.Buthewantstoknowwhatjustifies it.

Fregefirstemphasizesthat,whenweestablishacorrelationinthissense, wedonotcreateanythingbut“merelybringtoattention,apprehend,whatis alreadythere”(compareFrege,1884,§26).Hethenseizestheopportunityto takeaswipeatpsychologism.Havingslainthatfamiliarfoe,Fregeconsiders thequestionhowwemightformulate“apostulate,inthestyleofEuclid”,that permitssuchcorrelations,answering:

[It]wouldhavetobeunderstoodthisway:‘Anyobjectiscorrelatedwithanyobject’ or‘Thereisacorrelationbetweenanyobjectandanyobject’.Whatthenissucha correlationifitisnothingsubjective,createdonlybyourmaking?However,aparticular correlationofanobjecttoanobjectisnotwhatcanbeatissuehere…;ratherwerequire agenusofcorrelations,sotospeak,whatwehavesofarcalled,andwillcontinueto call,arelation. (Grundgesetze I,§66) Sowhattheactofcorrelatingbringstoattentionisarelation,andFregegoes ontodiscusshowtheconcept-scriptallowsustospecifysuchrelationspreciselyandtoshowthattheyhavethevariouspropertiesweneedthemtohave. Infact,inthiscase,Fregegoesontosuggest,theusualreasoningencourages ustooverlookcertainsubtleties,whichemergeinhiscareful,rigorouspresentationoftheproof.

WhatisatissuehereisaquestionfundamentaltoFrege’slogicism.The informalargumenthementionsinvolvesjustthesortoftoxicmixofreason andintuitionthattendstoobscuretheepistemologicalstatusoftheresult proved.Onecrucialquestion,inparticular,ishowweknowthatthe“correlations”weneedactuallyexist.And,solongasonethinksofcorrelatingas somethingwedo,ratherthanofcorrelationsassomethingwediscover,this willtendtomakeonesupposethattheexistenceofcorrelationsdependssomehowuponmentalactivity.Frege’sviewis,ofcourse,different.Inhissystem, theexistenceofconcepts,relations,andthelikeisguaranteedbyRule9on

14Inhisexposition,Fregespeaksofsuchthingsas“the u-concept”,bywhichhemeansthe conceptwhosevalue-range u is.Ihavesilentlyreplacedsuchtalkwithdirectreferencestoconcepts.

thelistgivenin§48.Thisruleallowsfortheuniformreplacementofafree variable,ofarbitrarytype,byanywell-formedexpressionoftheappropriate type.Suchasubstitutionprincipleisequivalenttothecomprehensionscheme ofsecond-orderlogic

R∀x∀y[Rxy ≡ A(x,y)] whichassertstheexistenceofarelationco-extensivewithanygivenformulaof thelanguage(solongasitdoesnotcontainthevariable R free).SoFrege’sview isthat logicitself commitsustotheexistenceofanyrelationwecandescribe. Since“correlations”arejustrelations,itisthuslogic,andnotpsychology,that affirmstheirexistence.

Toreturntotheproofsoftheaxioms,then,whatremaintobeprovedatthe endofChapterEpsilonaretheaxiomsconcerningthenotionofanatural(or finite)number.SowenowneedtoconsiderFrege’sdefinitionofthatnotion. Itis,again,essentiallythesameasthatgivenin DieGrundlagen,andituses Frege’sdefinitionoftheso-called ancestral,whichheintroducesin§23of Begriffsschrift.Givenarelation Q,wesaythataconcept F is hereditaryinthe Q-series justincase,whenever x is F ,eachobjecttowhich Q relatesitis F :

x∀y(Fx ∧ Qxy

Wenowsaythatanobject b follows anobject a inthe Q-seriesjustincase b fallsundereveryconceptthatishereditaryinthe Q-seriesandunderwhich eachobjecttowhich Q relates a falls.Formally,writing‘Q∗ab’for‘b follows a inthe Q-series’,Frege’sdefinitionofthe strong ancestral,15 whichhewrites as‘aS(bSMq)’,is(Grundgesetze I,§45):

