supernilpotence and higher dimensional congruences
play

Supernilpotence and Higher Dimensional Congruences Andrew Moorhead - PowerPoint PPT Presentation

Supernilpotence and Higher Dimensional Congruences Andrew Moorhead Vanderbilt University August 5, 2018 Overview of Talk 1. Commutator Theory, Nilpotence, and Supernilpotence Overview of Talk 1. Commutator Theory, Nilpotence, and


  1. Commutator Definition ◮ The modular commutator can be equivalently defined with either 1. the term condition, or 2. properties of a relation, usually called ∆. Definition (Term Condition) Let A be an algebra and take α, β, δ ∈ Con( A ). We say that α centralizes β modulo δ when the following condition is met: ◮ For all t ∈ Pol( A ) and a 0 ≡ α b 0 and a 1 ≡ β b 1 with | a 0 | + | a 1 | = σ ( t ) , � � t ( a 0 , a 1 ) ≡ δ t ( a 0 , b 1 ) = ⇒ t ( b 0 , a 0 ) ≡ δ t ( b 0 , b 1 )

  2. Commutator Definition ◮ The modular commutator can be equivalently defined with either 1. the term condition, or 2. properties of a relation, usually called ∆. Definition (Term Condition) Let A be an algebra and take α, β, δ ∈ Con( A ). We say that α centralizes β modulo δ when the following condition is met: ◮ For all t ∈ Pol( A ) and a 0 ≡ α b 0 and a 1 ≡ β b 1 with | a 0 | + | a 1 | = σ ( t ) , � � t ( a 0 , a 1 ) ≡ δ t ( a 0 , b 1 ) = ⇒ t ( b 0 , a 0 ) ≡ δ t ( b 0 , b 1 ) We write C TC ( α, β ; δ ) whenever this is true.

  3. Commutator Definition ◮ The modular commutator can be equivalently defined with either 1. the term condition, or 2. properties of a relation, usually called ∆. Definition (Term Condition) Let A be an algebra and take α, β, δ ∈ Con( A ). We say that α centralizes β modulo δ when the following condition is met: ◮ For all t ∈ Pol( A ) and a 0 ≡ α b 0 and a 1 ≡ β b 1 with | a 0 | + | a 1 | = σ ( t ) , � � t ( a 0 , a 1 ) ≡ δ t ( a 0 , b 1 ) = ⇒ t ( b 0 , a 0 ) ≡ δ t ( b 0 , b 1 ) We write C TC ( α, β ; δ ) whenever this is true. ◮ The term condition may be described as a condition that is quantified over a certain invariant relation of A which is called the algebra of ( α, β )-matrices and is denoted M ( α, β ).

  4. Matrices ◮ A square is the graph � 2 2 ; E � , where two functions f , g ∈ 2 2 are connected by an edge if and only if their outputs differ in exactly one argument.

  5. Matrices ◮ A square is the graph � 2 2 ; E � , where two functions f , g ∈ 2 2 are connected by an edge if and only if their outputs differ in exactly one argument. (0 , 1) (1 , 1) (0 , 0) (1 , 0) ◮ We say that a relation R on a set A is 2-dimensional if R ⊆ A 2 2 ( R is a set of squares whos vertices are labeled by elements of A .)

  6. Matrices ◮ A square is the graph � 2 2 ; E � , where two functions f , g ∈ 2 2 are connected by an edge if and only if their outputs differ in exactly one argument. (0 , 1) (1 , 1) (0 , 0) (1 , 0) ◮ We say that a relation R on a set A is 2-dimensional if R ⊆ A 2 2 ( R is a set of squares whos vertices are labeled by elements of A .) ◮ M ( α, β ) is the subalgebra of A 2 2 with generators �� x � � �� y � � � y y : x ≡ α y : x ≡ β y x y x x

  7. Matrices For δ ∈ Con( A ) we have that α centralizes β modulo δ if the implication c d c d → δ a b a b β α holds for all ( α, β )-matrices. This condition is abbreviated C TC ( α, β ; δ ).

  8. Matrices Similarly, we have that β centralizes α modulo δ if the implication δ c d c d → a b a b β α holds for all ( α, β )-matrices. This condition is abbreviated C TC ( β, α ; δ ).

