Professional Documents
Culture Documents
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
IA'1,where
f-i ~ ~ ~ )n,P +n
k-i
p = E(it- 1)P1? + (ik- 1)-P +i+-= (it - 1)Pt+1
t=l nk nk t=j+1
8-1
+(ii - 1)Pk+ + E (it - 1)1't+l + is.
t=k+l
upon a rule fororderingformulasof the same grade (as is easily done), we can
enumeratethe set of all formulasby arrangingthemin the orderof theirgrades.
For convenience,let us use the phrasefunctionofpositiveintegersto mean a
single-valuedfunctionof one variable,f(x), which takes on a value which is a
positive integer,wheneverx takes on a value which is a positive integer.We
can prove, by a familiarargument,that the set of all functionsof positive in-
tegers is not enumerable. For, iffi(x), f2(x), f3(x), * * * is any enumerationof
functionsof positiveintegers,then 1+f:(x) is a functionof positiveintegersnot
included in the enumeration.
Since, in any systemof symboliclogic, the set of all formulasis enumerable,
whereas the set of all functionsof positive integersis not enumerable,it seems
to followthat,in the case of any systemofsymboliclogic,thereexistsa function
of positive integerssuch that thereis no formulawhichstands forit. And surely
the existence of a functionof positive integerswhich has no representationas
a formulain the systemmeans that the systemis inadequate even forelemen-
tary numbertheory.
The Richard paradox can be said to consist in the followingproblem.How
is it possible that a systemof symboliclogic, in which the set of all formulasis
enumerable,should be adequate for any branch of mathematics which deals
withthe membersofa non-enumerableset (in particularforelementarynumber
theory)?
Given a system of symbolic logic, let us try to constructthe functionof
positive integerssuch that thereis no formulain the systemthat stands forit.
What we must do is firstto enumerateall formulas,by the method which we
have described,and then,going throughthis enumeration,to pick out in order
those formulaswhich stand forfunctionsof positive integers.The resultis an
enumerationof all formulaswhich stand forfunctionsof positiveintegers.And
ifwe letfn(x)be the functionof positiveintegersrepresentedby the nthformula
in this enumeration,then 1+fx(x) is the functionof positive integerssuch that
there is no formulain the systemthat stands forit.
But this function1+f$(x) is not, in general,definedin such a way that it is
always possible to calculate its value fora given positive integerx. For, in the
processofgoingthroughthe listofall formulasand pickingout thosewhichstand
for functionsof positive integers,we may at some stage finda formulaabout
which we do not know whetheror not it stands fora functionof positive in-
tegers. For example, we may finda formulawhose intuitivemeaning is, "The
least positive integern, greaterthan x, such that the equation Un+Vn= Wn has
a solution in positive integralvalues of u, v, w." And we could not determine
whether this formula stood for a functionof positive integerswithout first
proving,or disproving,Fermat's last theorem.Indeed, to be sure of always be-
ing able to determinewhethera given formulastands fora functionof positive
integers,we must have discovered a method of procedurewhich would enable
us to solve any problemof numbertheorywhatever. Thereforethe infinitese-
quence (about whichwe have been talking)ofall formulaswhichstand forfunc-