Continuum Hypothesis

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

VOL. 50, 1963 MATHEMATICS: P. J.

COHEN 1143

19 Gardner, R. S., A. J. Wahba, C. Basilio, R. S. Miller, P. Lengyel, and J. F. Speyer, these


PROCEEDINGS, 48, 2087 (1962).
20 Ralph, R. K., and H. G. Khorana, J. Am. Chem. Soc., 83, 2926 (1961).
21 Ames, B. N., and D. T. Dubin, J. Biol. Chem., 235, 769 (1960).
22 Chen, P. S., Jr., T. Y. Toribara, and H. Warner, Anal. Chem., 28, 1756 (1956).
23 Heppel, L. A., D. R. Harkness, and R. J. Hilmoe, J. Biol. Chem., 237, 841 (1962).
24 Razzell, W. E., and H. G. Khorana, J. Biol. Chem., 234, 2114 (1959).
'5 Kay, E. R. M., N. S. Simmons, and A. L. Dounce, J. Am. Chem. Soc., 74, 1724 (1952).
2' Sherman, J. R., and J. Adler, J. Biol. Chem., 238, 873 (1962).
27 Waley, S. G., and J. Watson, Biochem. J., 55, 328 (1953).
28 Akabori, S., K. Ohno, and K. Narita, Bull. Chem. Soc. Japan, 25, 214 (1952).
29 Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall, J. Biol. Chem., 193, 265
(1951).
30 Nakamoto, T., and S. B. Weiss, these PROCEEDINGS, 48, 880 (1962).
31 Fox, C. F., W. S. Robinson, and S. B. Weiss, Fed. Proc., 22, 463 (1963).
32 Krakow, J. S., and S. Ochoa, these PROCEEDINGS, 49, 88 (1963).

THE INDEPENDENCE OF THE CONTINUUM HYPOTHESIS


BY PAUL J. COHEN*
DEPARTMENT OF MATHEMATICS, STANFORD UNIVERSITY
Communicated by Kurt Godel, September 30, 1963
This is the first of two notes in which we outline a proof of the fact that the Con-
tinuum Hypothesis cannot be derived from the other axioms of set theory, including
the Axiom of Choice. Since G6dell has shown that the Continuum Hypothesis is
consistent with these axioms, the independence of the hypothesis is thus estab-
lished. We shall work with the usual axioms for Zermelo-Fraenkel set theory,2 and
by Z-F we shall denote these axioms without the Axiom of Choice, (but with the
Axiom of Regularity). By a model for Z-F we shall always mean a collection of
actual sets with the usual e-relation satisfying Z-F. We use the standard defini-
tions3 for the set of integers w, ordinal, and cardinal numbers.
THEOREM 1. There are models for Z-F in which the following occur:
(1) There is a set a, a C w such that a is not constructible in the sense of reference
3, yet the Axiom of Choice and the Generalized Continuum Hypothesis both hold.
(2) The continuum (i.e., @P(w) where (P means power set) has no well-ordering.
(3) The Axiom of Choice holds, but Ri $ 2^.
(4) The Axiom of Choice for countable pairs of elements in W(PQ(w)) fails.
Only part 3 will be discussed in this paper. In parts 1 and 3 the universe is well-
ordered by a single definable relation. Note that 4 implies that there is no simple
ordering of 6(6(P(c)). Since the Axiom of Constructibility implies the Generalized
Continuum Hypothesis,3 and the latter implies the Axiom of Choice,' Theorem 1
completely settles the question of the relative strength of these axioms.
Before giving details, we sketch the intuitive ideas involved. The starting point
is the realizations 4 that no formula a(x) can be shown from the axioms of Z-F
to have the property that the collection of all x satisfying it form a model for Z-F
in which the Axiom of Constructibility (V = L,I) fails. Thus, to find such models,
it seems natural to strengthen Z-F by postulating the existence of a set which is a
1144 M1 A THEMATICS: P. J. COHEN PROC. N. A. S.