  9. Matrices ◮ The binary commutator is defined to be � [ α, β ] TC = { δ : C ( α, β ; δ ) }

  10. Matrices ◮ The notions of matrices and centrality for three congruences are defined similarly.

  11. Matrices ◮ The notions of matrices and centrality for three congruences are defined similarly. ◮ A cube is the graph � 2 3 ; E � , where two functions f , g ∈ 2 3 are connected by an edge if and only if their outputs differ in exactly one argument.

  12. Matrices ◮ The notions of matrices and centrality for three congruences are defined similarly. ◮ A cube is the graph � 2 3 ; E � , where two functions f , g ∈ 2 3 are connected by an edge if and only if their outputs differ in exactly one argument. (0 , 1 , 0) (1 , 1 , 0) (0 , 1 , 1) (1 , 1 , 1) (0 , 0 , 0) (1 , 0 , 0) (0 , 0 , 1) (1 , 0 , 1) ◮ We say that a relation R on a set A is 3-dimensional if R ⊆ A 3 2 ( R is a set of cubes whos vertices are labeled by elements of A .)

  13. Matrices ◮ The notions of matrices and centrality for three congruences are defined similarly. ◮ A cube is the graph � 2 3 ; E � , where two functions f , g ∈ 2 3 are connected by an edge if and only if their outputs differ in exactly one argument. (0 , 1 , 0) (1 , 1 , 0) (0 , 1 , 1) (1 , 1 , 1) (0 , 0 , 0) (1 , 0 , 0) (0 , 0 , 1) (1 , 0 , 1) ◮ We say that a relation R on a set A is 3-dimensional if R ⊆ A 3 2 ( R is a set of cubes whos vertices are labeled by elements of A .)

  14. Matrices ◮ For congruences θ 0 , θ 1 , θ 2 ∈ Con( A ), set M ( θ 0 , θ 1 , θ 2 ) ≤ A 2 3 to be the subalgebra generated by the following labeled cubes: x y y y x x x y y y y y θ 1 θ 0 x y x x x x x y x x y y θ 2 M ( θ 0 , θ 1 , θ 2 ) is called the algebra of ( θ 0 , θ 1 , θ 2 )-matrices.

  15. Centrality ◮ For δ ∈ Con( A ), we say that θ 0 , θ 1 centralize θ 2 modulo δ if the following implication holds for all ( θ 0 , θ 1 , θ 2 )-matrices:

  16. Centrality ◮ For δ ∈ Con( A ), we say that θ 0 , θ 1 centralize θ 2 modulo δ if the following implication holds for all ( θ 0 , θ 1 , θ 2 )-matrices: c d θ 1 g h θ 0 a b θ 2 e f

  17. Centrality ◮ For δ ∈ Con( A ), we say that θ 0 , θ 1 centralize θ 2 modulo δ if the following implication holds for all ( θ 0 , θ 1 , θ 2 )-matrices: c d δ θ 1 g h θ 0 a b θ 2 e f

  18. Centrality ◮ For δ ∈ Con( A ), we say that θ 0 , θ 1 centralize θ 2 modulo δ if the following implication holds for all ( θ 0 , θ 1 , θ 2 )-matrices: c d δ θ 1 g h θ 0 a b θ 2 e f

  19. Centrality ◮ For δ ∈ Con( A ), we say that θ 0 , θ 1 centralize θ 2 modulo δ if the following implication holds for all ( θ 0 , θ 1 , θ 2 )-matrices: c d δ θ 1 g h θ 0 a b θ 2 e f ◮ This condition is abbreviated C TC ( θ 0 , θ 1 , θ 2 ; δ ).

  20. Centrality ◮ Here is a picture of C TC ( θ 1 , θ 2 , θ 0 ; δ ): δ c d 1 g h 0 a b 2 e f

  21. Matrices ◮ For congruences θ 0 , θ 1 , θ 2 we set � [ θ 0 , θ 1 , θ 2 ] TC = { δ : C TC ( θ 0 , θ 1 , θ 2 ; δ ) }

  22. Matrices ◮ For congruences θ 0 , θ 1 , θ 2 we set � [ θ 0 , θ 1 , θ 2 ] TC = { δ : C TC ( θ 0 , θ 1 , θ 2 ; δ ) } ◮ Higher centrality and the commutator for arity ≥ 4 are similarly defined.