∗ab df ≡∀F [∀x(Qax → Fx) ∧∀x∀y(Fx ∧ Qxy → Fy) → Fb]

Fregethendefinesthe weak ancestral,whichhewritesas‘aS(bSRq)’,thus (Grundgesetze I,§46):

Theconcept Nξ isthendefinableasP∗=0ξ.Soanobjectisanaturalnumber justincaseitbelongstotheP-seriesbeginningwith 0.Fregehasnospecial symbolforthis.Hedoes,however,regularlyread‘P∗=0ξ’as‘ξ isafinitenumber’(e.g.,in Grundgesetze I,§108).

Axioms1and2thenfollowfromgeneralfactsabouttheancestral,Theorems140and133,respectively.Famously,Axiom7,theinductionaxiom, alsofollowsfromFrege’sdefinition.Mattersaremorecomplicatedthanoften seemstobesupposed,however.Wecanquiteeasilyprove:

15Thestrongancestralisso-calledbecauseweneednothave Q∗aa,whereaswealwayshave Q∗= aa

fromwhich

followsbysubstitution.Sowehave

bysimplelogicalmanipulationsandthedefinitionof‘N’.Butthisisweaker thanAxiom7.Thehypothesisofinductionisnotthat, whenever x is F ,its successoris F ;itisonlythat,whenever x isanaturalnumber thatis F ,its successoris F .Thatis,inductionis:

Thisiseasyenoughtoprove—itfollowsbysubstitutionfromFrege’sTheorem 152—butthedifferencebetweenitandwhatthedefinitionoftheancestral deliversimmediatelyturnsouttobehistoricallysignificant,asweshallsee shortly.16

Theonlyremainingaxiom,then,isAxiom3,whichassertsthateverynumberhasasuccessor.LyingbehindFrege’sproofofAxiom3isapictureofhow thenaturalnumbersaregenerated.Thegenerativeprocessbegins,ofcourse, withzero.Fregeinsiststhatzeroexists,evenifnothingelsedoes,becausezero isthenumberofthingsthatarenon-self-identical,andthenon-self-identical thingsexistevenifnothingatalldoes.Butifzeroexists,thenthereisanumberthatisthenumberofthingsthatarelessthanorequaltozero,andthat numberwemaycall‘one’.ByHP,oneisnotzero:Therecanbenoone–one mapbetweenthethingsthatarenon-self-identicalandthethingsthatareless thanorequaltozero,sincethereisatleastonethinglessthanorequaltozero, namely,zero.Butnowbothzeroandoneexist,andsothereisanumberthat isthenumberofthingsthatarelessthanorequaltoone.Callthatnumber ‘two’.ByHP,twocanbeneitherzeronorone.Sozero,one,andtwoexist, andthereisanumberthatisthenumberofthingslessthanorequaltotwo… Formally,whatwewanttoproveis:

whichisthecentralresultofChapterH(Eta)andwhichsays,roughly,that everynaturalnumberissucceededbythenumberofnumbersintheP-series endingwithit(roughly,thenumberofnumberslessthanorequaltoit).The proofproceedsbyinduction,wherewesubstitute:

(ξ, Nx :P∗=xξ) for‘Fξ’.Soweneedtoprovethatzerofallsunderthisconcept:

16Thedifferenceisalsoofsometechnicalimport,sincetheweakerprincipleiseasilyprovable eveninpredicativesystemsthatdonotallowustoproveinduction(Heck,2011).

(154) andthatitishereditaryintheP-seriesbeginningwith0:

(150)

Theproofof(154)iseasy.

