Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
The direct product problem Vipul Naik The direct product cancellation problem The direct product problem The free product cancellation problem Conclusions Vipul Naik October 17, 2006 Outline The direct product problem Vipul Naik The direct product cancellation problem The direct product cancellation problem The free product cancellation problem Conclusions The free product cancellation problem Conclusions The problem statement The direct product problem Vipul Naik I Given: Groups G , H and K such that: G ×H ∼ =G ×K The direct product cancellation problem The free product cancellation problem Conclusions The direct product problem The problem statement Vipul Naik I Given: Groups G , H and K such that: G ×H ∼ =G ×K The direct product cancellation problem The free product cancellation problem Conclusions I To Prove: H∼ =K The direct product problem The problem statement Vipul Naik I Given: Groups G , H and K such that: G ×H ∼ =G ×K The direct product cancellation problem The free product cancellation problem Conclusions I To Prove: H∼ =K I Question: What hypotheses on G , H and K can take us from Given to To Prove? The direct product problem The problem statement Vipul Naik I Given: Groups G , H and K such that: G ×H ∼ =G ×K The direct product cancellation problem The free product cancellation problem Conclusions I To Prove: H∼ =K I Question: What hypotheses on G , H and K can take us from Given to To Prove?That is, when can we cancel G ? The direct product problem Vipul Naik The direct product cancellation problem The question really has two aspects: I What conditions on G ensure that G can be cancelled? The free product cancellation problem Conclusions The direct product problem Vipul Naik The direct product cancellation problem The question really has two aspects: I What conditions on G ensure that G can be cancelled? I What conditions on H (and/or K ) ensure that G can be cancelled? The free product cancellation problem Conclusions Monoidal interpretation The direct product problem Vipul Naik Consider a commutative and associative monoid whose: I Elements are isomorphism classes of groups I Multiplication map is the direct product operation. The direct product cancellation problem The free product cancellation problem Conclusions Monoidal interpretation The direct product problem Vipul Naik Consider a commutative and associative monoid whose: I Elements are isomorphism classes of groups I Multiplication map is the direct product operation. The big questions: I Which elements of the monoid are cancellative? The direct product cancellation problem The free product cancellation problem Conclusions Monoidal interpretation The direct product problem Vipul Naik Consider a commutative and associative monoid whose: I Elements are isomorphism classes of groups I Multiplication map is the direct product operation. The big questions: I Which elements of the monoid are cancellative? I Which elements of the monoid are cancelductive? That is, for which elements H is it true that G can always be cancelled from an expression of the form G ×H ∼ = G × K? The direct product cancellation problem The free product cancellation problem Conclusions Where the cancellation fails The direct product problem Vipul Naik The direct product cancellation problem Let H be any nontrivial group. Take the countable direct product of H with itself. Call that G . The free product cancellation problem Conclusions Where the cancellation fails The direct product problem Vipul Naik The direct product cancellation problem Let H be any nontrivial group. Take the countable direct product of H with itself. Call that G . Then: G ×H ∼ = G × {e} The free product cancellation problem Conclusions Where the cancellation fails The direct product problem Vipul Naik The direct product cancellation problem Let H be any nontrivial group. Take the countable direct product of H with itself. Call that G . Then: G ×H ∼ = G × {e} Even though H is nontrivial. The free product cancellation problem Conclusions Positive answer for finite groups The direct product problem Vipul Naik The direct product cancellation problem We’re going to prove that the collection of isomorphism classes of finite groups is cancellative. The free product cancellation problem Conclusions Positive answer for finite groups The direct product problem Vipul Naik The direct product cancellation problem We’re going to prove that the collection of isomorphism classes of finite groups is cancellative. In other words, if G , H and K