  23. Matrices ◮ An n -dimensional hypercube is the graph H n = � 2 n ; E � , where two functions f , g ∈ 2 n are connected by an edge if and only if their outputs differ in exactly one argument.

  24. Matrices ◮ An n -dimensional hypercube is the graph H n = � 2 n ; E � , where two functions f , g ∈ 2 n are connected by an edge if and only if their outputs differ in exactly one argument. ◮ We say that a relation R on a set A is n -dimensional if R ⊆ A 2 n

  25. Matrices ◮ An n -dimensional hypercube is the graph H n = � 2 n ; E � , where two functions f , g ∈ 2 n are connected by an edge if and only if their outputs differ in exactly one argument. ◮ We say that a relation R on a set A is n -dimensional if R ⊆ A 2 n ◮ Observation: The term condition definition of centrality involving n -many congruences θ 0 , . . . , θ n − 1 is a condition that is quantified over ( θ 0 , . . . , θ n − 1 ) -matrices , which are certain n -dimensional invariant relations M ( θ 0 , . . . , θ n − 1 ) ≤ A 2 n that have generators of the form ( n − 1)-dimensional cube f ∈ 2 n such that f ( i ) = 0 x θ i y f ∈ 2 n such that f ( i ) = 1

  26. ◮ Consider the n -dimensional hypercube H n = � 2 n ; E � . For any coordinate i ∈ n , there are two ( n − 1)-dimensional hyperfaces that are ‘perpendicular’ to i :

  27. ◮ Consider the n -dimensional hypercube H n = � 2 n ; E � . For any coordinate i ∈ n , there are two ( n − 1)-dimensional hyperfaces that are ‘perpendicular’ to i : i = �{ f ∈ 2 n : f ( i ) = 0 } ; E � and 1. ( H n ) 0

  28. ◮ Consider the n -dimensional hypercube H n = � 2 n ; E � . For any coordinate i ∈ n , there are two ( n − 1)-dimensional hyperfaces that are ‘perpendicular’ to i : i = �{ f ∈ 2 n : f ( i ) = 0 } ; E � and 1. ( H n ) 0 i = �{ f ∈ 2 n : f ( i ) = 1 } ; E � . 2. ( H n ) 1

  29. ◮ Consider the n -dimensional hypercube H n = � 2 n ; E � . For any coordinate i ∈ n , there are two ( n − 1)-dimensional hyperfaces that are ‘perpendicular’ to i : i = �{ f ∈ 2 n : f ( i ) = 0 } ; E � and 1. ( H n ) 0 i = �{ f ∈ 2 n : f ( i ) = 1 } ; E � . 2. ( H n ) 1 (0 , 1 , 0 , 0) (1 , 1 , 0 , 0) (0 , 1 , 0 , 1) (1 , 1 , 0 , 1) (0 , 1 , 1 , 0) (1 , 1 , 1 , 0) (0 , 1 , 1 , 1) (1 , 1 , 1 , 1) (0 , 0 , 0 , 1) (1 , 0 , 0 , 1) (0 , 0 , 1 , 1) (1 , 0 , 1 , 1) (0 , 0 , 0 , 0) (1 , 0 , 0 , 0) (0 , 0 , 1 , 0) (1 , 0 , 1 , 0)

  30. ◮ Consider the n -dimensional hypercube H n = � 2 n ; E � . For any coordinate i ∈ n , there are two ( n − 1)-dimensional hyperfaces that are ‘perpendicular’ to i : i = �{ f ∈ 2 n : f ( i ) = 0 } ; E � and 1. ( H n ) 0 i = �{ f ∈ 2 n : f ( i ) = 1 } ; E � . 2. ( H n ) 1 (0 , 1 , 0 , 0) (1 , 1 , 0 , 0) (0 , 1 , 0 , 1) (1 , 1 , 0 , 1) (0 , 1 , 1 , 0) (1 , 1 , 1 , 0) (0 , 1 , 1 , 1) (1 , 1 , 1 , 1) (0 , 0 , 0 , 1) (1 , 0 , 0 , 1) (0 , 0 , 1 , 1) (1 , 0 , 1 , 1) (0 , 0 , 0 , 0) (1 , 0 , 0 , 0) (0 , 0 , 1 , 0) (1 , 0 , 1 , 0) ( H n ) 0 3 and ( H n ) 1 3

