Subversions of forcing classes

Transcription

1 Subversions of forcing classes Gunter Fuchs July 19, st Girona conference on inner model theory CUNY College of Staten Island and the CUNY Graduate Center

2 Proper and subproper forcing

3 Proper forcing Here is an unfamiliar way of defining a familiar class of forcing notions: 1

4 Proper forcing Here is an unfamiliar way of defining a familiar class of forcing notions: Definition A forcing notion P is proper if the following holds: for all sufficiently large θ, if P P H θ N L τ ras ù ZFC, σ : N N with P P ranpσq, N countable and p P P X ranpσq, then there is a q p such that whenever G Q q is P-generic, then Ḡ σ 1 G is P σ 1 ppq-generic over N. 1

5 Proper forcing Here is an unfamiliar way of defining a familiar class of forcing notions: Definition A forcing notion P is proper if the following holds: for all sufficiently large θ, if P P H θ N L τ ras ù ZFC, σ : N N with P P ranpσq, N countable and p P P X ranpσq, then there is a q p such that whenever G Q q is P-generic, then Ḡ σ 1 G is P σ 1 ppq-generic over N. It follows in this situation that σ can be extended to an elementary embedding σ σ : NrḠs NrGs. 1

6 ...the sub(-)version The idea behind the sub-version of properness, called subproperness, is to weaken the definition of properness by requiring that there is some elementary embedding σ 1 in VrGs that is somewhat similar to σ and moves the generic correctly. 2

7 ...the sub(-)version The idea behind the sub-version of properness, called subproperness, is to weaken the definition of properness by requiring that there is some elementary embedding σ 1 in VrGs that is somewhat similar to σ and moves the generic correctly. Definition (first go) A forcing notion P is subproper if the following holds: for all sufficiently large θ, if P P H θ N L τ ras ù ZFC, σ : N N with P P ranpσq, N countable and p P P X ranpσq, 2

8 ...the sub(-)version The idea behind the sub-version of properness, called subproperness, is to weaken the definition of properness by requiring that there is some elementary embedding σ 1 in VrGs that is somewhat similar to σ and moves the generic correctly. Definition (first go) A forcing notion P is subproper if the following holds: for all sufficiently large θ, if P P H θ N L τ ras ù ZFC, σ : N N with P P ranpσq, N countable and p P P X ranpσq, and a 0, a 1,..., a n 1 P N, 2

9 ...the sub(-)version The idea behind the sub-version of properness, called subproperness, is to weaken the definition of properness by requiring that there is some elementary embedding σ 1 in VrGs that is somewhat similar to σ and moves the generic correctly. Definition (first go) A forcing notion P is subproper if the following holds: for all sufficiently large θ, if P P H θ N L τ ras ù ZFC, σ : N N with P P ranpσq, N countable and p P P X ranpσq, and a 0, a 1,..., a n 1 P N, then there is a q p such that whenever G Q q is P-generic, then in VrGs, there is an elementary embedding σ 1 : N N such that, letting P σ 1 ppq, 1. σ 1 p Pq P and σpa i q σ 1 pa i q, for all i n. 2

10 ...the sub(-)version The idea behind the sub-version of properness, called subproperness, is to weaken the definition of properness by requiring that there is some elementary embedding σ 1 in VrGs that is somewhat similar to σ and moves the generic correctly. Definition (first go) A forcing notion P is subproper if the following holds: for all sufficiently large θ, if P P H θ N L τ ras ù ZFC, σ : N N with P P ranpσq, N countable and p P P X ranpσq, and a 0, a 1,..., a n 1 P N, then there is a q p such that whenever G Q q is P-generic, then in VrGs, there is an elementary embedding σ 1 : N N such that, letting P σ 1 ppq, 1. σ 1 p Pq P and σpa i q σ 1 pa i q, for all i n. 2. Ḡ pσ 1 q 1 G is P-generic over N. 2

11 ...the sub(-)version The idea behind the sub-version of properness, called subproperness, is to weaken the definition of properness by requiring that there is some elementary embedding σ 1 in VrGs that is somewhat similar to σ and moves the generic correctly. Definition (first go) A forcing notion P is subproper if the following holds: for all sufficiently large θ, if P P H θ N L τ ras ù ZFC, σ : N N with P P ranpσq, N countable and p P P X ranpσq, and a 0, a 1,..., a n 1 P N, then there is a q p such that whenever G Q q is P-generic, then in VrGs, there is an elementary embedding σ 1 : N N such that, letting P σ 1 ppq, 1. σ 1 p Pq P and σpa i q σ 1 pa i q, for all i n. 2. Ḡ pσ 1 q 1 G is P-generic over N. 3. Hull N pδ Y ranpσqq Hull N pδ Y ranpσ 1 qq, where δ δppq is the minimal size of a dense subset of P. 2

12 Fullness The problem with this definition is that it is too restrictive. For example, if every element of ranpσq is definable in N from a 0,..., a n 1, then σ 1 would have to be equal to σ. 3

13 Fullness The problem with this definition is that it is too restrictive. For example, if every element of ranpσq is definable in N from a 0,..., a n 1, then σ 1 would have to be equal to σ. Thus, Jensen required the existence of a σ 1 only in the case that N is full, meaning that there is an ordinal γ 0 such that L γ p Nq ù ZFC and N is regular in M, i.e., if f : x ÝÑ N, where f P L γ p Nq, then ranpf q P N. 3

14 Subproperness Definition (Jensen) A forcing notion P is subproper if the following holds: for all sufficiently large θ, if P P H θ N L τ ras ù ZFC, σ : N N with P P ranpσq, N countable and full, p P P X ranpσq, and a 0, a 1,..., a n 1 P N, then there is a q p such that whenever G Q q is P-generic, then in VrGs, there is an elementary embedding σ 1 : N N such that, letting P σ 1 ppq, 1. σ 1 p Pq P and σpa i q σ 1 pa i q, for all i n. 2. Ḡ pσ 1 q 1 G is P-generic over N. 3. Hull N pδ Y ranpσqq Hull N pδ Y ranpσ 1 qq, where δ δppq. 4