are finite groups, we’ll prove that G ×H ∼ =K = G × K =⇒ H ∼ The free product cancellation problem Conclusions Category theoretic definition of direct product The direct product problem Vipul Naik The direct product cancellation problem Given any group L, the homomorphisms from L to G × H are in bijection with pairs of homomorphisms from L to G and L to H. The free product cancellation problem Conclusions Category theoretic definition of direct product The direct product problem Vipul Naik The direct product cancellation problem Given any group L, the homomorphisms from L to G × H are in bijection with pairs of homomorphisms from L to G and L to H. This follows from the category theoretic interpretation of direct product as a universal property with respect to maps to both the groups. The free product cancellation problem Conclusions The direct product problem A little working out Vipul Naik Let h(L, G ) denote the cardinality of Hom(L, G ) where L and G are finite groups. Then we have: G ×H ∼ = G ×K The direct product cancellation problem The free product cancellation problem Conclusions A little working out The direct product problem Vipul Naik Let h(L, G ) denote the cardinality of Hom(L, G ) where L and G are finite groups. Then we have: G ×H ∼ = G ×K =⇒ h(L, G × H) = h(L, G × K ) The direct product cancellation problem The free product cancellation problem Conclusions A little working out The direct product problem Vipul Naik Let h(L, G ) denote the cardinality of Hom(L, G ) where L and G are finite groups. Then we have: G ×H ∼ = G ×K =⇒ h(L, G × H) = h(L, G × K ) =⇒ h(L, G )h(L, H) = h(L, G )h(L, K ) The direct product cancellation problem The free product cancellation problem Conclusions A little working out The direct product problem Vipul Naik Let h(L, G ) denote the cardinality of Hom(L, G ) where L and G are finite groups. Then we have: G ×H ∼ = G ×K =⇒ h(L, G × H) = h(L, G × K ) =⇒ h(L, G )h(L, H) = h(L, G )h(L, K ) =⇒ h(L, G )(h(L, H) − h(L, K )) = 0 The direct product cancellation problem The free product cancellation problem Conclusions The direct product problem A little working out Vipul Naik Let h(L, G ) denote the cardinality of Hom(L, G ) where L and G are finite groups. Then we have: G ×H ∼ = G ×K =⇒ h(L, G × H) = h(L, G × K ) =⇒ h(L, G )h(L, H) = h(L, G )h(L, K ) =⇒ h(L, G )(h(L, H) − h(L, K )) = 0 =⇒ h(L, H) − h(L, K ) = 0 because h(L, G ) is finite nonzero =⇒ h(L, H) = h(L, K ) The direct product cancellation problem The free product cancellation problem Conclusions So what? The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem We showed that if G , H and K are finite groups and G ×H ∼ = G × K , then for any finite group L, h(L, H) = h(L, K ) Conclusions So what? The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem We showed that if G , H and K are finite groups and G ×H ∼ = G × K , then for any finite group L, h(L, H) = h(L, K ) Define the to-Hom statistics of a group H as the map L 7→ h(L, H). Then from the above discussion, H and K have the same to-Hom statistics. Conclusions to-Hom and injective to-Hom The direct product problem Vipul Naik The direct product cancellation problem Let i(L, H) denote the number of injective homomorphisms from L to H. The free product cancellation problem Conclusions The direct product problem to-Hom and injective to-Hom Vipul Naik The direct product cancellation problem Let i(L, H) denote the number of injective homomorphisms from L to H. Then we have: h(L, G ) = X NEL i(L/N, G ) (1) The free product cancellation problem Conclusions The direct product problem to-Hom and injective to-Hom Vipul Naik The direct product cancellation problem Let i(L, H) denote the number of injective homomorphisms from L to H. Then we have: h(L, G ) = X i(L/N, G ) NEL Manipulating this equation a bit, we get: H and K have the same to-Hom statistics =⇒ i(L, H) = i(L, K ) for all finite groups L. (1) The free product cancellation problem Conclusions The direct product problem to-Hom and injective to-Hom Vipul Naik The direct product cancellation problem Let i(L, H) denote the number of injective homomorphisms from L to H. Then we have: h(L, G ) = X i(L/N, G ) NEL Manipulating this equation a bit, we get: H and K have the same to-Hom statistics =⇒ i(L, H) = i(L, K ) for all finite groups L. And then, putting H = L, we are done! (1) The free product cancellation problem Conclusions to-Hom finite groups The direct product problem Vipul Naik The direct product cancellation problem Finiteness of G was not important to the above proof. All we needed