  31. ◮ Consider the n -dimensional hypercube H n = � 2 n ; E � . For any coordinate i ∈ n , there are two ( n − 1)-dimensional hyperfaces that are ‘perpendicular’ to i : i = �{ f ∈ 2 n : f ( i ) = 0 } ; E � and 1. ( H n ) 0 i = �{ f ∈ 2 n : f ( i ) = 1 } ; E � . 2. ( H n ) 1 (0 , 1 , 0 , 0) (1 , 1 , 0 , 0) (0 , 1 , 0 , 1) (1 , 1 , 0 , 1) (0 , 1 , 1 , 0) (1 , 1 , 1 , 0) (0 , 1 , 1 , 1) (1 , 1 , 1 , 1) (0 , 0 , 0 , 1) (1 , 0 , 0 , 1) (0 , 0 , 1 , 1) (1 , 0 , 1 , 1) (0 , 0 , 0 , 0) (1 , 0 , 0 , 0) (0 , 0 , 1 , 0) (1 , 0 , 1 , 0) ( H n ) 0 0 and ( H n ) 1 0

  32. ◮ Take h ∈ A 2 n . We consider h as a vertex labeled n -dimensional hypercube. For any coordinate i ∈ n , there are two ( n − 1)-dimensional vertex labeled hyperfaces that are perpendicular to i , which we denote

  33. ◮ Take h ∈ A 2 n . We consider h as a vertex labeled n -dimensional hypercube. For any coordinate i ∈ n , there are two ( n − 1)-dimensional vertex labeled hyperfaces that are perpendicular to i , which we denote 1. h 0 i and

  34. ◮ Take h ∈ A 2 n . We consider h as a vertex labeled n -dimensional hypercube. For any coordinate i ∈ n , there are two ( n − 1)-dimensional vertex labeled hyperfaces that are perpendicular to i , which we denote 1. h 0 i and 2. h 1 i .

  35. ◮ Take h ∈ A 2 n . We consider h as a vertex labeled n -dimensional hypercube. For any coordinate i ∈ n , there are two ( n − 1)-dimensional vertex labeled hyperfaces that are perpendicular to i , which we denote 1. h 0 i and 2. h 1 i . c d l k g s o p i j m n a b h ∈ A 2 n e f

  36. ◮ Take h ∈ A 2 n . We consider h as a vertex labeled n -dimensional hypercube. For any coordinate i ∈ n , there are two ( n − 1)-dimensional vertex labeled hyperfaces that are perpendicular to i , which we denote 1. h 0 i and 2. h 1 i . c d l k g s o p i j m n a b h ∈ A 2 n e f h 0 3 and h 1 3

  37. ◮ Take h ∈ A 2 n . We consider h as a vertex labeled n -dimensional hypercube. For any coordinate i ∈ n , there are two ( n − 1)-dimensional vertex labeled hyperfaces that are perpendicular to i , which we denote 1. h 0 i and 2. h 1 i . c d l k g s o p i j m n a b h ∈ A 2 n e f h 0 1 and h 1 1

  38. ◮ For R ⊆ A 2 n , set R i = {� h 0 i , h 1 i � : h ∈ R } .

  39. ◮ For R ⊆ A 2 n , set R i = {� h 0 i , h 1 i � : h ∈ R } . ◮ Fact: Suppose A is a member of a permutable variety, and take ( θ 0 , . . . , θ n − 1 ) ∈ Con( A ) n . Then, M ( θ 0 , . . . , θ n − 1 ) i is a congruence relation, for all i ∈ n .