15 Complete and subcomplete forcing

16 Completeness In chapter V of Shelah s Proper and improper forcing, he defines the concept of E-completeness, for a family E rµs ω. Jensen calls a special case of this condition just completeness, and one can express it as follows: 5

17 Completeness In chapter V of Shelah s Proper and improper forcing, he defines the concept of E-completeness, for a family E rµs ω. Jensen calls a special case of this condition just completeness, and one can express it as follows: Definition P is complete if there is a θ with P P H θ such that the following holds: if σ : N N where N is of the form L τ ras, H θ N, N ù ZFC, N is countable and transitive, P P ranpσq, 5

18 Completeness In chapter V of Shelah s Proper and improper forcing, he defines the concept of E-completeness, for a family E rµs ω. Jensen calls a special case of this condition just completeness, and one can express it as follows: Definition P is complete if there is a θ with P P H θ such that the following holds: if σ : N N where N is of the form L τ ras, H θ N, N ù ZFC, N is countable and transitive, P P ranpσq, and if Ḡ is P σ 1 ppq-generic over N, then there is a condition p P P such that whenever G Q p is P-generic, it follows that σ Ḡ G. 5

19 Subcompleteness Jensen showed that complete forcings are equivalent to countably closed ones. 6

20 Subcompleteness Jensen showed that complete forcings are equivalent to countably closed ones. The subversion of completeness is more interesting: 6

21 Subcompleteness Jensen showed that complete forcings are equivalent to countably closed ones. The subversion of completeness is more interesting: Definition (Jensen) P is subcomplete if there is a θ with P P H θ such that the following holds: if σ : N N where N is of the form L τ ras, H θ N, N ù ZFC, N is countable, transitive and full, P P ranpσq, a 0,..., a n 1 P N and Ḡ is P σ 1 ppq-generic over N, then there is a condition p P P such that whenever G Q p is P-generic, then in VrGs, there is a σ 1 : N N with 1. σ 1 p Pq P and σpa i q σ 1 pa i q, for all i n. 2. pσ 1 q Ḡ G. 3. Hull N pδ Y ranpσqq Hull N pδ Y ranpσ 1 qq, where δ δppq. 6

22 Basic properties All of these forcing classes preserve ω 1. 7

23 Basic properties All of these forcing classes preserve ω 1. Since σæω σ 1 æω, subcomplete forcing cannot add reals. 7

24 Basic properties All of these forcing classes preserve ω 1. Since σæω σ 1 æω, subcomplete forcing cannot add reals. Subcomplete forcing preserves Souslin trees, the principle, the set of cofinal branches through an ω 1 -tree, etc. 7

25 Basic properties All of these forcing classes preserve ω 1. Since σæω σ 1 æω, subcomplete forcing cannot add reals. Subcomplete forcing preserves Souslin trees, the principle, the set of cofinal branches through an ω 1 -tree, etc. Examples of subcomplete forcings: Namba forcing (assuming CH, Jensen) Příkrý forcing (Jensen) generalized Příkrý forcing (Minden) Magidor forcing (to change cofinality to ω 1, F.) Every ω 2 -distributive forcing (modulo forcing equivalence, F.) 7

26 Basic properties All of these forcing classes preserve ω 1. Since σæω σ 1 æω, subcomplete forcing cannot add reals. Subcomplete forcing preserves Souslin trees, the principle, the set of cofinal branches through an ω 1 -tree, etc. Examples of subcomplete forcings: Namba forcing (assuming CH, Jensen) Příkrý forcing (Jensen) generalized Příkrý forcing (Minden) Magidor forcing (to change cofinality to ω 1, F.) Every ω 2 -distributive forcing (modulo forcing equivalence, F.) Jensen showed that subcomplete and subproper forcing can be iterated with revised countable support in the style of Donder and Fuchs, with intermediate collapses to ω 1. 7

27 Relationships between the forcing classes ccc complete subcomplete proper subproper 8

28 Forcing principles

29 Forcing principles I want to look at the following forcing principles: 9

30 Forcing principles I want to look at the following forcing principles: Forcing axioms Bounded forcing axioms Maximality principles Focus: subcomplete forcing 9

31 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. 10

32 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. Familiar instances: 10

33 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. Familiar instances: Martin s Axiom at ω 1, MA ω1 (Γ the collection of all c.c.c. forcings) 10

34 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. Familiar instances: Martin s Axiom at ω 1, MA ω1 (Γ the collection of all c.c.c. forcings) PFA (Γ the collection of all proper forcings) 10

35 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. Familiar instances: Martin s Axiom at ω 1, MA ω1 (Γ the collection of all c.c.c. forcings) PFA (Γ the collection of all proper forcings) SPFA (Γ the collection of all semi-proper forcings) 10

36 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. Familiar instances: Martin s Axiom at ω 1, MA ω1 (Γ the collection of all c.c.c. forcings) PFA (Γ the collection of all proper forcings) SPFA (Γ the collection of all semi-proper forcings) MM (Γ the collection of all stationary set preserving forcings), equivalent to SPFA (Shelah) 10

37 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. Familiar instances: Martin s Axiom at ω 1, MA ω1 (Γ the collection of all c.c.c. forcings) PFA (Γ the collection of all proper forcings) SPFA (Γ the collection of all semi-proper forcings) MM (Γ the collection of all stationary set preserving forcings), equivalent to SPFA (Shelah) MA ω1 is equiconsistent with ZFC, while PFA, SPFA and MM have considerable consistency strength. 10

38 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. Familiar instances: Martin s Axiom at ω 1, MA ω1 (Γ the collection of all c.c.c. forcings) PFA (Γ the collection of all proper forcings) SPFA (Γ the collection of all semi-proper forcings) MM (Γ the collection of all stationary set preserving forcings), equivalent to SPFA (Shelah) MA ω1 is equiconsistent with ZFC, while PFA, SPFA and MM have considerable consistency strength. The instance I m mainly interested in here: 10

