You are on page 1of 5

Measure and Cardinality Author(s): James M.

Briggs and Thomas Schaffter Reviewed work(s): Source: The American Mathematical Monthly, Vol. 86, No. 10 (Dec., 1979), pp. 852-855 Published by: Mathematical Association of America Stable URL: http://www.jstor.org/stable/2320153 . Accessed: 18/11/2011 04:55
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Mathematical Association of America is collaborating with JSTOR to digitize, preserve and extend access to The American Mathematical Monthly.

http://www.jstor.org

852

JAMES M. BRIGGS AND THOMAS SCHAFFTER

[December

T. The weights Hadamard of and J. K. W. Schmidt Edward H. Wang, matrices, Combin. Theory A, 23 Ser. (1977)257-263. Math.Slov., (1976)329-335;MR 55 #7853;Zbl. 348.05114. 26 Jiri On graphs, Sedlacek, magic of and ZielonaGora, 1976, ,Someproperties magicgraphs, GraphsHypergraphs Block Systems, 247-253. A Y. Shiloach, minimum linear for arrangement algorithm undirected (tbp). trees A David Singmaster D. J.Souppouris, constrained and isoperimetric problem, Math.Proc.Cambridge Philos. Soc.,83 (1978)73-82. NeilJ.A. Sloane, Handbook Integer A of Sequences, Academic Press, 1973. W. F. Smyth, labeling A for sum algorithm minimum weight ofcomplete edge trees binary (tbp). in Herbert Taylor, pathsums an edge-labelled Math.Mag.,50 (1977)258-259. Odd tree, of in ZalmanUsiskin, thenumber 3-term On arithmetic progressions setswith 4-term no ones,Math.Mag. (tbp).

CLASSROOM NOTES
EDITEDBY DEBORAH TEPPER HAIMO ANDFRANKILN TEPPER HAIMo Materialforthisdepartment shouldbe sentto Professor Deborah Tepper Haimo, Department Mathematical of Sciences,University Missouri, Louis, MO 63121. of St.

The editors Advice prospective to authors: have recently beenreceiving abouttentimes many as Classroom our will Notesas can be used. It willsimplify workifauthors submit onlyespecially interesting manuscripts. R.P.B.
MEASUR1E AND CARDINAL1TY JAMES BRIGGS M. ANDTHOMAS SCHAFFrER

The purpose thispaperis to describe of how thesettheory axiomswe acceptaffect Lebesgue on knownamongset theorists easy to state measure thereal line.The results generally are and A but rather hardto getat forthenon-specialist. coursein real analysis would seem to be an students theseresults. to appropriate place to introduce One of thefirst learnedaboutLebesguemeasure thereal lineis thatcountable on properties sets of real numbersare measurableand theyhave measurezero. This is, of course,an of immediate consequence the factsthatsingleton have measurezero and thatLebesgue sets is measure countably To who knowsthere additive. thestudent of maybe setsof real numbers between (thecardinality thenatural of cardinality strictly and c (thecardinality of numbers) so thereals),thequestionnaturally arisesas to whatcan be said about their This measurability. from twoperspectives: whatmust can be answered hold in thereals,and (2) what question (1) hold in the reals.More precisely, distinction between whatis impliedby our the is might (1) basic axioms, and (2) whatis consistent withand whatis i.e.,whatcan be provedfrom them, of independent theaxioms.Of course, is notsufficient consider theusualaxiomsforthe it to just reals(i.e.,theaxiomsfora complete, ordered field);we mustalso consider axiomsof settheory. will the Clearly, answers givento theabove questions dependon whatsettheory axiomswe use. Thereis consensus amongmathematicians theaxiomsof Zermelo-Fraenkel theory that set (ZF) are acceptablebut thereare controversies the concerning Axiom of Choice (AC) and the Continuum Hypothesis (CH).

1979]

CLASSROOM NOTES

853

Fromthework Godel [31and Cohen[2] we knowthat of bothof thestatements, and CH, AC are independent ZF, i.e.,neither be proved refuted ZF. Another of can or in wayto putitis that there modelsof ZF in whichAC (respectively, holds and modelsin whichAC (CH) are CH) fails.For thosewho believein an objective of there strong is reality theobjectsof mathematics intuitive evidence AC. (Given a set of nonempty "surely" for there a set containing is sets, one element fromeach. Why wouldn'ttherebe? For a discussion thisand contrary of viewsof mathematics Bernays For a similar discussion related CH see Godel [4]. Thereis no to see [1]. such evidencefor CH (whyshould c be the first uncountable most cardinal?). Accordingly, mathematicians use unhesitatingly AC butdo nottakea standon CH. Thus,our original query (whatcan be said in ZF + AC (ZFC) about themeasurability sets of reals withcardinality of between and c) shouldbe meaningful mostmathematicians. to 40 It is wellknownamongpersonsspecializing thefoundations mathematics if there in of that are anysetsofrealnumbers cardinality of between and c, all suchsetswhichare measurable mo havemeasure zero.We present elementary an proof thisfact.Nextwe discusswhatis known of about thepossibility non-measurable in this"in between" of sets land. Finally, thosewho for consider up fordebate,we comment whatcan be said in theabsenceof thisaxiom. AC on
THEOREM.