  40. ◮ For R ⊆ A 2 n , set R i = {� h 0 i , h 1 i � : h ∈ R } . ◮ Fact: Suppose A is a member of a permutable variety, and take ( θ 0 , . . . , θ n − 1 ) ∈ Con( A ) n . Then, M ( θ 0 , . . . , θ n − 1 ) i is a congruence relation, for all i ∈ n . ◮ This leads to a nice characterization of the commutator for permutable varieties.

  41. Theorem (Binary Commutator) Let V be a permutable variety and let A ∈ V . For α, β ∈ Con( A ) , the following are equivalent:

  42. Theorem (Binary Commutator) Let V be a permutable variety and let A ∈ V . For α, β ∈ Con( A ) , the following are equivalent: 1. � x , y � ∈ [ α, β ] TC

  43. Theorem (Binary Commutator) Let V be a permutable variety and let A ∈ V . For α, β ∈ Con( A ) , the following are equivalent: 1. � x , y � ∈ [ α, β ] TC � x � y 2. ∈ M ( α, β ) x x

  44. Theorem (Binary Commutator) Let V be a permutable variety and let A ∈ V . For α, β ∈ Con( A ) , the following are equivalent: 1. � x , y � ∈ [ α, β ] TC � x � y 2. ∈ M ( α, β ) x x � a � y 3. ∈ M ( α, β ) for some a ∈ A a x

  45. Theorem (Binary Commutator) Let V be a permutable variety and let A ∈ V . For α, β ∈ Con( A ) , the following are equivalent: 1. � x , y � ∈ [ α, β ] TC � x � y 2. ∈ M ( α, β ) x x � a � y 3. ∈ M ( α, β ) for some a ∈ A a x � x � y 4. ∈ M ( α, β ) for some b ∈ A. b b

  46. ◮ Let V be a modular variety and let A ∈ V . For α, β ∈ Con( A ), define ∆ α,β to be the transitive closure of M ( α, β ) 0 .

  47. ◮ Let V be a modular variety and let A ∈ V . For α, β ∈ Con( A ), define ∆ α,β to be the transitive closure of M ( α, β ) 0 . a e 0 e 1 e 2 e n − 1 e n c b f 0 f 1 f 2 f n − 1 f n d a c 1 ∈ ∆ α,β 0 b d

  48. ◮ Let V be a modular variety and let A ∈ V . For α, β ∈ Con( A ), define ∆ α,β to be the transitive closure of M ( α, β ) 0 . a e 0 e 1 e 2 e n − 1 e n c b f 0 f 1 f 2 f n − 1 f n d a c 1 ∈ ∆ α,β 0 b d ◮ Fact: Both (∆ α,β ) 0 and (∆ α,β ) 1 are congruence relations.

  49. Theorem (Binary Commutator) Let V be a modular variety and let A ∈ V . For α, β ∈ Con( A ) , the following are equivalent: 1. � x , y � ∈ [ α, β ] TC � x � y 2. ∈ ∆ α,β x x � a � y 3. ∈ ∆ α,β for some a ∈ A a x � x � y 4. ∈ ∆ α,β for some b ∈ A. b b

  50. Theorem: Let V be a permutable variety. Take θ 0 , θ 1 , θ 2 ∈ Con( A ) for A ∈ V . The following are equivalent: (1) � x, y � ∈ [ θ 0 , θ 1 , θ 2 ] x x (2) x y ∈ M ( θ 0 , θ 1 , θ 2 ) x x x x There exist elements of A such that b a h x (3) (5) c y h y ∈ M ( θ 0 , θ 1 , θ 2 ) ∈ M ( θ 0 , θ 1 , θ 2 ) b a i j θ 1 c x i j d d θ 0 (4) ∈ M ( θ 0 , θ 1 , θ 2 ) x y θ 2 e e f f

  51. Theorem: Let V be a modular variety. Take θ 0 , θ 1 , θ 2 ∈ Con( A ) for A ∈ V . The following are equivalent: (1) � x, y � ∈ [ θ 0 , θ 1 , θ 2 ] x x (2) x y ∈ ∆ θ 0 ,θ 1 ,θ 2 x x x x There exist elements of A such that b a h x (3) c y (5) ∈ ∆ θ 0 ,θ 1 ,θ 2 h y ∈ ∆ θ 0 ,θ 1 ,θ 2 b a i j θ 1 c x i j d d θ 0 (4) ∈ ∆ θ 0 ,θ 1 ,θ 2 x y θ 2 e e f f

  52. Higher Dimensional Congruence Relations Definition Let R ⊆ A 2 n be an n -dimensional relation on some set A . R is called an n -dimensional equivalence relation if for all i ∈ n , each R i is an equivalence relation.