was that for every finite group L, h(L, G ) is finite. The free product cancellation problem Conclusions to-Hom finite groups The direct product problem Vipul Naik The direct product cancellation problem Finiteness of G was not important to the above proof. All we needed was that for every finite group L, h(L, G ) is finite. A group satisfying that property is termed to-Hom finite. Examples of to-Hom finite groups: torsion-free groups, quasicyclic groups. The free product cancellation problem Conclusions to-Hom finite groups The direct product problem Vipul Naik The direct product cancellation problem Finiteness of G was not important to the above proof. All we needed was that for every finite group L, h(L, G ) is finite. A group satisfying that property is termed to-Hom finite. Examples of to-Hom finite groups: torsion-free groups, quasicyclic groups. Reformulation: if G is to-Hom finite and H and K are finite, then G × H ∼ = G × K =⇒ H ∼ =K The free product cancellation problem Conclusions In greater generality We used three facts to prove: G ×H ∼ = G × K =⇒ H ∼ =K 1. For every L, we have: h(L, G × H) = h(L, G ) × h(L, H) The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem Conclusions The direct product problem In greater generality We used three facts to prove: Vipul Naik G ×H ∼ = G × K =⇒ H ∼ =K The direct product cancellation problem The free product cancellation problem 1. For every L, we have: h(L, G × H) = h(L, G ) × h(L, H) 2. There is a relationship between h and i, namely: X h(L, H) = i(L/N, H) NEL Conclusions The direct product problem In greater generality We used three facts to prove: Vipul Naik G ×H ∼ = G × K =⇒ H ∼ =K The direct product cancellation problem The free product cancellation problem 1. For every L, we have: h(L, G × H) = h(L, G ) × h(L, H) 2. There is a relationship between h and i, namely: X h(L, H) = i(L/N, H) NEL 3. If there is an injective homomorphism from H to K and there is an injective homomorphism from K to H, then H is isomorphic to K . Question: which of these is true for structures other than groups? Conclusions Variety of algebras The direct product problem Vipul Naik I An algebra is a set along with a collection of n-ary operations on the set) possibly with different n. The direct product cancellation problem The free product cancellation problem Conclusions Variety of algebras The direct product problem Vipul Naik I I An algebra is a set along with a collection of n-ary operations on the set) possibly with different n. A variety of algebras is described by a collection of operator symbols, each with an arity, and a collection of equations between expressions involving those operator symbols. An algebra belongs to the variety if it has operations for those operator symbols and the operations satisfy the equations universally. The direct product cancellation problem The free product cancellation problem Conclusions Variety of algebras The direct product problem Vipul Naik I I I An algebra is a set along with a collection of n-ary operations on the set) possibly with different n. A variety of algebras is described by a collection of operator symbols, each with an arity, and a collection of equations between expressions involving those operator symbols. An algebra belongs to the variety if it has operations for those operator symbols and the operations satisfy the equations universally. For instance, groups form a variety of algebras. Monoids form a variety of algebra. Rings with identity form a variety of algebras. Fields do not form a vareity of algebras. The direct product cancellation problem The free product cancellation problem Conclusions In the generality of varieties The general version says: If G , H and K are finite algebras of a variety with 0, and G ×H ∼ = G × K as algebras of the variety, then H ∼ = K as algebras of the variety. The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem Conclusions In the generality of varieties The general version says: If G , H and K are finite algebras of a variety with 0, and G ×H ∼ = G × K as algebras of the variety, then H ∼ = K as algebras of the variety. 1. For every L, we have: The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem Conclusions h(L, G × H) = h(L, G ) × h(L, H) This still holds. In the generality of varieties The general version says: If G , H and K are finite algebras of a variety with 0, and G ×H ∼ = G × K as algebras of the variety, then H ∼ = K as algebras of the variety. 1. For every L, we have: The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem Conclusions h(L, G × H) = h(L, G ) × h(L, H) This still holds. 