2.2.7: Summary
Sets of sentences in a sense describe the structures in which they are jointly true; these structures are their models . Conversely, if we start with a structure or set of structures, we might be interested in the set of sentences they are models of, this is the theory of the structure or set of structures. Any such set of sentences has the property that every sentence entailed by them is already in the set; they are closed . More generally, we call a set \(\Gamma\) a theory if it is closed under entailment, and say \(\Gamma\) is axiomatized by \(\Delta\) if \(\Gamma\) consists of all sentences entailed by \(\Delta\) .
Mathematics yields many examples of theories, e.g., the theories of linear orders, of groups, or theories of arithmetic, e.g., the theory axiomatized by Peano’s axioms. But there are many examples of important theories in other disciplines as well, e.g., relational databases may be thought of as theories, and metaphysics concerns itself with theories of parthood which can be axiomatized.
One significant question when setting up a theory for study is whether its language is expressive enough to allow us to formulate everything we want the theory to talk about, and another is whether it is strong enough to prove what we want it to prove. To express a relation we need a formula with the requisite number of free variables. In set theory , we only have \(\in\) as a relation symbol, but it allows us to express \(x \subseteq y\) using \(\lforall{u}{(u \in x \lif u \in y)}\) . Zermelo-Fraenkel set theory \(\Log{ZFC}\) , in fact, is strong enough to both express (almost) every mathematical claim and to (almost) prove every mathematical theorem using a handful of axioms and a chain of increasingly complicated definitions such as that of \(\subseteq\) .