  53. Higher Dimensional Congruence Relations Definition Let R ⊆ A 2 n be an n -dimensional relation on some set A . R is called an n -dimensional equivalence relation if for all i ∈ n , each R i is an equivalence relation. Definition Let A be an algebra with underlying set A . Let R ∈ A 2 n be an n -dimensional equivalence relation. R is called an n -dimensional congruence if R is preserved by the basic operations of A .

  54. Higher Dimensional Congruence Relations Definition Let R ⊆ A 2 n be an n -dimensional relation on some set A . R is called an n -dimensional equivalence relation if for all i ∈ n , each R i is an equivalence relation. Definition Let A be an algebra with underlying set A . Let R ∈ A 2 n be an n -dimensional equivalence relation. R is called an n -dimensional congruence if R is preserved by the basic operations of A . ◮ Fix n ≥ 1. The collection of all n -dimensional congruences of an algebra A is an algebraic lattice, which we denote by Con n ( A ).

  55. Higher Dimensional Congruence Relations Definition Let R ⊆ A 2 n be an n -dimensional relation on some set A . R is called an n -dimensional equivalence relation if for all i ∈ n , each R i is an equivalence relation. Definition Let A be an algebra with underlying set A . Let R ∈ A 2 n be an n -dimensional equivalence relation. R is called an n -dimensional congruence if R is preserved by the basic operations of A . ◮ Fix n ≥ 1. The collection of all n -dimensional congruences of an algebra A is an algebraic lattice, which we denote by Con n ( A ). ◮ There are n distinct embeddings from Con 1 ( A ) into Con n ( A ).

  56. Con 1 ( A )

  57. α φ 0 2 Con 1 ( A ) Con 2 ( A )

  58. α φ 0 2 Con 1 ( A ) Con 2 ( A ) � � x � � y φ 0 2 ( α ) = : � x, y � ∈ α x y

  59. φ 1 2 α β φ 0 2 Con 1 ( A ) Con 2 ( A ) � � x � � y φ 0 2 ( α ) = : � x, y � ∈ α x y � � x � � x φ 1 2 ( β ) = : � x, y � ∈ β y y

  60. Define ∆ α,β = φ 0 2 ( α ) ∨ φ 1 2 ( β ) ∆ α,β φ 1 2 α β φ 0 2 Con 1 ( A ) Con 2 ( A ) � � x � � y φ 0 2 ( α ) = : � x, y � ∈ α x y � � x � � x φ 1 2 ( β ) = : � x, y � ∈ β y y

  61. Higher Dimensional Congruence Relations ◮ Fix a dimension n and take i ∈ n . For a pair � x , y � ∈ A 2 , let Cube i ( � x , y � ) ∈ A 2 n be such that

  62. Higher Dimensional Congruence Relations ◮ Fix a dimension n and take i ∈ n . For a pair � x , y � ∈ A 2 , let Cube i ( � x , y � ) ∈ A 2 n be such that � � 0 1. Cube i ( � x , y � ) i is the ( n − 1)-dimensional cube with each vertex labeled by x .

  63. Higher Dimensional Congruence Relations ◮ Fix a dimension n and take i ∈ n . For a pair � x , y � ∈ A 2 , let Cube i ( � x , y � ) ∈ A 2 n be such that � � 0 1. Cube i ( � x , y � ) i is the ( n − 1)-dimensional cube with each vertex labeled by x .