model for Z-F, thus giving us greater flexibility in constructing new models. (In
the next paper we discuss how the question of independence, as distinct from that
of models, can be handled entirely within Z-F.) The Lbwenheim-Skolem theorem
yields the existence of a countable model on. Let 1, N2, etc., denote the correspond-
ing cardinals in 91M. Since 9 is countable, there exist distinct sets a5 C W, 0 < 6 <
N2. Put V = { (as, a5,) 6 < 6' }. We form the model 91 "generated" from 91M,
as, and V and hope to prove that in 9N the continuum has cardinality at least N2.
Of course, 91 will contain many new sets and, if the as are chosen indiscriminately,
the set N2 (in 911) may become countable in 91. Rather than determine the a5 di-
rectly, we first list all the countably many possible propositions concerning them
and decide in advance which are to be true. Only those properties which are true
in a "uniform" manner for "generic" subsets of w in M shall be true for the as in
91. For example, each as contains infinitely many primes, has no asymptotic
density, etc. If the as are chosen in such a manner, no new information will be
extracted from them in 9M which was not already contained in M, so that, e.g., N2
will remain the second uncountable cardinal. The primitive conditions n e as are
neither generically true nor false, and hence must be treated separately. Only
when given a finite set of such conditions will we be able to speak of properties
possibly being forced to hold for "generic" sets. The precise definition of forcing
will be given in Definition 6.
From now on, let 911 be a fixed countable model for Z-F, satisfying V = L, such
that x e 9 implies x c 911. If M' is a countable model without this property, de-
fine ' by transfinite induction on the rank of x, so that *I(x) = { y| 3 z e 911, z E x,
*1(z) = y}; the image 9 of 91' under ' is isomorphic to 91' with respect to E and
satisfies our requirement. Thus, the ordinals in 911 are truly ordinals. Let r > 1
be a fixed ordinal in 911, N, the corresponding cardinal in 91, and let as, 0 < 6 < NT
be subsets of w, not necessarily in 911, V = I (as, as,) 6 < 6'} .
LEMMA 1. There exist unique functions j, K1, K2, N, .-from ordinals to ordinals
definable in 911 such that
(1) j(a + 1) > j(a) and for all 1 such that j(a) + 1 < 3 < j(a + 1) the map
1 (N(13), K1(1), K2(13)) is a 1-1 correspondence between all such 13, and the set of all
triples (i, 'y, 6), 1 < i < 8, 'y < j(a), 6 < j(a). Furthermore, this map is order-
preserving if the triples are given the natural ordering S (Ref. 3, p. 36).
(2) j(0) = 3NT + 1, j(a) = supfj(13) 13 < a} if a is a limit ordinal.
(3) N(j(a)) = 0, N(j(a) + 1) = 9, Ki = O for these values.
(4) N(a), Ki(a) are zero if a < 3N,
(5) 1/f3 is as above, and N(13) = i, put J(i, K1(o), K2(1), j(a)) = 1. Also put
1(1) = j(a).
Definition 1: For a an ordinal in 911, define Fa by means of induction as follows:
(1) Fa"= if a < co.
(2) For w < a < MT, let Fa successively enumerate a5, the unordered pairs
(as, a5,) and the ordered pairs (as, as,) in any standard manner (e.g., the ordering
R on pairs of ordinals def. 7.813).
(3) For a = M71 Fa = V.
(4) For a > 3NT, if Ki(a) = 1, K2(a) = Y
if N(a) = 0, F. = {Fat| a' < a}
if 1 < i = N(a) < 8, Fa = ;i(F0, F.)
VOL. 50, 1963 MATHEMATICS: P. J. COHEN 1145

where 5, are defined as follows (Def. 9.13):

51(x, y) = {Xy}
5Y2(X, Y) = {(s, t) s E t, (s, t) E X}
53(X, Y) = - y
544(X, Y) = {(8, t) (s, t) e x, t E Y
55(X, Y) = sleX, 3t(t, s) ey}
936(x, y) = {(s, t) (s, t) E x, (t, s) e y}
57(x, y) = {(r, s, t) (r, s, t) e x, (t, r, s) e y}
5:8(x, y) = {(r, s, t)I (r, s, t) e x, (r, t, s) E Y}

if N(a) = 9, Fa = {Fa, a' < a, N(a') = 9}. We also define Ta = {F If3 < ca }.
We have introduced the case N = 9 for the convenience of later arguments. It
ensures that the ordinals in M1t are listed among the Fa in a canonical manner. Ob-
serve that since V = L holds in M, it is not difficult to see that this implies fM C91.
Denote by 9t, the set of all Fa for a e M. Note that in 9Z each Fa is a collection
of preceding FO. We shall often write a in place of Fa if a < w, and as in place of
Fs + + 1, etc., if there is no danger of confusion. If N(a) = 9, then the set Fa is
.

