You are on page 1of 2

CenterforPhilosophyofScienceoftheUniversityofLisbon

InternationalColloquiumThePhilosophersandMathematics

Abstract
TheIntensionalTakeontheAxiomofChoice:
ADialogicalPerspectiveonItsProof

ShahidRahman1
UniversitdeLille,UMR8163:STL

The present talk studies the interplay between a Philosophers reflections on mathematics, namely,
JaakkoHintikkaandaMathematicianreflectingonthephilosophicalandepistemologicalfoundationsof
his science. Despite our criticism of the former we do think that his work opened a new path for the
interchangebetweenphilosophyandmathematicsinawaythatwasalreadyprefiguredbyHenriPoincar
thoughwewillnotdiscussherePoincarsviewontheissue.Thetalk,Ithink,nicelyfitswiththebeautiful
titleofthismeetingthathonoursRoshdiRashedwhoseworkisalandmarkinthefield.

Ithasbeensaid,andrightlyso,thattheprincipleofsettheoryknownastheAxiomofChoice(AC)is
probablythemostinterestingandinspiteofitslateappearance,themostdiscussedaxiomofmathematics,
second only to Euclids Axiom of Parallels which was introduced more than two thousand years ago
(Fraenkel/BarHillelandLevy[1973]).
AccordingtoErnstZermelosformulationof1904ACamountstotheclaim;that,givenanyfamilyAof
nonemptysets,itispossibletoselectasingleelementfromeachmemberofA.Theselectionprocessis
carriedoutbyafunctionfwithdomaininM,suchthatforanynonemptysetMinA,thenf(M)isanelement
ofM.Theaxiomhasbeenresistedfromitsverybeginningsandtriggeredheatedfoundationaldiscussions
concerningamongothers,mathematicalexistenceandthenotionofmathematicalobjectingeneraland
offunctioninparticular.However,withthetime,thefoundationalandphilosophicalreticencefadedaway
and was replaced by a kind of praxisdriven view by the means of which AC is accepted as a kind of
postulate (rather than as an axiom the truth of which is manifest) necessary for the practice and
developmentofmathematics.
RecentlythefoundationaldiscussionsaroundACexperiencedanunexpectedrevivalwhenPerMartinLf,
showed(around1980)thatinconstructivelogic(thatdoesnotpresupposetheexcludedmiddle)theaxiom
ofchoiceislogicallyvalid(howeverinitsintensionalversion)andthatthislogicaltruthnaturally(almost
trivially) follows from the constructive meaning of the quantifiers involved it is this evidence, that
makesitanaxiomratherthanapostulate.Extensionalitycanalsobeprovedbutthen,eitherthirdexcluded
or unicity of the function must be assumed. MartinLfs proof, for which he was awarded with the
prestigiousKolmogorovprice,showedthatattherootoftheolddiscussionsanoldconceptualproblem
wasatstake,namelythetensionbetweenintensionandextension.

ThepresenttalkisbasedonClerbout/RahmanLinkingGamesandConstructiveTypeTheory:Dialogical
Strategies,CTTDemonstrationsandtheAxiomofchoice.(inprint).


AnevenmorerecentdevelopmentstudiesthegametheoreticalinterpretationofACbroughtforwardby
JaakkoHintikkaby19962,thoughhedidnotconsiderMartinLfsproofpresumablysobecauseHintikka
is not favorable to constructivist approaches. The aim of the paper is to develop a new constructivist
approachtothegametheoreticalinterpretationofACbasedontheCTTproof.Moreprecisely,Clerbout
andRahmanshowedthattheCTTunderstandingofAC,thatstressesthetypedependenceinvolvedby
thefunctionthatconstitutestheproofobjectoftheantecedent,canbeseenastheresultofanoutside
insideapproachtomeaning3.Itisthisapproachtomeaning,soweclaim,thatprovidesanaturaldialogical
interpretationtoAC,wherethe(intensional)functioninvolvedunderstoodasrulesofcorrespondence
producedbytheplayersinteractionconstitutesaplayobjectforthe(firstorder)universalquantifierthat
occurs in the antecedent of the formal expression of this axiom. Different to Hintikkas own game
theoreticalapproachthedialogicaltakeonACdoesnotrequireanotaxiomatisablelanguagesuchasthe
oneunderlyingIndependentFriendlyLogic(IFlogic).AspointedoutbyJovanovic(2014)thedialogical
approachtoCTTsupportsHintikkaclaimsthatagametheoreticaljustifiesZermelo'saxiomofchoiceina
firstorder way perfectly acceptable for the constructivists, however, no underlying IFsemantics is
required.Moreover,HintikkasownformulationofAC,whenspelledout,yieldstheCTTformulationof
MartinLf,thatisconstructivistafterall.Summingup,thoughHintikkaisrightinstressingtheperspicuity
ofthegametheoreticalinterpretationofACheiswronginrelationtothetheoryofmeaningrequiredfor
thisinterpretation.OneofthemainreasonsbehindHintikkascriticismoftheconstructivistapproachis
thatheassumesthattherejectionoftheclassicalunderstandingoftheACbytheconstructivistshasits
rootsintherejectionofafunctionthatisnotarecursiveone.However,asthoroughlydiscussedbyThierry
Coquand (2014), already Arend Heyting (1960) pointed out that recursive functions cannot (without
circularity)beusedtodefineconstructivityandfinallyErretBishop(1967)showedthatrecursivefunctions
are not at all needed to develop constructive mathematics. The very point of the rejection by the
constructivistsoftheclassicaltakeonACistheir(theclassical)assumptionthatthefunctionatstakeisan
extensionalone.Forshort,theCTTproofofACisbasedontheintensionaltakeonfunctionsandthisis
whatthedialogicalinterpretationdisplays.Wewillconcludewithsomereflectionsontheconceptuallink
betweentheconstructivistnotionoffunctionasruleofcorrespondenceanddialogicalinteractionwithin
humanplayablegames,thatmightrestatesomeofHintikkasremarksalbeitinadifferentframe.
I will finish by noticing that, as mentioned above, it is Henry Poincar (1905) who already compared
mathematicalknowledgewiththehabilitytounderstandthedevelopmentofplayofchess.According,to
thisview,andifwejoinandPoincarswithWittgensteinHintikkasremarksinthecontextofthenotion
ofproofaswinningstrategy,theknowledgeofaproofamountstograspingitsmeaningandthisamounts
onbeingabletoshowhowtoconstructthisproof:aproofbeyondourabilitiestoconstructitisnotproof
atall:humanplayableorreachableproofprovidesthefoundationtothenotionofinferenceandtruth.

SeeforexampleHintikka(1996,2001).
ForathoroughdiscussionontheissueseeJovanovich(2014,2015).

You might also like