On the two kinds of probability in algebra (Q583255): Difference between revisions
From MaRDI portal
Created a new Item |
ReferenceBot (talk | contribs) Changed an Item |
||
(5 intermediate revisions by 4 users not shown) | |||
Property / review text | |||
A groupoid table on the set \(\{\) 0,1,...,n-1\(\}\) is an \(n\times n\) matrix with entries from \(\{\) 0,1,...,n-1\(\}\). Let \({\mathcal T}_ n\) be the set of groupoid tables on the set \(\{\) 0,1,...,n-1\(\}\). If P is a groupoid property, we define the probability of P by \[ \Pr (P;{\mathcal T})=\lim_{n\to \infty}\Pr (P;{\mathcal T}_ n) \] where \[ \Pr (P;{\mathcal T}_ n)=| \{{\mathbb{T}}\in {\mathcal T}_ n:\quad {\mathbb{T}}\vDash P\}| /n^{n^ 2}. \] Of course, any given groupoid may be represented by many tables, so a more natural definition of probability would seem to be one which took account of this. More precisely, let \({\mathcal K}\) be any infinite set of finite algebras such that \({\mathcal K}_ n\) (the set of n- element mumbers of \({\mathcal K})\) is finite. Then the probability of a property P in \({\mathcal K}\) is defined to be \[ \Pr (P;{\mathcal K})=\lim_{n\to \infty}\Pr (P;{\mathcal K}_ n) \] where \[ \Pr (P;{\mathcal K}_ n)=| \{{\mathbb{A}}\in {\mathcal K}_ n:\quad {\mathbb{A}}\vDash P\}| /| {\mathcal K}_ n| \] and the limit is taken over those n for which \({\mathcal K}_ n\neq \emptyset.\) These probability concepts which, in analogy with graph theory, we might describe as labelled and unlabelled, can be generalized to any finite similarity type, \(\tau\). The interesting main result of this paper is that if \(\tau\) has an operation of arity at least 2, or has at least three unary operations, then these probabilities coincide. An example is given to show that these probabilities can differ for 2-unary algebras. \textit{J. F. Lynch} [Trans. Am. Math. Soc. 287, 543-568 (1985; Zbl 0526.03012)] has shown that, for unary algebras, the labelled probability of any property expressible in first order logic must exist, while \textit{K. J. Compton}, \textit{C. W. Henson} and \textit{S. Shelah} [Ann. Pure Appl. Logic 36, 207-224 (1987; Zbl 0632.03037)] have given an example showing that this is not the case for algebras with a binary operation. The results of this paper show that the same is true for unlabelled probability, provided that, in the unary case, there are at least three operations. | |||
Property / review text: A groupoid table on the set \(\{\) 0,1,...,n-1\(\}\) is an \(n\times n\) matrix with entries from \(\{\) 0,1,...,n-1\(\}\). Let \({\mathcal T}_ n\) be the set of groupoid tables on the set \(\{\) 0,1,...,n-1\(\}\). If P is a groupoid property, we define the probability of P by \[ \Pr (P;{\mathcal T})=\lim_{n\to \infty}\Pr (P;{\mathcal T}_ n) \] where \[ \Pr (P;{\mathcal T}_ n)=| \{{\mathbb{T}}\in {\mathcal T}_ n:\quad {\mathbb{T}}\vDash P\}| /n^{n^ 2}. \] Of course, any given groupoid may be represented by many tables, so a more natural definition of probability would seem to be one which took account of this. More precisely, let \({\mathcal K}\) be any infinite set of finite algebras such that \({\mathcal K}_ n\) (the set of n- element mumbers of \({\mathcal K})\) is finite. Then the probability of a property P in \({\mathcal K}\) is defined to be \[ \Pr (P;{\mathcal K})=\lim_{n\to \infty}\Pr (P;{\mathcal K}_ n) \] where \[ \Pr (P;{\mathcal K}_ n)=| \{{\mathbb{A}}\in {\mathcal K}_ n:\quad {\mathbb{A}}\vDash P\}| /| {\mathcal K}_ n| \] and the limit is taken over those n for which \({\mathcal K}_ n\neq \emptyset.\) These probability concepts which, in analogy with graph theory, we might describe as labelled and unlabelled, can be generalized to any finite similarity type, \(\tau\). The interesting main result of this paper is that if \(\tau\) has an operation of arity at least 2, or has at least three unary operations, then these probabilities coincide. An example is given to show that these probabilities can differ for 2-unary algebras. \textit{J. F. Lynch} [Trans. Am. Math. Soc. 287, 543-568 (1985; Zbl 0526.03012)] has shown that, for unary algebras, the labelled probability of any property expressible in first order logic must exist, while \textit{K. J. Compton}, \textit{C. W. Henson} and \textit{S. Shelah} [Ann. Pure Appl. Logic 36, 207-224 (1987; Zbl 0632.03037)] have given an example showing that this is not the case for algebras with a binary operation. The results of this paper show that the same is true for unlabelled probability, provided that, in the unary case, there are at least three operations. / rank | |||
Normal rank | |||
Property / Mathematics Subject Classification ID | |||
Property / Mathematics Subject Classification ID: 08A99 / rank | |||
