The Continuum Hypothesis - Logic at Harvard
The Continuum Hypothesis - Logic at Harvard
The Continuum Hypothesis - Logic at Harvard
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
the Ω Conjecture in the above arguments. This is the st<strong>at</strong>ement th<strong>at</strong> theΩ Conjecture is (non-trivially) Ω-s<strong>at</strong>isfiable, th<strong>at</strong> is, the st<strong>at</strong>ement: <strong>The</strong>reexists an ordinal α and a universe V ′ of the multiverse such th<strong>at</strong>andV ′ α |= ZFC+“<strong>The</strong>re is a proper class of Woodin cardinals”V ′ α |= “<strong>The</strong> Ω Conjecture”.This Σ 2 -st<strong>at</strong>ement is invariant under set forcing and hence is one adherentsto the generic multiverse view of truth must deem determin<strong>at</strong>e. Moreover,the key arguments above go through with this Σ 2 -st<strong>at</strong>ement instead of theΩ Conjecture. <strong>The</strong> person taking this second line of response would thusalso have to maintain th<strong>at</strong> this st<strong>at</strong>ement is false. But there is substantialevidence th<strong>at</strong> this st<strong>at</strong>ement is true. <strong>The</strong> reason is th<strong>at</strong> there is no knownexample of a Σ 2 -st<strong>at</strong>ement th<strong>at</strong> is invariant under set forcing rel<strong>at</strong>ive to largecardinal axioms and which cannot be settled by large cardinal axioms. (Sucha st<strong>at</strong>ement would be a candid<strong>at</strong>e for an absolutely undecidable st<strong>at</strong>ement.)So it is reasonable to expect th<strong>at</strong> this st<strong>at</strong>ement is resolved by large cardinalaxioms. However, recent advances in inner model theory—in particular,those in Woodin (2011b)—provide evidence th<strong>at</strong> no large cardinal axiom canrefute this st<strong>at</strong>ement. Putting everything together: It is very likely th<strong>at</strong> thisst<strong>at</strong>ement is in fact true; so this line of response is not promising.Third, one could reject either the Truth Constraint or the DefinabilityConstraint. <strong>The</strong> trouble is th<strong>at</strong> if one rejects the Truth Constraint then onthis view (assuming the Ω Conjecture) Π 2 truth in set theory is reducible inthe sense of Turing reducibility to truth in H(δ 0 ) (or, assuming the StrongΩ Conjecture, H(c + )). And if one rejects the Definability Constraint thenon this view (assuming the Ω Conjecture) Π 2 truth in set theory is reduciblein the sense of definability to truth in H(δ 0 ) (or, assuming the Strong ΩConjecture, H(c + )). On either view, the reduction is in tension with theacceptance of non-pluralism regarding the background theory ZFC+“<strong>The</strong>reis a proper class of Woodin cardinals”.Fourth, onecouldembracethecriticism, rejectthegenericmultiverse conceptionof truth, and admit th<strong>at</strong> there are some st<strong>at</strong>ements about H(δ + 0 ) (orH(c + ), granting, in addition, the AD + Conjecture) th<strong>at</strong> are true simpliciterbut not true in the sense of the generic-multiverse, and yet nevertheless continueto maintain th<strong>at</strong> CH is indetermin<strong>at</strong>e. <strong>The</strong> difficulty is th<strong>at</strong> any suchsentence ϕ is qualit<strong>at</strong>ively just like CH in th<strong>at</strong> it can be forced to hold and26