The continuum hypothesis was advanced byGeorg Cantor in 1878,[1] and establishing its truth or falsehood is the first ofHilbert's 23 problems presented in 1900. The answer to this problem isindependent of ZFC, so that either the continuum hypothesis or its negation can be added as an axiom to ZFC set theory, with the resulting theory being consistent if and only if ZFC is consistent. This independence was proved in 1963 byPaul Cohen, complementing earlier work byKurt Gödel in 1940.[2]
The name of the hypothesis comes from the termcontinuum for the real numbers.
Cantor believed the continuum hypothesis to be true and for many years tried in vain to prove it.[3] It became the first on David Hilbert'slist of important open questions that was presented at theInternational Congress of Mathematicians in the year 1900 in Paris.Axiomatic set theory was at that point not yet formulated.Kurt Gödel proved in 1940 that the negation of the continuum hypothesis, i.e., the existence of a set with intermediate cardinality, could not be proved in standard set theory.[2] The second half of the independence of the continuum hypothesis – i.e., unprovability of the nonexistence of an intermediate-sized set – was proved in 1963 byPaul Cohen.[4]
Two sets are said to have the samecardinality orcardinal number if there exists abijection (a one-to-one correspondence) between them. Intuitively, for two sets and to have the same cardinality means that it is possible to "pair off" elements of with elements of in such a fashion that every element of is paired off with exactly one element of and vice versa. Hence, the set has the same cardinality as despite the sets themselves containing different elements.
With infinite sets such as the set ofintegers orrational numbers, the existence of a bijection between two sets becomes more difficult to demonstrate. The rational numbers seemingly form a counterexample to the continuum hypothesis: the integers form a proper subset of the rationals, which themselves form a proper subset of the reals, so intuitively, there are more rational numbers than integers and more real numbers than rational numbers. However, this intuitive analysis is flawed since it does not take into account the fact that all three sets areinfinite. Perhaps more importantly, it in fact conflates the concept of "size" of the set with the order or topological structure placed on it. In fact, it turns out the rational numbers can actually be placed in one-to-one correspondence with the integers, and therefore the set of rational numbers is the same size (cardinality) as the set of integers: they are bothcountable sets.[5]
Cantor gave two proofs that the cardinality of the set ofintegers is strictly smaller than that of the set ofreal numbers (seeCantor's first uncountability proof andCantor's diagonal argument). His proofs, however, give no indication of the extent to which the cardinality of the integers is less than that of the real numbers. Cantor proposed the continuum hypothesis as a possible solution to this question.
In simple terms, the Continuum Hypothesis (CH) states that the set of real numbers has minimal possible cardinality which is greater than the cardinality of the set of integers. That is, every infinite set of real numbers can either be mapped one-to-one into the integers or the real numbers can be mapped one-to-one into. Since the real numbers areequinumerous with thepowerset of the integers, i.e., CH can be restated as follows:
Continuum Hypothesis—.
Assuming theaxiom of choice, there is a unique smallest cardinal number greater than, and the continuum hypothesis is in turn equivalent to the equality.[6][7]
Gödel[8][2]showed that CH cannot be disproved from ZF, even if theaxiom of choice (AC) is adopted, i.e. from ZFC. Gödel's proof shows that both CH and AC hold in theconstructible universe, aninner model of ZF set theory, assuming only the axioms of ZF. The existence of an inner model of ZF in which additional axioms hold shows that the additional axioms are (relatively)consistent with ZF, provided ZF itself is consistent. The latter condition cannot be proved in ZF itself, due toGödel's incompleteness theorems, but is widely believed to be true and can be proved in stronger set theories.
Cohen[4][9] showed that CH cannot be proven from the ZFC axioms, completing the overall independence proof. To prove his result, Cohen developed the method offorcing, which has become a standard tool in set theory. Essentially, this method begins with a model of ZF in which CH holds and constructs another model which contains more sets than the original in a way that CH does not hold in the new model. Cohen was awarded theFields Medal in 1966 for his proof.
Cohen's independence proof shows that CH is independent of ZFC. Further research has shown that CH is independent of all knownlarge cardinal axioms in the context of ZFC.[10] Moreover, it has been shown that thecardinality of the continuum can be any cardinal consistent withKőnig's theorem. A result of Solovay, proved shortly after Cohen's result on the independence of the continuum hypothesis, shows that in any model of ZFC, if is a cardinal of uncountablecofinality, then there is a forcing extension in which. However, per Kőnig's theorem, it is not consistent to assume is or or any cardinal with cofinality.
The continuum hypothesis is closely related to many statements inanalysis, point settopology andmeasure theory. As a result of its independence, many substantialconjectures in those fields have subsequently been shown to be independent as well.