Normal rank | |||
Property / Mathematics Subject Classification ID | |||
Property / Mathematics Subject Classification ID: 08A35 / rank | |||
Normal rank | |||
Property / Mathematics Subject Classification ID | |||
Property / Mathematics Subject Classification ID: 60B99 / rank | |||
Normal rank | |||
Property / zbMATH DE Number | |||
Property / zbMATH DE Number: 4132233 / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
probability of a groupoid property | |||
Property / zbMATH Keywords: probability of a groupoid property / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
groupoid table | |||
Property / zbMATH Keywords: groupoid table / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
finite similarity type | |||
Property / zbMATH Keywords: finite similarity type / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
2-unary algebras | |||
Property / zbMATH Keywords: 2-unary algebras / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
labelled probability | |||
Property / zbMATH Keywords: labelled probability / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
unlabelled probability | |||
Property / zbMATH Keywords: unlabelled probability / rank | |||
Normal rank | |||
Property / reviewed by | |||
Property / reviewed by: Sheila Oates-Williams / rank | |||
Normal rank | |||
Property / MaRDI profile type | |||
Property / MaRDI profile type: MaRDI publication profile / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: Nonconvergence, undecidability, and intractability in asymptotic problems / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: Probabilities on finite models / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: The number of finite relational structures / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: Probabilities of First-Order Sentences about Unary Functions / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: Q3770587 / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: Kombinatorische Anzahlbestimmungen für Gruppen, Graphen und chemische Verbindungen / rank | |||
Normal rank | |||
links / mardi / name | links / mardi / name | ||
Revision as of 13:00, 20 June 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | On the two kinds of probability in algebra |
scientific article |
Statements
On the two kinds of probability in algebra (English)
0 references
1990
0 references
A groupoid table on the set \(\{\) 0,1,...,n-1\(\}\) is an \(n\times n\) matrix with entries from \(\{\) 0,1,...,n-1\(\}\). Let \({\mathcal T}_ n\) be the set of groupoid tables on the set \(\{\) 0,1,...,n-1\(\}\). If P is a groupoid property, we define the probability of P by \[ \Pr (P;{\mathcal T})=\lim_{n\to \infty}\Pr (P;{\mathcal T}_ n) \] where \[ \Pr (P;{\mathcal T}_ n)=| \{{\mathbb{T}}\in {\mathcal T}_ n:\quad {\mathbb{T}}\vDash P\}| /n^{n^ 2}. \] Of course, any given groupoid may be represented by many tables, so a more natural definition of probability would seem to be one which took account of this. More precisely, let \({\mathcal K}\) be any infinite set of finite algebras such that \({\mathcal K}_ n\) (the set of n- element mumbers of \({\mathcal K})\) is finite. Then the probability of a property P in \({\mathcal K}\) is defined to be \[ \Pr (P;{\mathcal K})=\lim_{n\to \infty}\Pr (P;{\mathcal K}_ n) \] where \[ \Pr (P;{\mathcal K}_ n)=| \{{\mathbb{A}}\in {\mathcal K}_ n:\quad {\mathbb{A}}\vDash P\}| /| {\mathcal K}_ n| \] and the limit is taken over those n for which \({\mathcal K}_ n\neq \emptyset.\) These probability concepts which, in analogy with graph theory, we might describe as labelled and unlabelled, can be generalized to any finite similarity type, \(\tau\). The interesting main result of this paper is that if \(\tau\) has an operation of arity at least 2, or has at least three unary operations, then these probabilities coincide. An example is given to show that these probabilities can differ for 2-unary algebras. \textit{J. F. Lynch} [Trans. Am. Math. Soc. 287, 543-568 (1985; Zbl 0526.03012)] has shown that, for unary algebras, the labelled probability of any property expressible in first order logic must exist, while \textit{K. J. Compton}, \textit{C. W. Henson} and \textit{S. Shelah} [Ann. Pure Appl. Logic 36, 207-224 (1987; Zbl 0632.03037)] have given an example showing that this is not the case for algebras with a binary operation. The results of this paper show that the same is true for unlabelled probability, provided that, in the unary case, there are at least three operations.
0 references
probability of a groupoid property
0 references
groupoid table
0 references
finite similarity type
0 references
2-unary algebras
0 references
labelled probability
0 references
unlabelled probability
0 references