A "More Topological" Proof of The Tietze-Urysohn Theorem - Brian Scott

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

A "More Topological" Proof of the Tietze-Urysohn Theorem

Author(s): Brian M. Scott


Source: The American Mathematical Monthly, Vol. 85, No. 3 (Mar., 1978), pp. 192-193
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/2321064 .
Accessed: 29/12/2014 19:08

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

This content downloaded from 128.235.251.160 on Mon, 29 Dec 2014 19:08:43 PM


All use subject to JSTOR Terms and Conditions
192 B. M. SCOTT [March

A "MORE TOPOLOGICAL" PROOF OF THE TIETZE-URYSOHN THEOREM

BRIAN M. Scorr

In thestandard textbook proofoftheTietze-Urysohn Theorem, thedesiredextension is obtained


as theuniform limitof a sequenceof continuous approximations. (See e.g., [1]-[71.) This proofis
admittedly quiteeasy,butitmustoftenseem a bit"magical"(or at leastunnatural) to the student who
hasjustlearnttheconstruction usedin the proof of Urysohn's lemma. For thatreason, and because it
seemsmorein thespiritofgeneraltopology than theusual proof-which, after all, is just glorified
advancedcalculus!I prefer thefollowing argument, whichuses a moresophisticated versionofthe
Urysohn's lemmaconstruction. to
(I makenoclaim priority ofinvention; although I have neverseenit
in print,theproofmustbe part of thefolklore.)
Theheartoftheproofis thefollowing lemma,whichis ofsomeinterest initsownright. Although
we knowthatitis not in general possiblein a T4-spaceto finddisjoint open sets about two separated
sets,it is possibleto do so iftheseparatedsetsare "sufficiently like"closedsets:specifically, ifthey
are Fi-sets.
LEMMA1. LetX be T4,and letAo and A1 be separated open
F, -setsinX; thentherearedisjoint,
V0, V1 X suchthatAoC VOand AI C Vi.
Proof.For i E {O, 1}, let A, = U {Fi(n): n E N}, where,foreach n E N, Fi(n) is ciosed,and
Fi(n)C Fi(n+ 1). Usingnormality repeatedly,construct, on n,opensetsVi(n)suchthat
byinduction

(a) Fo(O)C Vo(O)C cl Vo(O)C X\clA l ;


(b) F1(O)C V1(O)C cl V1(O)C X\(clAo U cl Vo(O));
(c) foreach n E N, Fo(n+ 1) U cl Vo(n)C Vo(n+ 1) C cl Vo(n+ 1) C X\(clAI U cl Vi(n));
(d) foreach n E N, F1(n+ 1) U cl V1(n)C V1(n+ 1) C cl V1(n+ 1) C X\(clAo U cl Vo(n+ 1)).
Now let V,= U { Vi(n):n E N}, i E {O, 1}; clearlyAi C Vi fori E {O, 1}, and VOn V = 0.
(TheproofofLemma1 providesanotherniceexampleofthetechnique a chimney"
of"climbing
encounteredin theproofthat"regular+Lindel6fimpliesnormal.")
LetX
THEOREM1.(Tietze-Urysohn). bea T4-space, closedsubsetofX.
and letK bea (non-empty)
Iff: K-I( = [O,1]) is continuous, F: X-I suchthatF I K = f.
thereis a continuous
Proof.The idea is to tryto mimictheusualproofoftheUrysohn lemma.Recallthatintheproof
one beginsby constructing open sets W(q), one foreach rational
q thecondition
E I, satisfying
< cl C
(i) p q implies W(p) W(q).
The desiredfunction is thendefinedby
if x W(1)
(1) ~~~F(x)=
( inf{qEQ:xeW(q)}, otherwise,
(where0 isthesetofrationals therequirements
inI,) andis showntosatisfy oftheUrysohn lemma.
Those requirements are not verystringent:F mustbe contin'uous and mustassumespecified
values(O and 1) in justtwo"places".The Tietzetheorem is moredemanding, sinceit "tiesdown"
manyvaluesof F, so we imposeadditionalconditions on thesets W(q):