defined independently of as. We shall now examine statements concerning Fa


before the as are actually determined, and thus the Fa for a while shall be con-
sidered as merely formal symbols.
Definition 2: (1) x e y, x E F Fa E x, Fa E F, are formulas; (2) if so and ,6 are
formulas, so are 7po and so & q,; and (3) only (1) and (2) define formulas.
Definition 3: A Limited Statement is a formula a(xi, . . ., xn) in which all variables
are bound by a universal quantifier (Xi)a or an existential quantifier 3aXi placed in
front of it, where a is an ordinal in on. An Unlimited Statement is the same except
that no ordinals are attached to the quantifiers.
Our intention is that the variable x in (x)a or 3,x is restricted to range over all
F, with 3 < a. The symbol = is not used since by means of the Axiom of Exten-
sionality it can be avoided. We only consider statements in prenex form. Since
it is clear how to reduce negations, conjunctions, etc., of such statements to prenex
form, we shall not do so if there is no risk of confusion.
Definition 4: The rank of a limited statement a is (a, r) if r is the number of
quantifiers and a is the least ordinal such that for all fl, ,B < a if F# occurs in a, and
,B < a if (x), or 3,x occurs in a. We write (a, r) < (,B, s) if a < , or a = and
r < s.
Thus, if rank a = (a, r), a can be formulated in { F,B| , < c }.
Definition 5: Let P denote a finite set of conditions of the form n e a5 or m n e as
such that no condition and its negation are both included.
In the following definition, which is the key point of the paper, we shall define
a certain concept for all limited statements by means of transfinite induction. The
well-ordering we use is not, however, precisely the corresponding ordering of the
ranks, but requires a slight modification. We say a is of type a, if rank a = (a +
1, r), (x)a + 1 and 3a + ix do not occur in a, and no expression of the form FaE ( )
occurs in a. We order the limited statements by saying, if rank a = (a, r) and
rank b = (,B, s), a precedes b if and only if rank a < rank b, unless a = 8 and one of
1146 MATHEMATICS: P. J. COHEN PROC. N. A. S.

the two statements a, b is of type (R and the other is not of type CR, in which case the
former precedes the latter.
Definition 6: By induction, we define the concept of "P forces a" as follows:
I. If r > 0, P forces a = (x)ab(x) if for all P' v P) P' does not force 7b(Fp)
for 13 < a. P forces 3axb(X) if for some 13 < a, P forces b(FB).
II. If r = 0, and a has propositional connectives, P forces a if for each compo-
nent Fa, E Few or 7Fa e Fl appearing in a, these, by case III of this definition, are
forced to be true or their negations are forced to be true so that in the usual sense
of the propositional calculus a is true.
III. If a is of the form Fa f Fe or 7 Fa E F, we define P forces a as follows:
(i) If a, 13 < 3NT, then a must hold as a formal consequence of P, i.e., P forces
a, if a is true whenever as are distinct subsets of w, satisfying P, different from any
integer and c.
(ii) 7 Fa e Fa is always forced.
(iii) If a < 3, N(1) = i < 9, 3 > 3R,, P forces a, where aFa e F,6 or Fa e Fe,
if P forces 4ti or 7 4,j, respectively, where j,6 is the limited statement expressing the
definition of Fe. That is, if K1(13) = oy, K2(13) = 6:
(0) it0 is vacuous and always forced.
(1) 4t' Fa = F V Fa = F5.
(2) V2 3,6x 3ey (Fa = (x, y) & x e y& Fa e F).
(3) 4I3 -Fa E Fly& 7Fae Fs.
(4) 44 3ax3Y (Fa = (x, y) & Fa e Fy & y E Fs).
(5) 445- 3]x (Fa e F, & (x, Fa) e F5).
(6), (7), (8), similarly.
Here the use of ordered pairs must eventually be replaced by their definition, and
the use of equality in x = y is replaced by (z)#(z e x (===) z e y).
(iv) If a <1, N(13) = 9, ,8 > 3M, P forces a - Fa e Ff if for some 13' < ,# N(13')
= 9, P forces Fa = Ffl. P forces 7 F., e Fe, if for all 13' <1, N(13') = 9 and all
P' v P) P' does not force Fa = F,. Again the symbol = is treated as before.
(v) If a > 13, we reduce the case Fa e Fa to cases (iii) and (iv) treated above.
We say P forces Fa e Fe if for some 13' <1, P forces Fag e Fee and P forces Fa =
Ff, (i.e., (x) a(x e Fa (K) x e F,6,) which is a statement of type (R and hence precedes
Fa E Ffi). We say P forces 7 FaE F,6 if for all 1' <1 and P' D P, P' does not force
both Fa, 6 F,9 and F,> = Fa.
The most important part of Definition 6 is I, the other parts are merely obvious
derivatives of it.
Definition 7: If a is an unlimited statement with r quantifiers, we define "P
forces a" by induction on r. If r = 0, then a is a limited statement. If a
(x) b(x), P forces a, if for all P' D P, and a, P' does not force 7 b(Fa). If a
3x b(x), P forces a if for some a, P forces b(Fa).
In the proofs of Lemmas 2, 3, 4, and 5, we keep the same well-ordering on limited
statements as in Definition 6, and proceed by induction.
LEMMA 2. P does not force a and 7 a, for any a and P.
Proof: Let a be a limited statement with r quantifiers. If r > 0, and P forces
both 3ax b(x) and (x) a 7 b(x), then P must force b(F,6) for 13 < a which means P
cannot force (x) a 7 b(x). Case II of Definition 6 will clearly follow from case III.
Parts (i) and (ii) are trivial. If a is in part (iii), then P forces a if and only if P
VOL. 50, 1963 MATHEMATICS: P. J. COHEN 1147

