


Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
An introduction to first-order logic, focusing on its syntax and semantics. It covers the extension of propositional logic through quantification, the formation of formulas, the concept of free and bound variables, and the definition of valuations and truth sets. The document also explains the differences between boolean and first-order valuations.
Typology: Study notes
1 / 4
This page cannot be seen from the preview
Don't miss anything!



Applied Logic Lecture 15: First-Order Logic CS 4860 Spring 2009 Tuesday, March 10, 2009
First-Order Logic is the calculus one usually has in mind when using the word “logic”. It is expres- sive enough for all of mathematics, except for those concepts that rely on a notion of construction or computation. However, dealing with more advanced concepts is often somewhat awkward and researchers often design specialized logics for that reason. Our account of first-order logic will be similar to the one of propositional logic. We will present
The syntax of first-order logic is essentially an extension of propositional logic by quantification ∀ and ∃. Propositional variables are replaced by n-ary predicate symbols (P , Q, R) which may be instantiated with either variables (x, y, z, ...) or parameters (a, b, ...). Here is a summary of the most important concepts.
Every atomic formula is a formula. If A and B are formulas and x is a variable then (A), ∼A, A ∧ B, A ∨ B, A⊃B, (∀x)A, and (∃x)A are formulas.
In (∀x)P x ∨ Qx the scope of (∀x) is just P x, while Qx is outside the scope of the quantifier. To include Qx in the scope of (∀x) one has to add parentheses: (∀x)(P x ∨ Qx). Note that the conventions about the scope of quantifiers differ in the literature.
That is, the root of the tree is F. The sucessor of a formula of the form ∼A is A. The successors of A ∧ B, A ∨ B, A⊃B are A and B. The successors of (∀x)A and (∃x)A are A|x ai for all parameters ai. Note that quantifiers usually have infinitely many successors. Atomic formulas have no successors.
The semantics of first-order logic, like the one of propositional logic and P 2 , is based on a concept of valuations. In propositional logic, it was sufficient to assign values to all propositional vari- ables and then extend the evaluation from atoms to formulas in a canonical fashion. In P 2 , the semantics of quantified formulas was defined in terms of the values of all immediate subformulas: v[(∀p)A] = (v|p f )[A] ∧B (v|p t )[A]. In first-order logic, we will proceed in the same way. However, since we don’t have propositional variables anymore, we have to explain the meaning of atomic formulas first.
The standard approach is to interpret parameters by elements of some universe U and n-ary pred- icates by subsets of U n. A closed formula P a 1 ..an then expresses the fact that the interpretations ki ∈^ U of the ai, taken together as n-tuple (k 1 , .., kn), form an element of the interpretation of P.
Smullyan’s approach is similar to the above idea but avoids set theory altogether. Instead, he introduces U -formulas , where the elements of the universe U are used as parameters and defines first-order valuations as canonical extensions of boolean valuations on the set EU^ of all closed U-formulas. The semantics of arbitrary formulas is then defined by a mapping ϕ from the set of parameters into U.
An atomic sentence P a 1 ..an is true under I if (ϕ(a 1 ), ..ϕ(an)) ∈^ ι(P ). In this manner, every inter- pretation induces an atomic valuation v 0 (together with ϕ) and vice versa and from now on we will use whatever notion is more convenient.
A formula A is called satisfiable if it is true under at least one interpretation I (i.e. under at least one universe U, one mapping ϕ, and one interpretation of the predicate symbols). I is also called a model of A. A is valid if A is true under every interpretation. These notions can be extended to sets of formula sin a canonical fashion.
It should be noted that there is a fine distinction between boolean valuations and first-order val- uations. Boolean valuations can only analyze the propositional structure of formulas. They can- not evaluate quantified formulas and therefore have to treat them like propositional variables. In contrast to that first-order valuations can analyze the internals of quantified formulas and extract information that is unaccessible to boolean valuations.
For instance, a boolean valuations would interpret the logical structure of the formula (∀x)(P x ∧ Qx)⊃(∀x)P x as P Q⊃P , which is obviously not a tautology. In contrast to that, every first-order valuation would go into the details of (∀x)(P x ∧ Qx) and (∀x)P x and evaluate to true. Thus the formula is valid, but not a tautology.
For the same reason, the formula (∀x)(P x ∧ Qx) ∧ (∃x)(∼P x) is truth-functionally satisfiable but not first-order satisfiable, since there is no first-order valuation (with a non-empty universe) that can make it true.
First-order valuations provide a more specific analysis than boolean valuations can give. They agree on quantifier-free formulas, however (Exercise!), and in that sense first-order logic is a canonical extension of propositional logic.