If E is a measurable of realsof cardinality thanc, then has measure set less E 0.

Proof.We showthatif E has positive measurethenthecardinality E is c. Accordingly, of supposemE >0 and let F be a closed subsetof E of positivemeasure.Then thereare two disjoint, closed intervals F I,JI of lengthless than mF which intersect in sets of positive measure.Now within thereexist two disjoint, closed intervals Io IoJIo1 of lengthless than m(F n IO) which also intersect in setsof positive F measure. we Similarly, can findintervals Ilo and III within withthecorresponding properties. I, in Continuing thismanner, produceat the nthstep2" disjoint, we closed intervals Ij.,. ,,, F ji E f0, each of whichintersects in a set of positivemeasure.In the limiteach infinite 1), sequence to sequenceIj,,Ij,j2,... of nested, closed jlIJ2.... from{0,1} corresponds an infinite intervals, of whichintersect By thecompleteness the realsthe intersection such a all F. of of sequenceis non-empty, hencedetermines leastone pointofF. The pointsof F produced and at by these intersections clearly are and distinct, therefore process this 2'o identifies = c elements of F, giving thedesired us result. At this point we are assumingAC. However,we note that it is not needed for the constructions thisproof:theinterval in selection could be made without use; itsuse clearly its can also be avoided when choosinga point fromeach intersection nestedintervals of by requiring lengths theintervals go to zero,thereby the of to producing singletons. Thus,in all modelsof thereals(within ZFC) whereCH fails(i.e.,wheresetsof realsof "in between" cardinality all exist), measurable of "in between" sets cardinality have measure zero. But can therebe non-measurable of "in between"cardinality? classicalproofof the sets The existence a non-measurable due to Vitali(1905),is of no help sinceit producesa set of of set, cardinality (see [5,pp. 69-701).However, c RobertSolovayhas provided answer. theone the On sidehe has shown thatthere modelsofZFC in which failswhere are CH every of cardinality set less thanc is measurable. Hence,assuming setsof "in between" that cardinality exist, cannot we prove,usingZFC, thatany such setsare non-measurable. theterminology Foundations, In of ZFC + - CH cannotprovethere non-measurable of "in between" are sets cardinality. On theother side,Solovayhas shownthatthere modelsof ZFC in whichCH failswhere are there setsofcardinality (thefirst are 81 uncountable cardinal) which notmeasurable. are Hence, ZFC cannotprovethere no non-measurable of thekindin question are sets (see [9]). Thus,just as we cannotproveor refute ZFC) thatthere setsof realsof "in between" (in are we cardinality, cannot prove or refutethat among such sets (if thereare any) some are non-measurable. Stateddifferently, as ZFC does not decide the Continuum just Hypothesis, ZFC does notdecidewhether setsof realsof cardinality thanc are measurable. all less

854

JAMES M. BRIGGS AND THOMAS SCHAFFTER

[December

of on Finally,we comment whatcan happenin the absenceof AC. Some form theaxiom In measure. theabsenceof AC it can happen a to in be must retained order preserve reasonable us sets(Levy [71),forcing intoa measure unionof countable thatthesetof realsis a countable to A additive. weak formof AC, whichis sufficient or trivial not countably whichis either Choices(DC). of is of properties Lebesguemeasure theprinciple Dependent the preserve main is R relation on a set S such thatforall x E S there a y E S for givena binary DC statesthat, of number choicescan be made, witha particular e S, a countable whichxRy,thenstarting xo givenxoe S thereexistsa on to relative R, each one depending the previous;moreexactly, impliesthe Countable countablesequencex I,x2,... such that xoRxRX2..-. This statement and is sets of collection nonempty has a choice function) Axiomof Choice (everycountable implication.) by implied AC. (See Jech[6, p. 791fora proofof thefirst clearly Solovay has shownin [81 that if thereis a model of ZF whichcontainsan inaccessible is withZF), thenthere a cardinalis consistent of cardinal(i.e.,if theexistence an inaccessible cardinal an (Roughly, inaccessible set modelof ZF +DC in whichevery of realsis measurable. unionsor powersets; belowby taking whichcannotbe reachedfrom cardinal is an uncountable less thanX sets,each of cardinality thans, has thentheunionof fewer i.e.,if K is inaccessible, less less and thepowersetof a set of cardinality thanK has cardinality less cardinality thanK, of thatthe existence such a cardinalis believeit is quite unlikely than". Many set theorists to Thus,if one is willing give is the with inconsistent ZF; however, question open at thistime.) to are up AC, which most mathematicians not, it is consistent assume that thereare no of of the sets non-measurable of realsat all (assuming consistency theexistence an inaccessible of a of the here Solovay'smotivation was notto present possibility considering theory cardinal). the of to but realsin whichall setsweremeasurable, rather showthenecessity AC in proving set: existence a non-measurable "Of course,the axiom of choice is true,and so thereare of of about thetruth AC, the if sets non-measurable [8,p. 3]." However, one is notquiteso certain is above interpretationpossible. are that cardinalassumption, there undertheinaccessible In thesamepaper Solovayshows, sets whereall definable of modelsof ZFC, somein whichCH holdsand somein whichit fails, within of The realsare measurable. adequacyof an interpretation a conceptsuchas definability as Solovay's translation is of the formalism set theory always open to question; however, This the close to capturing notion. seemsto comevery a from sequenceof ordinals" "definable set. a explicitly non-measurable thatwe willneverbe able to exhibit evidence is very strong it we In summary, have seenthatiftheAxiomofChoiceis accepted can be provedthatthere sets sets c; are non-measurable of realsof cardinality and if thereare uncountable of reals of zero. Further, as it just theymusthave measure less cardinality thanc whichare measurable, the including Axiomof axiomsof set theory, the (from ordinary cannotbe provedor refuted that it sets Choice) thatsuchuncountable ofrealsexist, also cannotbe provedor refuted among ones. are are thatthere some) there non-measurable thesesets(assuming to We have also seen thatif one is willing giveup theaxiomof choice,it cannotbe proved of the sets are anynon-measurable of realsat all (assuming consistency theexistence thatthere are there modelsof ZFC, some in whichCH holds cardinal).Furthermore, of an inaccessible Thus it is very sets and some in whichit fails,in whichall definable of reals are measurable. set. a we unlikely willeverbe able to produceexplicitly non-measurable
for We Acknowledgment: thankLooy Simonoff his contributions.