39 Martin s Axiom For a class Γ of forcings, FA Γ says that for any ω 1 -sized collection of dense subsets of a forcing in Γ, there is a filter that meets each of the dense sets in the collection. Familiar instances: Martin s Axiom at ω 1, MA ω1 (Γ the collection of all c.c.c. forcings) PFA (Γ the collection of all proper forcings) SPFA (Γ the collection of all semi-proper forcings) MM (Γ the collection of all stationary set preserving forcings), equivalent to SPFA (Shelah) MA ω1 is equiconsistent with ZFC, while PFA, SPFA and MM have considerable consistency strength. The instance I m mainly interested in here: The Subcomplete Forcing Axiom, SCFA (Γ the collection of all subcomplete forcings) 10

40 Forcing SCFA Assuming the existence of a supercompact cardinal κ, one can iterate proper forcings with countable support, with iterands given by a Laver function for the supercompactness of κ, producing a model in which PFA κ ω 2 2 ω holds. (Baumgartner) 11

41 Forcing SCFA Assuming the existence of a supercompact cardinal κ, one can iterate proper forcings with countable support, with iterands given by a Laver function for the supercompactness of κ, producing a model in which PFA κ ω 2 2 ω holds. (Baumgartner) This can be modified to work for SPFA, by iterating semi-proper forcings with rcs, inserting collapses to ω 1 after each step in the iteration. (Foreman-Magidor-Shelah) 11

42 Forcing SCFA Assuming the existence of a supercompact cardinal κ, one can iterate proper forcings with countable support, with iterands given by a Laver function for the supercompactness of κ, producing a model in which PFA κ ω 2 2 ω holds. (Baumgartner) This can be modified to work for SPFA, by iterating semi-proper forcings with rcs, inserting collapses to ω 1 after each step in the iteration. (Foreman-Magidor-Shelah) This can be modified to work for SCFA, by iterating subcomplete forcings. During the iteration, CH, and even will be forced, and since no reals are added, and is preserved, the final model will satisfy SCFA κ ω 2. (Jensen) 11

43 Bounded forcing axioms For a forcing class Γ, the bounded forcing axiom BFA Γ, introduced by Goldstern-Shelah, says that if P P Γ and is a collection of ω 1 many maximal antichains in B r.o.ppq, each of which has size at most ω 1, then there is a filter in B that meets them all. 12

44 Bounded forcing axioms For a forcing class Γ, the bounded forcing axiom BFA Γ, introduced by Goldstern-Shelah, says that if P P Γ and is a collection of ω 1 many maximal antichains in B r.o.ppq, each of which has size at most ω 1, then there is a filter in B that meets them all. Bagaria has characterized this principle by saying that whenever P P Γ, then xh ω2, Py Σ1 xh ω2, Py VP which means that for every a P H ω2 and every Σ 1 -formula ϕpxq, xh ω2, Py ù ϕras iff, P pxh ω2, Py ù ϕrasq. 12

45 Bounded forcing axioms For a forcing class Γ, the bounded forcing axiom BFA Γ, introduced by Goldstern-Shelah, says that if P P Γ and is a collection of ω 1 many maximal antichains in B r.o.ppq, each of which has size at most ω 1, then there is a filter in B that meets them all. Bagaria has characterized this principle by saying that whenever P P Γ, then xh ω2, Py Σ1 xh ω2, Py VP which means that for every a P H ω2 and every Σ 1 -formula ϕpxq, xh ω2, Py ù ϕras iff, P pxh ω2, Py ù ϕrasq. If Γ has the property that whenever p P P P Γ, then P p is equivalent to some Q P Γ, then this characterization can be equivalently expressed by saying that whenever P P Γ and G is P-generic, it follows that xh ω2, Py Σ1 xh ω2, Py VrGs 12

46 Consistency strengths Theorem (Goldstern-Shelah) The consistency strength of BPFA, the bounded proper forcing axiom, is a reflecting cardinal. 13

47 Consistency strengths Theorem (Goldstern-Shelah) The consistency strength of BPFA, the bounded proper forcing axiom, is a reflecting cardinal. Theorem (F.) The same is true of the bounded subcomplete forcing axiom, BSCFA. 13

48 Consistency strengths Theorem (Goldstern-Shelah) The consistency strength of BPFA, the bounded proper forcing axiom, is a reflecting cardinal. Theorem (F.) The same is true of the bounded subcomplete forcing axiom, BSCFA. Observation The same is true of the bounded subproper forcing axiom. 13

49 Maximality principles A way to generalize the bounded forcing axiom for Γ by 14

50 Maximality principles A way to generalize the bounded forcing axiom for Γ by 1. not only talking about satisfaction in H ω2, but in V, 14

51 Maximality principles A way to generalize the bounded forcing axiom for Γ by 1. not only talking about satisfaction in H ω2, but in V, 2. not only talking about Σ 1 -formulas, but any statements that persist to forcing extensions by forcings in Γ. 14

52 Maximality principles A way to generalize the bounded forcing axiom for Γ by 1. not only talking about satisfaction in H ω2, but in V, 2. not only talking about Σ 1 -formulas, but any statements that persist to forcing extensions by forcings in Γ. Definition (Väänänen-Stavi, Hamkins) Thus, MP Γ pxq is the scheme expressing that for every formula ϕ with parameters in X, if ϕ can be forced to be true by a forcing in Γ in such a way that it stays true in any further forcing extension by a forcing in Γ, then ϕ is true. 14

53 Maximality principles A way to generalize the bounded forcing axiom for Γ by 1. not only talking about satisfaction in H ω2, but in V, 2. not only talking about Σ 1 -formulas, but any statements that persist to forcing extensions by forcings in Γ. Definition (Väänänen-Stavi, Hamkins) Thus, MP Γ pxq is the scheme expressing that for every formula ϕ with parameters in X, if ϕ can be forced to be true by a forcing in Γ in such a way that it stays true in any further forcing extension by a forcing in Γ, then ϕ is true. Depending on Γ, different parameter sets X are reasonable. 14