The independence from ZFC means that proving or disproving the CH within ZFC is impossible. However, Gödel and Cohen's negative results are not universally accepted as disposing of all interest in the continuum hypothesis. The continuum hypothesis remains an active topic of research: seeWoodin[11][12] andKoellner[13] for an overview of the current research status.
The continuum hypothesis and theaxiom of choice were among the first genuinely mathematical statements shown to be independent of ZF set theory. Although the existence of some statements independent of ZFC had already been known more than two decades prior: for example, assuminggood soundness properties and the consistency of ZFC,Gödel's incompleteness theorems published in 1931 establish that there is a formal statement Con(ZFC) (one for each appropriateGödel numbering scheme) expressing the consistency of ZFC, that is also independent of it. The latter independence result indeed holds for many theories.
Arguments for and against the continuum hypothesis
Gödel believed that CH is false, and that his proof that CH is consistent with ZFC only shows that theZermelo–Fraenkel axioms do not adequately characterize the universe of sets. Gödel was aPlatonist and therefore had no problems with asserting the truth and falsehood of statements independent of their provability. Cohen, though aformalist,[14] also tended towards rejecting CH.
Historically, mathematicians who favored a "rich" and "large"universe of sets were against CH, while those favoring a "neat" and "controllable" universe favored CH. Parallel arguments were made for and against theaxiom of constructibility, which implies CH. More recently,Matthew Foreman has pointed out thatontological maximalism can actually be used to argue in favor of CH, because among models that have the same reals, models with "more" sets of reals have a better chance of satisfying CH.[15]
Another viewpoint is that the conception of set is not specific enough to determine whether CH is true or false. This viewpoint was advanced as early as 1923 bySkolem, even before Gödel's first incompleteness theorem. Skolem argued on the basis of what is now known asSkolem's paradox, and it was later supported by the independence of CH from the axioms of ZFC since these axioms are enough to establish the elementary properties of sets and cardinalities. In order to argue against this viewpoint, it would be sufficient to demonstrate new axioms that are supported by intuition and resolve CH in one direction or another. Although theaxiom of constructibility does resolve CH, it is not generally considered to be intuitively true any more than CH is generally considered to be false.[16]
At least two other axioms have been proposed that have implications for the continuum hypothesis, although these axioms have not currently found wide acceptance in the mathematical community. In 1986, Chris Freiling[17] presented an argument against CH by showing that the negation of CH is equivalent toFreiling's axiom of symmetry, a statement derived by arguing from particular intuitions aboutprobabilities. Freiling believes this axiom is "intuitively clear"[17] but others have disagreed.[18][19]
A difficult argument against CH developed byW. Hugh Woodin has attracted considerable attention since the year 2000.[11][12]Foreman does not reject Woodin's argument outright but urges caution.[20] Woodin proposed a new hypothesis that he labeled the"(*)-axiom", or "Star axiom". The Star axiom would imply that is, thus falsifying CH. The Star axiom was bolstered by an independent May 2021 proof showing the Star axiom can be derived from a variation ofMartin's maximum. However, Woodin stated in the 2010s that he now instead believes CH to be true, based on his belief in his new "ultimate L" conjecture.[21][22]
Solomon Feferman argued that CH is not a definite mathematical problem.[23] He proposed a theory of "definiteness" using a semi-intuitionistic subsystem of ZF that acceptsclassical logic for bounded quantifiers but usesintuitionistic logic for unbounded ones, and suggested that a proposition is mathematically "definite" if the semi-intuitionistic theory can prove. He conjectured that CH is not definite according to this notion, and proposed that CH should, therefore, be considered not to have a truth value.Peter Koellner wrote a critical commentary on Feferman's article.[24]
Joel David Hamkins proposes amultiverse approach to set theory and argues that "the continuum hypothesis is settled on the multiverse view by our extensive knowledge about how it behaves in the multiverse, and, as a result, it can no longer be settled in the manner formerly hoped for".[25] In a related vein,Saharon Shelah wrote that he does "not agree with the pure Platonic view that the interesting problems in set theory can be decided, that we just have to discover the additional axiom. My mental picture is that we have many possible set theories, all conforming to ZFC".[26]