References and ed. of in in On 1. P. Bemays, platonism mathematics, Philosophy Mathematics, by P. Benacerraf H. N.J. Englewood Cliffs, 1964. Prentice-Hall, Putnam, Part hypothesis, I, II, Proc.Nat. Acad. Sci. U.S.A.,50 of The independence thecontinuum 2. P. J.Cohen, 51 (1964)105-110. (1963)1143-1148;

1979]

CLASSROOM NOTES

855

3. K. Godel,The consistency theaxiomof choiceand thegeneralized of continuum hypothesis the with axioms settheory, of Annals Mathematics, No. 3, Princeton of Study Univ.Press, Princeton, 1940. N.J., in 4. __ is of ed. , What Cantor's continuum problem?, Philosophy Mathematics, byP. Benacerraf H. and Putnam, Prentice-hall, Englewood Cliffs, N.J.,1964. Van 5. P. R. Halmos, Measure Theory, Nostrand, Princeton, 1965. N.J., in Lecture Notesin Mathematics, 1971. 6. T. J.Jech, Lectures SetTheory, Springer-Verlag, of of 7. A. Levy,The independence certain consequences theaxiomof choice,Fund. Math.,54 (1964) 135-137. A in set Ann. 8. R. Solovay, modelof settheory which every of realsis Lebesgue measurable, ofMath., 92, No. 1 (July 1970)1-56. 2 Ann. Internal Cohenextensions, Math.Logic, (1970)143-178. 9. R. Solovay D. Martin, and
DEPARTMENTOF MATHEMATICS,UNIVERSITY OF NEVADA, LAS VEGAS, NV 89154.

THE MATRIC EQUATION X2 =A W. R. UTZ

X2 for the equation =A. procedure solving matric of The purpose thisnoteis to givea useful in a of In response the curiosity sophomores linearalgebracoursesI have accumulated to suitableforan equationsin matrices involving polynomial and theorems number exercises of a of course.Chapter7 of a book [5] of C. C. MacDuffeeprovides summary basic elementary equationsup to 1933.It is adequate for for results and is a good sourceof references matric field(cf.[2],[4],for is course.Matricequations stillan activeresearch enriching elementary an of to 2], According [6, Chapter thesolution the survey. example)butI am unawareof a recent IX-X of 27..0AiV. divisors to matric equation 2!OAIX' =0 is equivalent thesearchforright In thisconnection paper[3] is of interest. the X2 a [1, to quadratic = A, a classicaltheorem p. 299] ensures (nonsinguReturning thesimple I to once suggested me a lar) solution theequationifA is nonsingular.now describe method, of X2 of of regardless the the by W. E. Roth,thatreduces problem solving =2A to linearequations rankofA. are for there no solution easilydevised), is If X is a solution theequation(examples which of then

hence

X2-X2I=A -21; (X-XI)(X+XI)=A -2I. (*)

then function a solution, of If +(X) denotes characteristic the 4(X)4( -X) = det(A-A2I)

if of of mustbe a divisor det(A-2I). function a solution, one exists, and so thecharacteristic the all For each ofthepossiblesolutions +(X) of(*) one secures equation+(X) = 0 in which even A powersof X may be replacedby powersof theknownmatrix to reducethe equationto a equation. linearmatrix the As an example, consider equation 9 0 -81 5 4 -5 X2 For this A,

_ 0(X)0(-X) = det(A -X2I) =- X2(X2 9)(X2 - 4)

You might also like