54 Maximality principles A way to generalize the bounded forcing axiom for Γ by 1. not only talking about satisfaction in H ω2, but in V, 2. not only talking about Σ 1 -formulas, but any statements that persist to forcing extensions by forcings in Γ. Definition (Väänänen-Stavi, Hamkins) Thus, MP Γ pxq is the scheme expressing that for every formula ϕ with parameters in X, if ϕ can be forced to be true by a forcing in Γ in such a way that it stays true in any further forcing extension by a forcing in Γ, then ϕ is true. Depending on Γ, different parameter sets X are reasonable. In the case where Γ is the class of subcomplete forcing, the maximal parameter set possible is H ω2, and the resulting boldface principle is MP SC ph ω2 q. The lightface principle allows for no parameters and is denoted MP SC. But parameters from H ω1 are free, it turns out. 14

55 Consistency strengths As with maximality principles for other canonical forcing classes, the consistency strength of MP SC is just ZFC, while MP SC ph ω2 q is equiconsistent with ZFC, together with the scheme expressing that δ is regular and V δ V (δ is a constant symbol here). 15

56 Consistency strengths As with maximality principles for other canonical forcing classes, the consistency strength of MP SC is just ZFC, while MP SC ph ω2 q is equiconsistent with ZFC, together with the scheme expressing that δ is regular and V δ V (δ is a constant symbol here). The original arguments due to Hamkins work in the context of subcomplete forcing without difficulty (Minden). 15

57 Consistency strengths As with maximality principles for other canonical forcing classes, the consistency strength of MP SC is just ZFC, while MP SC ph ω2 q is equiconsistent with ZFC, together with the scheme expressing that δ is regular and V δ V (δ is a constant symbol here). The original arguments due to Hamkins work in the context of subcomplete forcing without difficulty (Minden). It also works for subproper forcing. 15

58 Consequences The principle MP SC implies CH, and even, because each can be forced to be true by a subcomplete forcing (for example, by Addpω 1, 1q), and once true, it persists to further subcomplete forcing extensions. 16

59 Consequences The principle MP SC implies CH, and even, because each can be forced to be true by a subcomplete forcing (for example, by Addpω 1, 1q), and once true, it persists to further subcomplete forcing extensions. Observe that the usual forcing axiom FA Γ gets stronger as Γ is enlarged. Thus, MM implies SCFA. So, since MM implies CH, SCFA does not imply CH. 16

60 Consequences The principle MP SC implies CH, and even, because each can be forced to be true by a subcomplete forcing (for example, by Addpω 1, 1q), and once true, it persists to further subcomplete forcing extensions. Observe that the usual forcing axiom FA Γ gets stronger as Γ is enlarged. Thus, MM implies SCFA. So, since MM implies CH, SCFA does not imply CH. The maximality principles don t have this monotonicity property. 16

61 Geology

62 Motivation There is a long history of results on the existence of definable well-orders of Ppωq or even Ppω 1 q, assuming the (bounded) proper forcing axiom. 17

63 Motivation There is a long history of results on the existence of definable well-orders of Ppωq or even Ppω 1 q, assuming the (bounded) proper forcing axiom. I want to present a result that goes in a similar direction, but assumes MP SC instead, together with a (natural?) assumption on the set-theoretic geology of the ambient universe. 17

64 Motivation There is a long history of results on the existence of definable well-orders of Ppωq or even Ppω 1 q, assuming the (bounded) proper forcing axiom. I want to present a result that goes in a similar direction, but assumes MP SC instead, together with a (natural?) assumption on the set-theoretic geology of the ambient universe. Since MP SC implies CH, we are in a very different situation than with BPFA. The methods used are also very different. 17

65 Geology Definition A ground is an inner model W such that V Wrgs, for some filter g W-generic for a poset P P W. 18

66 Geology Definition A ground is an inner model W such that V Wrgs, for some filter g W-generic for a poset P P W. Fact (Ground model enumerability; Laver, F.-Hamkins-Reitz) There is a formula ϕpx, yq such that for every p P V, the class W p tx ϕpx, pqu is a ground, and such that every ground is of the form W p, for some parameter p. 18

67 Geology Definition A ground is an inner model W such that V Wrgs, for some filter g W-generic for a poset P P W. Fact (Ground model enumerability; Laver, F.-Hamkins-Reitz) There is a formula ϕpx, yq such that for every p P V, the class W p tx ϕpx, pqu is a ground, and such that every ground is of the form W p, for some parameter p. Hence the following is a natural first order definition: Definition The mantle M is the intersection of all grounds, i.e., M p W p. 18

68 In the beginning, it was unclear whether the mantle is a model of ZFC. Together with Hamkins and Reitz, we isolated the following crucial hypothesis: Strong Downward Directedness of Grounds hypothesis Any set-sized collection of grounds has a common ground. That is, if a is a set, then there is a q such that W q ppa W p. 19

69 In the beginning, it was unclear whether the mantle is a model of ZFC. Together with Hamkins and Reitz, we isolated the following crucial hypothesis: Strong Downward Directedness of Grounds hypothesis Any set-sized collection of grounds has a common ground. That is, if a is a set, then there is a q such that W q ppa W p. This was a crucial hypothesis, which would answer many initially open questions. 19

70 Fact (F.-Hamkins-Reitz) Assume the strong DDG. Then: 1. The mantle is constant across the grounds (directedness suffices). 2. The mantle is a model of ZF (directedness suffices). 3. The mantle satisfies ZFC (directedness + local set-directedness suffice). 20

71 Fact (F.-Hamkins-Reitz) Assume the strong DDG. Then: 1. The mantle is constant across the grounds (directedness suffices). 2. The mantle is a model of ZF (directedness suffices). 3. The mantle satisfies ZFC (directedness + local set-directedness suffice). Initially, we had defined the generic mantle to be the intersection of the mantles of all set-forcing extensions, and proved that it satisfies ZF, and is the largest forcing-invariant class. 20