Thegeneralized continuum hypothesis (GCH) states that if an infinite set's cardinality lies between that of an infinite setS and that of thepower set ofS, then it has the same cardinality as eitherS or. That is, for anyinfinite cardinal there is no cardinal such that. GCH is equivalent to:
Thebeth numbers provide an alternative notation for this condition: for every ordinal. The continuum hypothesis is the special case for the ordinal. GCH was first suggested byPhilip Jourdain.[27] For the early history of GCH, see Moore.[28]
Like CH, GCH is also independent of ZFC, butSierpiński proved that ZF + GCH implies theaxiom of choice (AC) (and therefore the negation of theaxiom of determinacy, AD), so choice and GCH are not independent in ZF; there are no models of ZF in which GCH holds and AC fails. To prove this, Sierpiński showed GCH implies that every cardinality n is smaller than somealeph number, and thus can be ordered. This is done by showing that n is smaller than which is smaller than its ownHartogs number—this uses the equality; for the full proof, see Gillman.[29]
Kurt Gödel showed that GCH is a consequence of ZF +V=L (the axiom that every set is constructible relative to the ordinals), and is therefore consistent with ZFC. As GCH implies CH, Cohen's model in which CH fails is a model in which GCH fails, and thus GCH is not provable from ZFC. W. B. Easton used the method of forcing developed by Cohen to proveEaston's theorem, which shows it is consistent with ZFC for arbitrarily large cardinals to fail to satisfy. Much later,Foreman andWoodin proved that (assuming the consistency of very large cardinals) it is consistent that holds for every infinite cardinal. Later Woodin extended this by showing the consistency of for every. Carmi Merimovich[30] showed that, for eachn ≥ 1, it is consistent with ZFC that for each infinite cardinalκ,2κ is thenth successor ofκ (assuming the consistency of some large cardinal axioms). On the other hand, László Patai[31] proved that ifγ is an ordinal and for each infinite cardinalκ,2κ is theγth successor ofκ, thenγ is finite.
For any infinite setsA andB, if there is an injection fromA toB then there is an injection from subsets ofA to subsets ofB. Thus for any infinite cardinalsA andB,. IfA andB are finite, the stronger inequality holds. GCH implies that this strict, stronger inequality holds for infinite cardinals as well as finite cardinals.
Although the generalized continuum hypothesis refers directly only to cardinal exponentiation with 2 as the base, one can deduce from it the values of cardinal exponentiation in all cases. GCH implies that for ordinalsα andβ:[32]
whenα ≤β+1;
whenβ+1 <α and, wherecf is thecofinality operation; and
^Goodman, Nicolas D. (1979). "Mathematics as an objective science".The American Mathematical Monthly.86 (7):540–551.doi:10.2307/2320581.JSTOR2320581.MR0542765.This view is often calledformalism. Positions more or less like this may be found in Haskell Curry [5], Abraham Robinson [17], and Paul Cohen [4].
^abFreiling, Chris (1986). "Axioms of Symmetry: Throwing darts at the real number line".Journal of Symbolic Logic.51 (1). Association for Symbolic Logic:190–200.doi:10.2307/2273955.JSTOR2273955.S2CID38174418.
^Hamkins, Joel David (January 2015). "Is the Dream Solution of the Continuum Hypothesis Attainable?".Notre Dame Journal of Formal Logic.56 (1).arXiv:1203.4026.doi:10.1215/00294527-2835047.
^Rittberg, Colin J. (March 2015). "How Woodin changed his mind: new thoughts on the Continuum Hypothesis".Archive for History of Exact Sciences.69 (2):125–151.doi:10.1007/s00407-014-0142-8.S2CID122205863.
^Moore, Gregory H. (2011). "Early history of the generalized continuum hypothesis: 1878–1938".Bulletin of Symbolic Logic.17 (4):489–532.doi:10.2178/bsl/1318855631.MR2896574.
^Patai, L. (1930). "Untersuchungen über die א-reihe".Mathematische und naturwissenschaftliche Berichte aus Ungarn (in German).37:127–142.
^Hayden, Seymour; Kennison, John F. (1968).Zermelo-Fraenkel Set Theory. Columbus, Ohio: Charles E. Merrill. p. 147, exercise 76.
Maddy, Penelope (June 1988). "Believing the axioms, [part I]".Journal of Symbolic Logic.53 (2). Association for Symbolic Logic:481–511.doi:10.2307/2274520.JSTOR2274520.
Dales, H.G.; Woodin, W.H. (1987).An Introduction to Independence for Analysts. Cambridge.
Enderton, Herbert (1977).Elements of Set Theory. Academic Press.
Gödel, K.:What is Cantor's Continuum Problem?, reprinted in Benacerraf and Putnam's collectionPhilosophy of Mathematics, 2nd ed., Cambridge University Press, 1983. An outline of Gödel's arguments against CH.
Martin, D. (1976). "Hilbert's first problem: the continuum hypothesis," inMathematical Developments Arising from Hilbert's Problems, Proceedings of Symposia in Pure Mathematics XXVIII, F. Browder, editor. American Mathematical Society, 1976, pp. 81–92.ISBN0-8218-1428-1