Complete theories

From MaRDI portal





This is a continuation of the author's work on the mathematics of algebra, i. e. the use of mathematical logic to derive general algebraic theorems. As in his earlier work [The metamathematics of algebra. Amsterdam: North-Holland (1950; Zbl 0043.24702)] he defines a structure as a non-empty set \(C\) of constants together with a non-empty set \(P\) of relations defined over \(C\). The diagram of such a structure \(M\) is the set of statements containing, for each relation \(R\) of \(P\) and each set \(a_1, \ldots, a_n\) of elements of \(C\) that one of the statements \(R(a_1, \ldots, a_n)\), \(\sim R(a_1, \ldots, a_n)\) which holds in \(M\). The logical system used is a version of the lower predicate calculus; a set \(K\) of statements is said to be complete if for each statement \(X\), containing no constants or relations not occurring in \(K\), either \(X\) or \(\sim X\) is deducible form \(K\). Most of the book is devoted to proving the completeness of axiom systems for various kinds of algebras. The first chapter is a brief survey of the basic concepts and of the logical theorems which are used, chief of which are Gödel's completeness theorem and its corollary that a necessary and sufficient condition for a set of statements \(K\) to possess a model is that every finite subset of \(K\) possesses a model. In the second chapter a notion of model-completeness is introduced, a non-empty consistent set \(K\) of statements being called model complete if, for each model \(M\) for \(K\), the set \(K \cup N\) is complete, \(N\) being the diagram of \(K\). It is shown that model-completeness neither implies nor is implied by completeness. It is easily seen that a necessary and sufficient condition for the model completeness of \(K\) is that for every pair of models \(M, M'\) of \(K\) such that \(M \subseteq M'\), every statement \(Y\) containing only constants and relations from \(M\), holds in \(M\) if and only if it holds in \(M'\). The main theorem of Chapter II is that the sufficiency still holds under the weaker assumption that the condition is satisfied for formulae or their negations. This considerably reduces the labour of proving model completeness. Chapter II contains other general theorems on model completeness and various related concepts. These are applied in Chapter III to prove the model completeness of axiom systems for various algebraic systems e. g. completely divisible torsion-free abelian groups, algebraically closed fields (four proofs of this are given, each one using more logic and less algebra than the preceding one), real-closed fields, algebraically closed fields with non-trivial valuation in an ordered abelian group etc. In Chapter IV it is shown that under certain conditions model-completeness entails completeness, e. g. if \(K\) has a prime model \(M_0\) such that every model \(M\) for \(K\) has a submodel isomorphic to \(M_0\). These results are used to prove the completeness of the axiom systems for various algebraic systems including completely divisible torsion-free abelian groups, real-closed fields, infinite abelian groups in which all elements except 0 are of order \(p\) \((p\) being a fixed prime), algebraically closed fields of specified characteristic, algebraically closed fields of specified prime characteristic with non-trivial valuation in an ordered abelian group. Chapter V contains applications of these metamathematical results to the proving of purely mathematical theorems. e. g. ``If \(F_1, F_2\) are two commutative ordered fields then there exists a commutative ordered field which includes subfields isomorphic to \(F_1\) and \(F_2\). ``Let \(M\) be an algebraically closed field and let \(M^*\) be the algebraic closure of \(M(\xi_1, \ldots, \xi_n)\), obtained by adjoining the indeterminates \(\xi_1, \ldots, \xi_n\) to \(M\). Let \(Q(x_1, \ldots, x_n)\) be a predicate formulated in terms of addition, multiplication and equality and in terms of the constants of \(M\). Suppose that \(Q(\xi_1, \ldots, \xi_n)\) holds in \(M^*\). Then there exists a polynomial \(p(x_1, \ldots, x_n)\) with coefficients in \(M\), not all zero, such that \(Q((\xi_1, \ldots, \xi_n))\) holds in \(M\) for all elements \(a_1, \ldots, a_n \in M \) for which \(p(a_1, \ldots, a_n) \neq 0\). The last chapter is concerned with a particular syntactical operator which associates with each statement \(X\) of a particular form a statement \(B\{X\}\) called the bounding transform of \(X\). A particular axiom set \( \Sigma \) is introduced and it is proved that, under certain conditions if \(X\) is deducible from a set of axioms \(K\) then \(B\{X\}\) is deducible from \(B\{K\} \cup \Sigma\). By taking a particular interpretation of the relations in \(\Sigma\) this theorem is used to obtain general transfer principles allowing one to derive theorems on normed algebras from theorems on the corresponding purely algebraic structures e.g. the statement that the sum of two elements \(x, y\) exists is transformed by \(B\) into the statement that the sum exists and is bounded uniformly, for given bounds on \(x\) and \(y\). The author promises to deal elsewhere with a similar continuity transformation allowing one to derive statements about the continuity of solutions to certain problems from statements of their mere existence.





Cited in
(72)








This page was built for publication: Complete theories

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q766973)