72 Fact (F.-Hamkins-Reitz) Assume the strong DDG. Then: 1. The mantle is constant across the grounds (directedness suffices). 2. The mantle is a model of ZF (directedness suffices). 3. The mantle satisfies ZFC (directedness + local set-directedness suffice). Initially, we had defined the generic mantle to be the intersection of the mantles of all set-forcing extensions, and proved that it satisfies ZF, and is the largest forcing-invariant class. But note that if the strong DDG holds in every set-forcing extension (this is what we called the generic strong DDG), then by 1. above, the generic mantle is equal to the mantle, and the mantle is a forcing-invariant inner model of ZFC. 20

73 Usuba s contributions Using a forgotten theorem of Bukovsky ingeniously, Usuba proved: Theorem (Usuba) The strong DDG holds. 21

74 Usuba s contributions Using a forgotten theorem of Bukovsky ingeniously, Usuba proved: Theorem (Usuba) The strong DDG holds. Thus, the generic strong DDG also holds, and we have: Theorem The mantle is a model of ZFC, and it is invariant under set-forcing. 21

75 Usuba s contributions Using a forgotten theorem of Bukovsky ingeniously, Usuba proved: Theorem (Usuba) The strong DDG holds. Thus, the generic strong DDG also holds, and we have: Theorem The mantle is a model of ZFC, and it is invariant under set-forcing. Also, the mantle is equal to the generic mantle, and is the intersection of the generic multiverse, so it is the largest forcing-invariant class. 21

76 Usuba s contributions, cont d Initially, Usuba proved the following theorem under a stronger large cardinal assumption, but he was able to lower it recently: Theorem (Usuba) If there is an extendible cardinal, then there are only set-many grounds. 22

77 Usuba s contributions, cont d Initially, Usuba proved the following theorem under a stronger large cardinal assumption, but he was able to lower it recently: Theorem (Usuba) If there is an extendible cardinal, then there are only set-many grounds. Observation If there are only set-many grounds, then the mantle is a ground. 22

78 Usuba s contributions, cont d Initially, Usuba proved the following theorem under a stronger large cardinal assumption, but he was able to lower it recently: Theorem (Usuba) If there is an extendible cardinal, then there are only set-many grounds. Observation If there are only set-many grounds, then the mantle is a ground. This is because by the strong DDG, there is a ground that is contained in every ground, in this situation, and this is the minimal ground, and the mantle. 22

79 Recap We have the following properties of the mantle: 1. pzfcq M, 23

80 Recap We have the following properties of the mantle: 1. pzfcq M, 2. M is a forcing-invariant inner model, 23

81 Recap We have the following properties of the mantle: 1. pzfcq M, 2. M is a forcing-invariant inner model, 3. if there is an extendible cardinal, then M is a ground. 23

82 Recap We have the following properties of the mantle: 1. pzfcq M, 2. M is a forcing-invariant inner model, 3. if there is an extendible cardinal, then M is a ground. In particular, M is very close to V: above the size of the forcing leading from M to V, V and M agree about cardinal arithmetic. 23

83 Recap We have the following properties of the mantle: 1. pzfcq M, 2. M is a forcing-invariant inner model, 3. if there is an extendible cardinal, then M is a ground. In particular, M is very close to V: above the size of the forcing leading from M to V, V and M agree about cardinal arithmetic. Of course, M has no canonical structure, in general. But just having the properties listed above is very useful. 23

84 Recap We have the following properties of the mantle: 1. pzfcq M, 2. M is a forcing-invariant inner model, 3. if there is an extendible cardinal, then M is a ground. In particular, M is very close to V: above the size of the forcing leading from M to V, V and M agree about cardinal arithmetic. Of course, M has no canonical structure, in general. But just having the properties listed above is very useful. I ll give an application in the following - it seems like there should be many similar uses of these techniques. 23

85 The goal theorem I want to show: Theorem (F.) Assume MP SC and there are only set-many grounds. Then there is a well-order of Ppω 1 q, definable without parameters, of order type ω 2. 24

86 The goal theorem I want to show: Theorem (F.) Assume MP SC and there are only set-many grounds. Then there is a well-order of Ppω 1 q, definable without parameters, of order type ω 2. Before giving a proof of this theorem, I need a crucial coding device, and I should remark: Observation If ZFC MP SC ph ω2 q is consistent, then so is the theory ZFC MP SC ph ω2 q there are only set-many grounds. 24

87 The extended Namba problem The following remarkable theorem of Jensen serves as a coding tool in my application. 25

88 The extended Namba problem The following remarkable theorem of Jensen serves as a coding tool in my application. Theorem (Jensen) Let κ be an inaccessible cardinal, and assume that GCH holds below κ. Let A κ be a set of regular cardinals. Then there is a subcomplete, κ-c.c. forcing P of size κ such that if G is P-generic, then κ ω VrGs 2 and for every regular τ P pω 1, κq, cf VrGs pτq # ω 1 ω if τ P A otherwise. 25

89 The proof Now all the tools needed for the argument are assembled, and the proof is actually quite easy. 26

90 The proof Now all the tools needed for the argument are assembled, and the proof is actually quite easy. Recall the theorem we want to prove: Theorem (F.) Assume MP SC and there are only set-many grounds. Then there is a well-order of Ppω 1 q, definable without parameters, of order type ω 2. 26

91 The proof Now all the tools needed for the argument are assembled, and the proof is actually quite easy. Recall the theorem we want to prove: Theorem (F.) Assume MP SC and there are only set-many grounds. Then there is a well-order of Ppω 1 q, definable without parameters, of order type ω 2. Proof. First, recall that since there are only set-many grounds, M is a ground. 26

92 The proof Now all the tools needed for the argument are assembled, and the proof is actually quite easy. Recall the theorem we want to prove: Theorem (F.) Assume MP SC and there are only set-many grounds. Then there is a well-order of Ppω 1 q, definable without parameters, of order type ω 2. Proof. First, recall that since there are only set-many grounds, M is a ground. Let V Mrgs g being P-generic over M, for some poset P P M. 26

93 The proof Let δ ω 2. 27

94 The proof Let δ ω 2. Since M is forcing-absolute, it follows from MP SC ph ω2 q that V M δ M by verifying the Tarski-Vaught criterion for elementary substructures. 27