2. There is a relationship between h and i, namely: X h(L, H) = i(L/ρ, H) ρ a congruence on L In the generality of varieties The general version says: If G , H and K are finite algebras of a variety with 0, and G ×H ∼ = G × K as algebras of the variety, then H ∼ = K as algebras of the variety. 1. For every L, we have: The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem Conclusions h(L, G × H) = h(L, G ) × h(L, H) This still holds. 2. There is a relationship between h and i, namely: X h(L, H) = i(L/ρ, H) ρ a congruence on L 3. If there is an injective homomorphism from H to K and there is an injective homomorphism from K to H, then H is isomorphic to K . Outline The direct product problem Vipul Naik The direct product cancellation problem The direct product cancellation problem The free product cancellation problem Conclusions The free product cancellation problem Conclusions Meaning of free product The direct product problem Vipul Naik The direct product cancellation problem The free product of two groups G and H is denoted as G ∗ H and is defined as the set of formal strings whose letters are elements from G and elements from H modulo the product relations in G and in H. The free product cancellation problem Conclusions Meaning of free product The direct product problem Vipul Naik The direct product cancellation problem The free product of two groups G and H is denoted as G ∗ H and is defined as the set of formal strings whose letters are elements from G and elements from H modulo the product relations in G and in H. The collection of isomorphism classes of all groups forms a commutative monoid with respect to the free product operation. Finite groups do not form a submonoid: the free product of any two nontrivial groups is infinite. However, finitely generated groups do form a submonoid. The free product cancellation problem Conclusions The problem statement The direct product problem Vipul Naik I Given: Groups G , H and K such that: G ∗H ∼ =G ∗K The direct product cancellation problem The free product cancellation problem Conclusions The direct product problem The problem statement Vipul Naik I Given: Groups G , H and K such that: G ∗H ∼ =G ∗K The direct product cancellation problem The free product cancellation problem Conclusions I To Prove: H∼ =K The direct product problem The problem statement Vipul Naik I Given: Groups G , H and K such that: G ∗H ∼ =G ∗K The direct product cancellation problem The free product cancellation problem Conclusions I To Prove: H∼ =K I Question: What hypotheses on G , H and K can take us from Given to To Prove? The direct product problem The problem statement Vipul Naik I Given: Groups G , H and K such that: G ∗H ∼ =G ∗K The direct product cancellation problem The free product cancellation problem Conclusions I To Prove: H∼ =K I Question: What hypotheses on G , H and K can take us from Given to To Prove?That is, when can we cancel G ? Where the cancellation fails The direct product problem Vipul Naik The direct product cancellation problem Let H be any nontrivial group. Take the countable free product of H with itself. Call that G . The free product cancellation problem Conclusions Where the cancellation fails The direct product problem Vipul Naik The direct product cancellation problem Let H be any nontrivial group. Take the countable free product of H with itself. Call that G . Then: G ∗H ∼ = G ∗ {e} The free product cancellation problem Conclusions Where the cancellation fails The direct product problem Vipul Naik The direct product cancellation problem Let H be any nontrivial group. Take the countable free product of H with itself. Call that G . Then: G ∗H ∼ = G ∗ {e} Even though H is nontrivial. The free product cancellation problem Conclusions Products and coproducts The direct product problem Vipul Naik The direct product cancellation problem Given groups (or algebras of a variety) L, G , and H, there is a natural bijection between the homomorphisms from L to G × H and pairs of homomorphisms L to G , L to H. We can in fact define G × H from this. The free product cancellation problem Conclusions Products and coproducts The direct product problem Vipul Naik The direct product cancellation problem Given groups (or algebras of a variety) L, G , and H, there is a natural bijection between the homomorphisms from L to G × H and pairs of homomorphisms L to G , L to H. We can in fact define G × H from this. Is it true that h(G × H, L) = h(G , L)h(H, L)? No. But if we replace direct product by free