sf89e,7/8/1996PROBABILITYINQUANTUMTHEORYyE.T.JaynesWaymanCrowProfessorofPhysicsWashingtonUniversity,St.LouisMO63130Abstract:Forsomesixtyyearsithasappearedtomanyphysiciststhatprobabilityplaysafundamentallydi erentroleinquantumtheorythanitdoesinstatisticalmechanicsandanalysisofmeasurementerrors.Itisacommonlyheardstatementthatprobabilitiescalculatedwithinapurestatehaveadi erentcharacterthantheprobabilitieswithwhichdi erentpurestatesappearinamixture,ordensitymatrix.AsPauliputit,theformerrepresents\EineprinzipielleUnbestimmtheit,nichtnurUnbekanntheit.Butthisviewpointleadstosomanyparadoxesandmysteriesthatweexploretheconsequencesoftheuni edview,thatallprobabilitysigni esonlyincompletehumaninformation.Weexamineindetailonlyoneoftheissuesthisraises:therealityofzero{pointenergy.CONTENTSINTRODUCTION:HOWWELOOKATTHINGS1HOWDOWELOOKATGRAVITATIONANDQED?2HOWDOWELOOKATBASICQUANTUMTHEORY?4PROBABILITYTHEORYASTHELOGICOFSCIENCE7HOWWOULDQUANTUMTHEORYBEDIFFERENT?9ISZERO{POINTENERGYREAL?11THELAMBSHIFTINCLASSICALMECHANICS13CLASSICALSUBTRACTIONPHYSICS15CONCLUSION18REFERENCES19INTRODUCTION:HOWWELOOKATTHINGSInthisworkshopweareventuringintoasmokyareaofsciencewherenobodyknowswhattherealtruthis.Such eldsarealwaysdominatedbythecompensationphenomenon:supremeself{con dencetakestheplaceofrationalarguments.Thereforeweshalltrytoavoiddogmaticasser-tions,andonlypointoutsomeofthewaysinwhichquantumtheorywouldappeardi erentifweweretoadoptadi erentviewpointaboutthemeaningandfunctionaluseofprobabilitytheory.WethinkthattheoriginalviewpointofJamesBernoulliandLaplaceo erssomeadvantagestodayinbothconceptualclarityandtechnicalresultsforcurrentlymysteriousproblems.yArevisedandextendedversionofapaperpresentedattheWorkshoponComplexity,Entropy,andthePhysicsofInformation,SantaFe,NewMexico,May29{June2,1989.TheoriginalversionisintheProceedingsVolume,Complexity,EntropyandthePhysicsofInformation,W.H.Zurek,Editor,Addison{WesleyPublishingCo.,Reading,MA(1990).2Howwelookatatheorya ectsourjudgmentastowhetheritismysteriousorirrationalontheonehand;orwhetheritissatisfactoryandreasonableontheother.Thusita ectsthedirectionofourresearche orts;andafortioritheirresults.Indeed,whetherwetheoristscaneveragainmanagetogetaheadofexperimentwilldependonhowwechoosetolookatthings,becausethatdeterminesthepossibleformsofthefuturetheoriesthatwillgrowoutofourpresentones.Oneviewpointmaysuggestnaturalextensionsofatheory,whichcannotevenbestatedintermsofanother.Whatseemsaparadoxononeviewpointmaybecomeaplatitudeonanother.Forexample,100yearsagoitwasamuchdiscussedproblemhowmaterialobjectscanmovethroughtheaetherwithoutresistance.Yetadi erentwayoflookingatitwouldhavemadethemysterydisappearwithoutanyneedtodispensewiththeaether.Onecanregardmaterialobjects,notasimpedimentstothe\ owofaether,butaspartsoftheaether(\knotsinitsstructure)whicharepropagatingthroughit.Onthiswayoflookingatit,thereisnomysterytobeexplained.AsastudentatPrincetonmanyyearsago,IwasfascinatedtolearnfromJohnWheelerhowmuchofphysicscanberegardedasreallyonlygeometry,inthisway.Todaywearebeginningtorealizehowmuchofallphysicalscienceisreallyonlyinformation,organizedinaparticularway.Butwearefarfromunravellingtheknottyquestion:\Towhatextentdoesthisinformationresideinus,andtowhatextentisitapropertyofNature?Surely,almosteveryconceivableopiniononthiswillbeexpressedatthisworkshop.Isthisvariabilityofviewpointsomethingtobedeplored?Well,eventuallyweshouldhopetopresentauni edpicturetotherestoftheworld.Butforthemomentthisisneitherpossiblenordesirable.Wearealllookingatthesamerealityandtryingtounderstandwhatitis.Butwecouldneverunderstandthestructureofamountainifwelookedatitonlyfromoneside.Therealitywearestudyingisfarmoresubtlethanamountain,andsoitisnotonlydesirable,butnecessarythatitbeexaminedfrommanydi erentviewpoints,ifweareevertoresolvethemysteryofwhatitis.Herewepresentoneofthoseviewpoints,whichwethinkhasnotbeensu cientlyrecognizedintherecentliterature.Firstwenoteamoreimmediateexampleofthee ectofhowwelookatthings,tosupportbyphysicalargumentswhatissuggestedlaterfromprobabilityconsiderations.HOWDOWELOOKATGRAVITATIONANDQED?Inteachingrelativitytheory,onemayencounterabrightstudentwhoraisesthisobjection:\Whyshouldsuchafundamentalthingasthemetricofspaceandtimebedeterminedonlybygravitational elds|theweakestofallinteractions?Thisseemsirrational.Wepointouttohimadi erentwayoflookingatit,whichmakestheirrationalitydisappear:\Oneshouldnotthinkofthegravitational eldasakindofpreexistingforcewhich‘causes’themetric;rather,thegravitational eldisthemainobservableconsequenceofthemetric.Thestronginteractionshavenotbeenignored,becausethe eldequationsshowthatthemetricisdeterminedbyalltheenergypresent.Accordingtothe rstviewpoint,onemightthinkitapressingresearchproblemtoclearupthemysteryofwhythemetricdependsonlyongravitationalforces.Onthesecondviewpointthisproblemdoesnotexist.Butthenifthestudentisverybright,hewillbebackthenextdaywithanothercriticism:\Ifthegravitational eldisonlyakindofbootstrape ectoftheotherforces,itraisesthequestionwhetherthegravitational eldshouldbequantizeds