95 The proof Let δ ω 2. Since M is forcing-absolute, it follows from MP SC ph ω2 q that V M δ M by verifying the Tarski-Vaught criterion for elementary substructures. Since δ is regular, it follows that δ is inaccessible in M. 27

96 The proof Let δ ω 2. Since M is forcing-absolute, it follows from MP SC ph ω2 q that V M δ M by verifying the Tarski-Vaught criterion for elementary substructures. Since δ is regular, it follows that δ is inaccessible in M. This, in turn, implies that there is a proper class of inaccessible cardinals in M, and since V is a set forcing extension of M, also in V. 27

97 The proof Let δ ω 2. Since M is forcing-absolute, it follows from MP SC ph ω2 q that V M δ M by verifying the Tarski-Vaught criterion for elementary substructures. Since δ is regular, it follows that δ is inaccessible in M. This, in turn, implies that there is a proper class of inaccessible cardinals in M, and since V is a set forcing extension of M, also in V. Let α P, and let xκ i i ω 1 y enumerate the next ω 1 1 inaccessible cardinals above α. We can perform an Easton iteration of at least countably closed forcing notions in order to reach an extension in which GCH holds below κ ω1, such that if h is generic, then each κ i is still inaccessible in Vrhs. 27

98 The proof Now, given a subset A of ω 1 in V, let à tκ i i P Au. 28

99 The proof Now, given a subset A of ω 1 in V, let à tκ i i P Au. Then let Q N Vrhs Ã,κ ω1 be the extended Namba forcing for à in Vrhs. Thus, Q is κ ω1 -c.c., and if G is Q-generic over Vrhs, then for every µ P Ã, VrhsrGs thinks that cfpµq ω 1, and for every ν P κ ω1 zã which is regular in Vrhs, VrhsrGs thinks that cfpνq ω. In particular, the latter is true for every κ j with j ω 1, j R A, since κ j remains regular in Vrhs. κ ω1 becomes ω 2 in VrhsrGs. 28

100 The proof Now, given a subset A of ω 1 in V, let à tκ i i P Au. Then let Q N Vrhs Ã,κ ω1 be the extended Namba forcing for à in Vrhs. Thus, Q is κ ω1 -c.c., and if G is Q-generic over Vrhs, then for every µ P Ã, VrhsrGs thinks that cfpµq ω 1, and for every ν P κ ω1 zã which is regular in Vrhs, VrhsrGs thinks that cfpνq ω. In particular, the latter is true for every κ j with j ω 1, j R A, since κ j remains regular in Vrhs. κ ω1 becomes ω 2 in VrhsrGs. So in VrhsrGs, we have: cfpκ i q # ω 1 ω if i P A if i P ω 1 za 28

101 The proof Now if VrhsrGsrIs is a further subcomplete forcing extension of VrhsrGs, then the cofinality of κ i, for i ω 1, cannot change, since subcomplete forcing preserves ω 1. 29

102 The proof Now if VrhsrGsrIs is a further subcomplete forcing extension of VrhsrGs, then the cofinality of κ i, for i ω 1, cannot change, since subcomplete forcing preserves ω 1. Moreover, the sequence xκ i i ω 1 y is definable in VrhsrGsrIs from the parameter α, as the enumeration of the next ω 1 many inaccessible cardinals in M beyond α. 29

103 The proof Now if VrhsrGsrIs is a further subcomplete forcing extension of VrhsrGs, then the cofinality of κ i, for i ω 1, cannot change, since subcomplete forcing preserves ω 1. Moreover, the sequence xκ i i ω 1 y is definable in VrhsrGsrIs from the parameter α, as the enumeration of the next ω 1 many inaccessible cardinals in M beyond α. Hence, A is definable in VrhsrGsrIs as the set of i cfpκ i q ω 1. ω 1 such that 29

104 The proof Now if VrhsrGsrIs is a further subcomplete forcing extension of VrhsrGs, then the cofinality of κ i, for i ω 1, cannot change, since subcomplete forcing preserves ω 1. Moreover, the sequence xκ i i ω 1 y is definable in VrhsrGsrIs from the parameter α, as the enumeration of the next ω 1 many inaccessible cardinals in M beyond α. Hence, A is definable in VrhsrGsrIs as the set of i cfpκ i q ω 1. ω 1 such that Let ψpa, αq be the statement expressing that if xλ i i ω 1 y enumerates the next ω 1 many inaccessible cardinals of M beyond α, then for all i ω 1, i P A iff cfpλ i q ω 1, and i R A iff cfpλ i q ω. Then the statement ϕpaq, expressing that there is an α with ψpa, αq, holds in VrhsrGsrIs. 29

105 The proof Since I was generic for an arbitrary subcomplete forcing notion in VrhsrGs, this means that ϕpaq is necessary with respect to subcomplete forcing extensions in VrhsrGs, and hence it is forceably necessary with respect to subcomplete forcing in V. 30

106 The proof Since I was generic for an arbitrary subcomplete forcing notion in VrhsrGs, this means that ϕpaq is necessary with respect to subcomplete forcing extensions in VrhsrGs, and hence it is forceably necessary with respect to subcomplete forcing in V. It follows by MP SC ph ω2 q that ϕpaq already holds in V. 30

107 The proof Since I was generic for an arbitrary subcomplete forcing notion in VrhsrGs, this means that ϕpaq is necessary with respect to subcomplete forcing extensions in VrhsrGs, and hence it is forceably necessary with respect to subcomplete forcing in V. It follows by MP SC ph ω2 q that ϕpaq already holds in V. Since A was an arbitrary subset of ω 1 in V, it follows that ϕpbq holds, for every B ω 1. 30