product, we get: h(G ∗ H, L) = h(G , L)h(H, L) The free product can be defined this way. The free product cancellation problem Conclusions The direct product problem A little working out Vipul Naik Let h(G , L) denote the cardinality of Hom(G , L) where L and G are finite groups. Then we have: The direct product cancellation problem The free product cancellation problem G ∗H ∼ = G ∗K Conclusions A little working out The direct product problem Vipul Naik Let h(G , L) denote the cardinality of Hom(G , L) where L and G are finite groups. Then we have: The direct product cancellation problem The free product cancellation problem G ∗H ∼ = G ∗K =⇒ h(G ∗ H, L) = h(G ∗ K , L) Conclusions A little working out The direct product problem Vipul Naik Let h(G , L) denote the cardinality of Hom(G , L) where L and G are finite groups. Then we have: The direct product cancellation problem The free product cancellation problem G ∗H ∼ = G ∗K =⇒ h(G ∗ H, L) = h(G ∗ K , L) =⇒ h(G , L)h(H, L) = h(G , L)h(K , L) Conclusions A little working out The direct product problem Vipul Naik Let h(G , L) denote the cardinality of Hom(G , L) where L and G are finite groups. Then we have: The direct product cancellation problem The free product cancellation problem G ∗H ∼ = G ∗K =⇒ h(G ∗ H, L) = h(G ∗ K , L) =⇒ h(G , L)h(H, L) = h(G , L)h(K , L) =⇒ h(G , L)(h(H, L) − h(K , L)) = 0 Conclusions The direct product problem A little working out Vipul Naik Let h(G , L) denote the cardinality of Hom(G , L) where L and G are finite groups. Then we have: The direct product cancellation problem The free product cancellation problem G ∗H ∼ = G ∗K =⇒ h(G ∗ H, L) = h(G ∗ K , L) =⇒ h(G , L)h(H, L) = h(G , L)h(K , L) =⇒ h(G , L)(h(H, L) − h(K , L)) = 0 =⇒ h(H, L) − h(K , L) = 0 because h(G , L) is finite nonzero =⇒ h(H, L) = h(K , L) Conclusions So what? The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem We showed that if G , H and K are finite groups and G ∗H ∼ = G ∗ K , then for any finite group L, h(H, L) = h(K , L) Conclusions So what? The direct product problem Vipul Naik The direct product cancellation problem The free product cancellation problem We showed that if G , H and K are finite groups and G ∗H ∼ = G ∗ K , then for any finite group L, h(H, L) = h(K , L) Define the from-Hom statistics of a group H as the map L 7→ h(H, L). Then from the above discussion, H and K have the same from-Hom statistics. Conclusions to-Hom and injective to-Hom The direct product problem Vipul Naik The direct product cancellation problem Let s(H, L) denote the number of surjective homomorphisms from H to H. The free product cancellation problem Conclusions The direct product problem to-Hom and injective to-Hom Vipul Naik The direct product cancellation problem Let s(H, L) denote the number of surjective homomorphisms from H to H. Then we have: h(G , L) = X M≤L s(G , M) (2) The free product cancellation problem Conclusions The direct product problem to-Hom and injective to-Hom Vipul Naik The direct product cancellation problem Let s(H, L) denote the number of surjective homomorphisms from H to H. Then we have: h(G , L) = X s(G , M) M≤L Manipulating this equation a bit, we get: H and K have the same fom-Hom statistics =⇒ s(H, L) = s(K , L) for all finite groups L. (2) The free product cancellation problem Conclusions The direct product problem to-Hom and injective to-Hom Vipul Naik The direct product cancellation problem Let s(H, L) denote the number of surjective homomorphisms from H to H. Then we have: h(G , L) = X s(G , M) M≤L Manipulating this equation a bit, we get: H and K have the same fom-Hom statistics =⇒ s(H, L) = s(K , L) for all finite groups L. And then, putting H = L, we are done! (2) The free product cancellation problem Conclusions from-Hom finite groups The direct product problem Vipul Naik The direct product cancellation problem Finiteness of G was not important to the above proof. All we needed was that for every finite group L, h(G , L) is finite. The free product cancellation problem Conclusions from-Hom finite groups The direct product problem Vipul Naik The direct product cancellation problem Finiteness of G was not important to the above proof. All we needed was that for every finite group L, h(G , L) is finite. A group satisfying that property is termed from-Hom finite. Examples of from-Hom finite groups: finitely generated groups, simple groups, groups with only finitely many normal subgroups of finite index. Reformulation: if G is from-Hom finite and H and K are finite, then G ∗ H ∼ = G ∗ K =⇒ H ∼ =K The free product