Theinterestliesintheproofof(150).Itfollowsbygeneralizationfrom:

Toprovethis,supposethat d isanaturalnumber,that d precedesthenumber ofmembersoftheP-seriesendingwith d,andthat d precedes a.Wemust showthat a precedesthenumberofmembersoftheP-seriesendingwith a. Todoso,wemust,bythedefinitionofP,findsomeconcept F andsome object x fallingunder F suchthat a isthenumberof F sotherthan x andthe numberof F sisthesameasthenumberofmembersoftheP-seriesending in a.Thatis,wemustshowthat:

TheconceptinquestionistobeP∗=ξa;theobjectinquestionistobe a itself. Hence,wemustshowthat:

Thelasttwoconjunctsaretrivial.Thefirstwemayderivefrom:

bythetransitivityofidentity.TheformerfollowsfromthefactthatPissinglevalued,since,byhypothesis,wehavebothPda andP(d, Nx :P∗=xd).The latter,inturn,istheconsequentof:

whichFregederivesfrom:

andtheextensionalityofthecardinalityoperator:

HederivesTheorem149α fromthefollowingtworesults:

whichareTheorems148α and148ζ.Forthelatter,weneedthecentralresult ofChapterZeta:

Thissaysthatthereareno“loops”inthenaturalseriesofnumbers,andittoo isprovedbyinduction.

TheargumentherewilllookfamiliartoanyonewhohasstudiedFrege’s informalproof,in§§82–3of DieGrundlagen,thateverynaturalnumberhas asuccessor.Theproofin Grundgesetze hasmuchthesamestructure,andmany ofthestepsintheproofarealsomentionedin DieGrundlagen.Forexample, (145)isthelastpropositionmentionedbyFregein§83.Butthereareimportantdifferencesbetweenthetwoproofs,aswell.AclosereadingofFrege’s discussionoftheproofin§114of Grundgesetze shows,infact,thatFregewas awarethattheproofsketchedin DieGrundlagen isactuallyincorrect(Boolos andHeck,2011;Heck,2012,§6.7).Theearlierproofpurportstorelyonly uponwhat,asmentionedbefore,triviallyfollowsfromthedefinitionofnaturalnumber:

ratherthanuponmathematicalinductionproper:

Inparticular,theproofin DieGrundlagen wassupposedtogovia:

whichisadirectformalizationofthepropositionmarked‘1.’in§82of Die Grundlagen andwhichisalso(150ε)minusitsfirstconjunct,P

=0d.Frege mentionsthisformulaexplicitlyin§114—itistheformulalabeled(α)—only thentosay,inafootnote,thatit“is,itseems,unprovable…”.Itishardto seewhyFregewouldsomuchashavediscussedthispropositionifitdidnot figurecruciallyinhisearlierargument.Andthat,tome,isthemostimpressive evidencethatFregeknewhisearlierproofwasflawed.

Itfollows,presumably,thatFregecannotactuallyhavehadafullyworked out,formalproofoftheexistenceofsuccessorswhenhewrote DieGrundlagen. Hesimplycouldnothavemadesuchamistakeotherwise.17 Moreinterestingly,itmakesitplainthatFregewasawarethattheremightbepropositions ofhisformallanguagethathewouldregardastruebutthatwerenonetheless unprovablefromtheBasicLawshewasthenpreparedtoaccept.Forthere isgoodreasontothinkFregeregardedthe“unprovable”formula(α)asbeingtrue.ThereisaverysimpleargumentforitthatusesDedekind’sresult thateveryinfinitesetisDedekindinfinite,aclaimthatFregeacceptedastrue, though,likemanymathematiciansofhistime,heregardedDedekind’sproof asinsufficientlyrigorous(Frege,1892,271).I’llreturntothispointbelow.

Finally,carefulanalysisofFrege’sproofin Grundgesetze revealsthatthe onlyfactsaboutnumberstowhichitessentiallyappealsaretheextensionality

17Andthat,inturn,suggeststhatthemanusciptmentionedintheletterfrom1882cannot havecontainedsuchaproof,whichmakesitaninterestingquestionwhatitdidcontain.Frege saysintheletterthathehassetouttoprove“thefirstprinciplesofcomputation”,andtheexistence ofsuccessorsisnotnaturallysodescribed.

Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.