PREFACEThe purpose of this monograph is to give an axiomaticfoundationforthetheoryofprobability.Theauthorsethimself
thetask of puttingin their natural place, among the general
notionsofmodernmathematics,thebasicconceptsofprobability
theory—conceptswhichuntilrecentlywereconsideredtobequite
peculiar.
Thistaskwouldhavebeenaratherhopelessonebefore
theintroductionofLebesgue'stheories ofmeasureandintegration.
However,after Lebesgue'spublicationofhisinvestigations,the
analogiesbetweenmeasureofasetandprobabilityofanevent,
andbetweenintegralofafunctionandmathematicalexpectation
ofarandomvariable,becameapparent.Theseanalogiesallowed
offurtherextensions; thus,forexample,variousproperties of
independent
randomvariableswere
seentobeincompleteanalogywiththecorrespondingproperties oforthogonal functions.But
ifprobabilitytheorywastobebasedontheaboveanalogies, itstillwasnecessarytomakethetheoriesofmeasureandintegra-
tion independentof thegeometric elementswhichwerein the
foregroundwithLebesgue. ThishasbeendonebyFrechet.
Whileaconceptionofprobabilitytheorybasedontheabovegeneralviewpointshasbeencurrentforsometimeamongcertain
mathematicians,there waslackingacompleteexpositionofthe
wholesystem, freeofextraneous complications. (Cf.,however,thebookbyFrechet,
[2]inthebibliography.)Iwishtocallattentiontothosepointsofthepresentexpositionwhichareoutsidetheabove-mentionedrangeofideasfamiliartothespecialist.They
arethefollowing: Probabilitydistributionsininfinite-dimensional
spaces (ChapterIII,§
4);differentiationandintegration ofmathematicalexpectationswithrespecttoaparameter(ChapterIV,
§5);andespeciallythetheoryofcondi-tional probabilities and conditional expectations (Chapter V).Itshouldbeemphasizedthatthesenewproblemsarose,ofneces-sity,fromsomeperfectlyconcretephysicalproblems.11Cf.,e.g.,thepaperbyM.Leontovichquotedinfootnote 6 on
p.46;alsothejointpaperbytheauthorandM.Leontovich,ZurStatistikderkontinuier-lichen
Systemeunddeszeitlichen VerlaufesderphysikalischenVorgdnge.Phys.Jour,oftheUSSR,Vol.3,1933,
pp.35-63.