cancellation problem Conclusions This also generalizes The direct product problem Vipul Naik The direct product cancellation problem The notion of free product makes sense for any variety of algebras. And the earlier result generalizes to saying: If G , H and K are finite algebras of a variety then G ∗H ∼ =K = G ∗ K =⇒ H ∼ The free product cancellation problem Conclusions Outline The direct product problem Vipul Naik The direct product cancellation problem The direct product cancellation problem The free product cancellation problem Conclusions The free product cancellation problem Conclusions Crude summary The direct product problem Vipul Naik When does G × H ∼ = G × K imply that H ∼ = K? For groups (and also for varieties of algebras): I When all of G , H and K are finite. The direct product cancellation problem The free product cancellation problem Conclusions Crude summary The direct product problem Vipul Naik When does G × H ∼ = G × K imply that H ∼ = K? For groups (and also for varieties of algebras): I When all of G , H and K are finite. I when H and K are finite and G is to-Hom finite. For instance, G may be torsion-free, or its torsion part may be finite. The direct product cancellation problem The free product cancellation problem Conclusions Crude summary The direct product problem Vipul Naik When does G × H ∼ = G × K imply that H ∼ = K? For groups (and also for varieties of algebras): I When all of G , H and K are finite. I when H and K are finite and G is to-Hom finite. For instance, G may be torsion-free, or its torsion part may be finite. ∼ K? When does G ∗ H ∼ = G ∗ K imply that H = For groups (and also for varieties of algebras): I When all of G , H and K are finite. The direct product cancellation problem The free product cancellation problem Conclusions Crude summary The direct product problem Vipul Naik When does G × H ∼ = G × K imply that H ∼ = K? For groups (and also for varieties of algebras): I When all of G , H and K are finite. I when H and K are finite and G is to-Hom finite. For instance, G may be torsion-free, or its torsion part may be finite. ∼ K? When does G ∗ H ∼ = G ∗ K imply that H = For groups (and also for varieties of algebras): I When all of G , H and K are finite. I When H and K are finite and G is from-Hom finite. For instance, G may be finitely generated, simple, or have only finitely many finite index normal subgroups. The direct product cancellation problem The free product cancellation problem Conclusions Sophisticated summary The direct product problem Vipul Naik I The monoid of isomorphism classes of finite algebras of a variety (such as groups) with respect to the direct product operation, is cancellative. The direct product cancellation problem The free product cancellation problem Conclusions Sophisticated summary The direct product problem Vipul Naik I I The monoid of isomorphism classes of finite algebras of a variety (such as groups) with respect to the direct product operation, is cancellative. Even better, in the monoid of isomorphism classes of algebras of a variety, the to-Hom finite members can always be cancelled when multiplying with the finite members. The direct product cancellation problem The free product cancellation problem Conclusions Sophisticated summary The direct product problem Vipul Naik I The monoid of isomorphism classes of finite algebras of a variety (such as groups) with respect to the direct product operation, is cancellative. I Even better, in the monoid of isomorphism classes of algebras of a variety, the to-Hom finite members can always be cancelled when multiplying with the finite members. I In the monoid of isomorphism classes of algebras with the free product as multiplication, the finite elements cancel with the finite elements. The direct product cancellation problem The free product cancellation problem Conclusions Sophisticated summary The direct product problem Vipul Naik I The monoid of isomorphism classes of finite algebras of a variety (such as groups) with respect to the direct product operation, is cancellative. I Even better, in the monoid of isomorphism classes of algebras of a variety, the to-Hom finite members can always be cancelled when multiplying with the finite members. I In the monoid of isomorphism classes of algebras with the free product as multiplication, the finite elements cancel with the finite elements. I More generally, the from-Hom finite elements cancel when multiplying with the finite elements. The direct product cancellation problem The free product cancellation problem Conclusions