forces a statement of lower rank and in this case the lemma follows by induction.
In part (iv), if P forces Fa e F,, then for some (3' < fl, N(,3') = 9, and P forces Fa =
Fit which means P can not force 7 Fa E Fa. In part (v) if P forces Fa e F,6, for
some (3' <(,, P forces F,6, e Fe and Fa = Fe, which again violates P forcing 7 Fc. e Fi.
If a is an unlimited statement, the lemma follows in the same manner by induction
on the number of quantifiers.
LEMMA 3. If P forces a and P' D P, then P' forces a.
Proof by induction as in Lemma 2.
LEMMA 4. For any statement a and condition P, there is P' D P such that either )I"
forces a or P'forces 7 a.
Proof: Let a be a limited statement with r quantifiers. If r > 0 and P does not
force a = (x)a b(x), then for some P' v P, P' forces 7 b(F,9), (3 < a, which means
P' forces 7 a. If r = 0, we may restrict ourselves to III, for if we enumerate the
components of a, by defining a finite sequence Pa, Po = P and P, + 1 D Pn we may
successively force each component or its negation so that finally either a or 7 a is
forced. Again, cases (i) and (ii) are trivially disposed of. Case (iii) is handled by
induction as before. If a = Fa e F, is in case (iv) then if P does not force 7 a, for
some P' v P and (3' <(, N((') = 9, P' forces Fa = F,3, so P' forces a. If a-
Fa e F, is in case (v) if P does not force 7 a, then for some P' D P, (' < f, P' forces
Fir e F, and F,3' = Fa' hence P' forces a. Unlimited statements are handled as
before.
Definition 8: Enumerate all statements an, both limited and unlimited, and all
ordinals a. in M11. Define P2n as the first extension of P2n - 1 which forces either an
or 7 an. Define P2n + 1 as the first extension of P2n which has the property that it
forces F,6 e Fan where ( is the least possible ordinal for which there exists such an
extension of P2n, whereas if no such ( exists, put P2n + 1 = P2n.
The sequence P. is not definable in M. Since all statements of the form n e as
are enumerated, Pn clearly approach in an obvious sense, sets as of integers. With
this choice of as, let 91 be defined as the set of all Fa defined by Definition 1.
LEMMA 5. All statements in 91 which are forced by some P, are true in 91 and con-
versely.
Proof: Let a be a limited statement with r quantifiers. If r > 0, then if P, forces
(x)a b(x), if (3 < a, then some Pm must force b(F06) since no Pm can force 71 b(F,3).
By induction we have that b(F,6) holds, so that (x)a b(x) holds in 91. If P. forces
3ax b(x), for some (3 < a, P, forces b(F,) so by induction b(F,6) holds and hence
3Sx b(x) holds in 9Z. Case II will clearly follow from case III and (i) and (ii) are
trivial. If a is Fa e F,3 or 7 Fa E F,3 in case (iii) then if Pn forces a, Pn forces pre-
cisely the statement which because of the definition of F,3 is equivalent to a. In
case (iv) if Pn forces Fa i F,6, for some (3' <(, N((3') = 9, P,, forces F,3t = Fa, which
therefore holds by induction in 91. If P, forces 7 Fa e F,3, then for each (' < 3,y
N(,(') = 9, Fa = F#,3 is not forced by any Pm so some Pm must force Fa id F,3d
which proves 7 Fat e F,3 holds in 91. Similarly for case (v) and for unlimited state-
ments. Since every statement or its negation is forced eventually, the converse is
also true.
Lemma 5 is the justification of the definition of forcing since we can now throw
back questions about 91 to questions about forcing which can be formulated in M11.
1148 PATHOLOGY: BLACK ET AL. PROC. N. A. S.