(ii) K n W(q)= V(q), and


(iii)K n cl W(q)= C(q),
by(1), does
whereV(q) = f1[[O,q)], and C(q)= f1[[O,q]]. Condition(ii) ensuresthatF, defined
extendf. Condition(iii) is technical:the sets W(q) will be constructed usingan
recursively,
enumerationof 0, and (iii) is neededto maketherecursion work.
We beginbydefining W(O)= 0 and W(1) = X\-1 [{1}].At thegeneralstepoftheconstruction,

This content downloaded from 128.235.251.160 on Mon, 29 Dec 2014 19:08:43 PM


All use subject to JSTOR Terms and Conditions
19781 MATHEMATICAL EDUCATION 193

weshallhaverationals p,q, r E I withW(p) and W(q) alreadydefined so as tosatisfy


(i)-(iii)andwith
p < q < r,and we shallconstruct W(q) as follows.
V(q) andK\C(q) areseparatedF,,-sets in X, so, byLemma1, thereis an openH C X suchthat
V(q) C H and K n clH C C(q). Let G = H n [(X\K) U V(q)], and note thatG is open and
satisfies
(ii) and(iii).Next,usingnormality twice,findan openHoC X suchthatcl W(p) U cl V(q) 5
Ho 5 clHo 5 W(r) and an openH1 C X suchthatcl W(p) 5 H1 5 clH1 5 Ho\(K\V(q)). Put W(q) =
H1 U (G n Ho); it is theneasilycheckedthatW(q) satisfies (i)-(iii).
Actually,the fullstatement of the Tietze theoremis slightly stronger thanTheorem1: any
continuous f: K -E1 can be extendedto a continuous F: X-) E1 in sucha waythatanyopenray
containing therangeoff alsocontains therangeofF. Thisstrengthening followsfromTheorem1 as in
theusualproof(cf.[1]).
References

1. J.Dugundji,Topology,Allynand Bacon,Boston,1966.
2. R. Engelking,Outlineof GeneralTopology,Wiley,New York,1968.
3. J.Greever,Theoryand Examplesof Point-Set Topology,Brooks/Cole,Belmont,Calif.,1967.
4. J.L. Kelley,GeneralTopology,Van Nostrand, Princeton,
NJ,1955.
5. J.R. Munkres, Topology,a FirstCourse,Prentice-Hall,
EnglewoodCliffs,
NJ,1975.
6. G. F. Simmons, Introduction
to Topologyand ModernAnalysis, New York,1963.
McGraw-Hill,
7. S. Willard,GeneralTopology,Addison-Wesley,Reading,Mass.,1970.

DEPARTMENT OF MATHEMATICS, CLEVELAND STATE UNIVERSITY, CLEVELAND, OH 44115.

MATHEMATICALEDUCATION

EDITED BY PAUL T. MIELKE AND W. E. MASTROCOLA

Material for thisDepartmentshould be sent to W. E. Mastrocola,Departmentof Mathematics,Colgate


University,Hamilton,NY 13346.

A MODELING PROBLEM FOR THE CLASSROOM

J. GLENN BROOKSHEAR

Theincrease incourse intheareaofmathematical


offerings modeling hasresultedina universal
inproblems
interest forstudents tosolveontheir
own.Theseproblems must beinvolvedenoughtobe
challengingyetsimpleenoughto be statedin a fewpages.Moreover, thediversedisciplines
representedbytheenrollment inmostmodeling classesdictate
thattheseproblems be basedon a
background common to a widevariety Suchproblems
ofstudents. beingusedto
aresuccessfully
developa student's
ability ofa problem,
toisolatethekeypoints tomakemeaningful assumptionsin
orderto obtaina manageable model,andto evaluatetheeffects theseassumptionshaveon the
It is myhopethattheMONTHLY
model'spredictions. can serveas a forumfortheexchangeofsuch
problems.Thefollowing "prerequisite
problem"hassucceededinstimulating inmystudents,
interest
and I hope it can do thesameforothers.
In 1969an increaseintuition
at BunkerUniversity
anda strongrecruitingcampaignbyother
nearbyuniversities
resultedin an unusually
smallfreshman class at BunkerUniversity.
The
enrollmentproblem wassolvedthenextyear.In fact,admissionapplicationshavesoaredto the
extentthattheuniversityhasbeenabletoadoptmorerigorous entrance requirements
andstillfill

This content downloaded from 128.235.251.160 on Mon, 29 Dec 2014 19:08:43 PM


All use subject to JSTOR Terms and Conditions

You might also like