108 The proof Since I was generic for an arbitrary subcomplete forcing notion in VrhsrGs, this means that ϕpaq is necessary with respect to subcomplete forcing extensions in VrhsrGs, and hence it is forceably necessary with respect to subcomplete forcing in V. It follows by MP SC ph ω2 q that ϕpaq already holds in V. Since A was an arbitrary subset of ω 1 in V, it follows that ϕpbq holds, for every B ω 1. Given A ω 1, let β be such that ψpa, βq holds. In V Colpω1,βq, and in any further subcomplete forcing extension, ψpa, βq continues to hold, and β ω 2 there. Thus, if we let α A be least such that ψpa, α A q holds, it is subcomplete forceably necessary that α A ω 2, and so, it is already less than ω 2 in V. 30

109 The proof Since I was generic for an arbitrary subcomplete forcing notion in VrhsrGs, this means that ϕpaq is necessary with respect to subcomplete forcing extensions in VrhsrGs, and hence it is forceably necessary with respect to subcomplete forcing in V. It follows by MP SC ph ω2 q that ϕpaq already holds in V. Since A was an arbitrary subset of ω 1 in V, it follows that ϕpbq holds, for every B ω 1. Given A ω 1, let β be such that ψpa, βq holds. In V Colpω1,βq, and in any further subcomplete forcing extension, ψpa, βq continues to hold, and β ω 2 there. Thus, if we let α A be least such that ψpa, α A q holds, it is subcomplete forceably necessary that α A ω 2, and so, it is already less than ω 2 in V. This shows that if we define A B iff α A α B, for A, B ω 1, then this is a well-ordering of Ppω 1 q of order type ω 2, as wished. 30

110 Variations So the key was that the forcing-invariance of M, together with its closeness to V, enabled us to define the coding points in a generically absolute way. 31

111 Variations So the key was that the forcing-invariance of M, together with its closeness to V, enabled us to define the coding points in a generically absolute way. Other inner models can serve this purpose, under anti-large cardinal assumptions which ensure that they are close to V. 31

112 Variations So the key was that the forcing-invariance of M, together with its closeness to V, enabled us to define the coding points in a generically absolute way. Other inner models can serve this purpose, under anti-large cardinal assumptions which ensure that they are close to V. Theorem (F.) Assume MP SC ph ω2 q, and assume there is no inner model with an inaccessible limit of measurable cardinals. Then there is a well-order of Ppω 1 q of order type ω 2, definable from a subset of ω 1. 31

113 Variations So the key was that the forcing-invariance of M, together with its closeness to V, enabled us to define the coding points in a generically absolute way. Other inner models can serve this purpose, under anti-large cardinal assumptions which ensure that they are close to V. Theorem (F.) Assume MP SC ph ω2 q, and assume there is no inner model with an inaccessible limit of measurable cardinals. Then there is a well-order of Ppω 1 q of order type ω 2, definable from a subset of ω 1. Theorem (F.) Suppose 0 # does not exist, and that BSCFA holds. Then there is a well-order of Ppω 1 q of order type ω 2, definable without parameters. This well-order is in LpPpω 1 qq, is 1 -definable from a subset I 1 of ω 1 there, and it is xhω 2,Py 1 -definable in I 1. 31

114 Thus, we get definable well-orders from subcomplete forcing principles both from large cardinal assumptions and from anti-large cardinal assumptions. 32

115 The role of CH

116 Back to forcing axioms What makes SCFA so interesting is that it is consistent with CH, and even with. 33

117 Back to forcing axioms What makes SCFA so interesting is that it is consistent with CH, and even with. Even more interestingly, SCFA has many of the same consequences as MM. 33

118 Back to forcing axioms What makes SCFA so interesting is that it is consistent with CH, and even with. Even more interestingly, SCFA has many of the same consequences as MM. I will exemplify this by considering the effects of these forcing axioms on (the failure of) square principles. 33

119 Squares Definition (Jensen) Let κ be a cardinal. l κ says that there is a l κ -sequence, that is, a sequence xc α κ α κ, α limity such that each C α is club in α, otppc α q κ and for each β that is a limit point of C α, C β C α X β. 34

120 Squares Definition (Jensen) Let κ be a cardinal. l κ says that there is a l κ -sequence, that is, a sequence xc α κ α κ, α limity such that each C α is club in α, otppc α q κ and for each β that is a limit point of C α, C β C α X β. If λ is also a cardinal, then l κ,λ is the assertion that there is a l κ,λ -sequence, i.e., a sequence xc α κ α κ, α limity such that each C α has size at most λ, and each C P C α is club in α, has order-type at most κ, and satisfies the coherency condition that if β is a limit point of C, then C X β P C β. 34

121 Squares Definition (Jensen) Let κ be a cardinal. l κ says that there is a l κ -sequence, that is, a sequence xc α κ α κ, α limity such that each C α is club in α, otppc α q κ and for each β that is a limit point of C α, C β C α X β. If λ is also a cardinal, then l κ,λ is the assertion that there is a l κ,λ -sequence, i.e., a sequence xc α κ α κ, α limity such that each C α has size at most λ, and each C P C α is club in α, has order-type at most κ, and satisfies the coherency condition that if β is a limit point of C, then C X β P C β. l κ,κ is known as weak square, denoted by l κ. 34

122 Squares Definition (Jensen) Let κ be a cardinal. l κ says that there is a l κ -sequence, that is, a sequence xc α κ α κ, α limity such that each C α is club in α, otppc α q κ and for each β that is a limit point of C α, C β C α X β. If λ is also a cardinal, then l κ,λ is the assertion that there is a l κ,λ -sequence, i.e., a sequence xc α κ α κ, α limity such that each C α has size at most λ, and each C P C α is club in α, has order-type at most κ, and satisfies the coherency condition that if β is a limit point of C, then C X β P C β. l κ,κ is known as weak square, denoted by l κ. l κ,κ holds trivially, so l κ is the weakest nontrivial principle here. 34

123 Other kinds of squares Definition (Todorcevic, Jensen (?)) Let λ be a limit of limit ordinals. A sequence C xc α α λ, α limity is coherent if for every limit α λ, C α H and for every C P C α, C is club in α, and for every limit point β of C, C X β P C β. 35

124 Other kinds of squares Definition (Todorcevic, Jensen (?)) Let λ be a limit of limit ordinals. A sequence C xc α α λ, α limity is coherent if for every limit α λ, C α H and for every C P C α, C is club in α, and for every limit point β of C, C X β P C β. A thread through C is a set T such that C " ttu is coherent. 35