  64. Higher Dimensional Congruence Relations ◮ Fix a dimension n and take i ∈ n . For a pair � x , y � ∈ A 2 , let Cube i ( � x , y � ) ∈ A 2 n be such that � � 0 1. Cube i ( � x , y � ) i is the ( n − 1)-dimensional cube with each vertex labeled by x . � � 1 2. Cube i ( � x , y � ) i is the ( n − 1)-dimensional cube with each vertex labeled by y . ◮ Define φ i n : Con 1 ( A ) → Con n ( A ) by φ i n ( α ) = { Cube i ( � x , y � ) : � x , y � ∈ α }

  65. i φ i Define ∆ θ 0 ,...,θ n − 1 = � n ( θ i ) ∆ θ 0 ,...,θ n − 1 φ n − 1 n θ 0 θ n − 1 φ 0 n Con 1 ( A ) Con n ( A )

  66. Characterizing Joins ◮ Let A be an algebra and let θ be an equivalence relation on A . Then, θ is an admissible relation if and only if θ is compatible with the unary polynomials of A .

  67. Characterizing Joins ◮ Let A be an algebra and let θ be an equivalence relation on A . Then, θ is an admissible relation if and only if θ is compatible with the unary polynomials of A . ◮ This generalizes to: Theorem Let A be an algebra and let n ≥ 1 . An n-dimensional equivalence relation θ is admissible if and only if θ is compatible with the n-ary polynomials of A .

  68. Proof Idea a 0 a 1 a 2 a 3 b 0 b 1 b 2 b 3 ∈ θ Take d 0 , d 1 , d 2 , c 0 c 1 c 2 c 3 d 3

  69. Proof Idea a 0 a 1 a 2 a 3 b 0 b 1 b 2 b 3 ∈ θ Take Then, d 0 , d 1 , d 2 , c 0 c 1 c 2 c 3 d 3 a 1 a 1 a 0 b 0 b 0 b 0 b 0 b 1 b 1 b 1 c 0 a 1 a 1 d 0 b 1 b 1 b 1 d 0 d 0 d 0 c 0 d 0 d 0 c 1 c 1 d 0 d 0 d 1 d 1 d 1 c 1 c 0 c 1 d 0 d 0 d 0 d 0 d 1 d 1 d 1 c 1 c 0 c 1 d 0 d 0 d 0 d 1 d 1 d 1 d 0 ∈ θ a 2 a 2 a 2 a 3 a 3 a 3 a 3 b 2 b 2 b 3 a 2 a 3 a 3 a 3 a 3 a 2 a 2 b 2 b 2 b 3 a 2 a 2 a 3 a 3 a 3 a 3 a 2 b 2 b 2 b 3 c 2 c 2 a 3 a 3 a 3 a 3 c 2 d 2 d 2 b 3 c 2 c 2 c 2 c 3 c 3 c 3 c 3 d 3 d 2 d 2

  70. Proof Idea a 0 a 1 a 2 a 3 b 0 b 1 b 2 b 3 Take ∈ θ Then, d 0 , d 1 , d 2 , c 0 c 1 c 2 c 3 d 3 a 1 a 1 a 0 b 0 b 1 b 1 b 0 b 0 b 0 b 1 c 0 a 1 a 1 d 0 d 0 d 0 d 0 b 1 b 1 b 1 c 0 c 1 c 1 d 0 d 0 d 0 d 0 d 1 d 1 d 1 c 0 c 1 d 0 d 0 d 0 d 0 c 1 d 1 d 1 d 1 c 1 c 1 c 0 d 0 d 0 d 0 d 1 d 1 d 0 d 1 ∈ θ a 2 a 2 a 2 a 3 a 3 a 3 a 3 b 2 b 2 b 3 Compatibility a 3 a 2 a 2 a 2 b 2 a 3 a 3 a 3 b 3 b 2 with binary polynomials is a 3 a 2 a 2 a 2 b 2 a 3 a 3 a 3 b 2 b 3 sufficient to show compatibility c 2 c 2 a 3 a 3 a 3 a 3 c 2 b 3 d 2 d 2 with a 4-ary operation. c 2 c 2 c 2 c 3 c 3 c 3 c 3 d 2 d 2 d 3

  71. Characterizing Joins ◮ ∆ θ 0 ,...,θ n − 1 = � i φ i n ( θ i ) is therefore obtained by 1. Closing � φ i n ( θ i ) under all n -ary polynomials and then

Recommend


More recommend