In the next paper, we shall prove that 9N is a model for Z-F in which part 3 of
Theorem 1 holds.
* The author is a fellow of the Alfred P. Sloan Foundation.
1 Cohen, P. J., "A minimal model for set theory," Bull. Amer. Math. Soc., 69, 537-540 (1963).
2Fraenkel, A., and Y. Bar-Hillel, Foundations of Set Theory (1958).
3 Gbdel, K., The Consistency of the Continuum Hypothesis (Princeton University Press, 1940).
4 Shepherdson, J. C., "Inner models for set theory," J. Symb. Logic, 17, 225-237 (1957).
5 Sierpinski, W., "L'hypothlse generalisee du continu et l'axiome du choix," Fund. Math., 34
1-5 (1947)

A SPECIFIC COMPLEMENT-FIXING ANTIGEN PRESENT IN


SV40 TUMOR AND TRANSFORMED CELLS
BY PAUL H. BLACK, WALLACE P. ROWE, HORACE C. TURNER, AND
ROBERT J. HUEBNER
LABORATORY OF INFECTIOUS DISEASES, NATIONAL INSTITUTES OF HEALTH, BETHESDA, MARYLAND
Communicated October 4, 1963
Many experimental tumors, both carcinogen-induced' and virus-induced,2-5
contain new cellular antigens, generally demonstrable by transplant rejection
procedures. Huebner et al.6 first demonstrated the presence of new, noninfectious,
complement-fixing (CF) antigens clearly under viral genetic control, in adenovirus-
induced tumors in hamsters and rats.
A new transplantation antigen(s) has been found in SV40-induced hamster tu-
mors,7-9 but it is not established whether its synthesis is under viral or host cell
genetic control. This paper presents evidence for a new CF antigen in SV40 tumors
and transformed tissue culture cells, formed from information contained within the
viral genome. Preliminary results of these studies were presented by Huebner et al.6
Materials and Methods.-The CF procedure was identical with that used by Huebner et al. ;6
this is a Bengtson procedure done in microplates using overnight fixation at 40C, with two exact
units of complement. Tumor extracts consisted of 10% suspensions in Eagle's basal medium,
clarified by centrifugation at 2500 rpm for 30 min, and stored at - 60'C. The extracts were
tested for antigens only if the undiluted extract was not anticomplementary (AC). The primary
SV40-induced hamster tumors used in these studies are from experiments described in detail
elsewhere.10 Tumor extracts used as standard CF antigens were selected for having high titer
reactivity with sera from tumorous hamsters.
Suspensions of both normal and transformed tissue culture cells of various species,11 as well as
tissue culture cells from a variety of hamster tumors, were prepared in the following manner.
Cells grown in 32-oz Blake bottles were washed with phosphate-buffered saline (PBS) (pH 7.2),
scraped off the glass with a rubber policeman, centrifuged at 150 g for 8 min, and resuspended in
9 volumes of PBS. These suspensions were stored at - 60'C before use.
"Viral antigen" was prepared by inoculating SV40 strain 77612 into acontinuous tissue culture
cell line (strain BSC-1) of African green monkey kidney (AGMK) at a multiplicity of about 10-4,
and harvesting the cells and fluid together when cytopathogenicity was maximal. The cell
suspension was stored at - 20'C, thawed, and used without further processing. 13
Four types of sera were used as standard reagents; to avoid undue heterogeneity of antibodies,
sera of individual animals generally were used: (1) serum from tumorous hamsters, selected for
having high CF antibody titer against SV40 tumor extracts, and no reaction with viral antigen;
(2) serum from similar hamsters, but having high CF antibody titers for both tumor extracts and

You might also like