125 Other kinds of squares Definition (Todorcevic, Jensen (?)) Let λ be a limit of limit ordinals. A sequence C xc α α λ, α limity is coherent if for every limit α λ, C α H and for every C P C α, C is club in α, and for every limit point β of C, C X β P C β. A thread through C is a set T such that C " ttu is coherent. A coherent sequence is maximal if it has no thread. 35

126 Other kinds of squares Definition (Todorcevic, Jensen (?)) Let λ be a limit of limit ordinals. A sequence C xc α α λ, α limity is coherent if for every limit α λ, C α H and for every C P C α, C is club in α, and for every limit point β of C, C X β P C β. A thread through C is a set T such that C " ttu is coherent. A coherent sequence is maximal if it has no thread. If κ is a cardinal, then the principle lpλ, κq says that there is a maximal coherent sequence of length λ all of whose elements have size less than κ, and such a sequence is called a lpλ, κq-sequence. 35

127 Other kinds of squares Definition (Todorcevic, Jensen (?)) Let λ be a limit of limit ordinals. A sequence C xc α α λ, α limity is coherent if for every limit α λ, C α H and for every C P C α, C is club in α, and for every limit point β of C, C X β P C β. A thread through C is a set T such that C " ttu is coherent. A coherent sequence is maximal if it has no thread. If κ is a cardinal, then the principle lpλ, κq says that there is a maximal coherent sequence of length λ all of whose elements have size less than κ, and such a sequence is called a lpλ, κq-sequence. The principle lpλ, κq says that there is a maximal coherent sequence of length λ all of whose elements have size at most κ, and such a sequence is called a lpλ, κq-sequence. 35

128 Other kinds of squares Definition (Todorcevic, Jensen (?)) Let λ be a limit of limit ordinals. A sequence C xc α α λ, α limity is coherent if for every limit α λ, C α H and for every C P C α, C is club in α, and for every limit point β of C, C X β P C β. A thread through C is a set T such that C " ttu is coherent. A coherent sequence is maximal if it has no thread. If κ is a cardinal, then the principle lpλ, κq says that there is a maximal coherent sequence of length λ all of whose elements have size less than κ, and such a sequence is called a lpλ, κq-sequence. The principle lpλ, κq says that there is a maximal coherent sequence of length λ all of whose elements have size at most κ, and such a sequence is called a lpλ, κq-sequence. The principle lpλ, 1q is denoted lpλq. 35

129 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails 36

130 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold 36

131 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails 36

132 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold 36

133 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold cfpλq ω l λ fails fails 36

134 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold cfpλq ω l λ fails fails λ ω 1 l ω 1 fails may hold 36

135 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold cfpλq ω l λ fails fails λ ω 1 l ω 1 fails may hold cfpλq ω 3, µ ω 1 lpλ, ω 1 q fails fails 36

136 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold cfpλq ω l λ fails fails λ ω 1 l ω 1 fails may hold cfpλq ω 3, µ ω 1 lpλ, ω 1 q fails fails cfpλq ω 3, µ ω 2 lpλ, ω 2 q may hold may hold 36

137 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold cfpλq ω l λ fails fails λ ω 1 l ω 1 fails may hold cfpλq ω 3, µ ω 1 lpλ, ω 1 q fails fails cfpλq ω 3, µ ω 2 lpλ, ω 2 q may hold may hold cfpλq ω 2, µ ω lpλ, ωq fails fails 36

138 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold cfpλq ω l λ fails fails λ ω 1 l ω 1 fails may hold cfpλq ω 3, µ ω 1 lpλ, ω 1 q fails fails cfpλq ω 3, µ ω 2 lpλ, ω 2 q may hold may hold cfpλq ω 2, µ ω lpλ, ωq fails fails cfpλq ω 2, µ ω 1 lpλ, ω 1 q fails may hold 36

139 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold cfpλq ω l λ fails fails λ ω 1 l ω 1 fails may hold cfpλq ω 3, µ ω 1 lpλ, ω 1 q fails fails cfpλq ω 3, µ ω 2 lpλ, ω 2 q may hold may hold cfpλq ω 2, µ ω lpλ, ωq fails fails cfpλq ω 2, µ ω 1 lpλ, ω 1 q fails may hold Thus, the only instances of square principles that are consistent with SCFA but not with MM are actually consequences of CH. 36

140 Consequences of MM and SCFA Conditions on µ, λ Principle Under MM Under SCFA cfpλq ω 2, µ cfpλq l λ,µ fails fails cfpλq ω 2, µ cfpλq l λ,µ may hold may hold cfpλq ω 1, µ λ l λ,µ fails fails cfpλq ω 1 λ µ l λ,µ may hold may hold cfpλq ω l λ fails fails λ ω 1 l ω 1 fails may hold cfpλq ω 3, µ ω 1 lpλ, ω 1 q fails fails cfpλq ω 3, µ ω 2 lpλ, ω 2 q may hold may hold cfpλq ω 2, µ ω lpλ, ωq fails fails cfpλq ω 2, µ ω 1 lpλ, ω 1 q fails may hold Thus, the only instances of square principles that are consistent with SCFA but not with MM are actually consequences of CH. It is then a natural question to ask whether their negations follow from SCFA CH. 36

141 SCFA CH More generally, one is led to compare SCFA CH with MM. 37

142 SCFA CH More generally, one is led to compare SCFA CH with MM. It may seem somewhat counter-intuitive that one might be interested in models of SCFA CH, since one of the main appeals of SCFA is that it is consistent with CH. 37

143 SCFA CH More generally, one is led to compare SCFA CH with MM. It may seem somewhat counter-intuitive that one might be interested in models of SCFA CH, since one of the main appeals of SCFA is that it is consistent with CH. In fact, recall that MP SC implies CH, and even. The same is true of higher resurrection axioms for subcomplete forcing. Thus, the status of l ω 1 is clear under these principles. 37