Is propositional logic the same as Boolean logic?

The terms are closely related in common usage, particularly for true/false formulas, though their contexts can emphasize different applications.

Can it represent uncertainty?

Ordinary classical propositional logic does not assign probabilities. A separate probabilistic layer can reason about uncertain propositions.

Is a satisfiable rule set guaranteed to describe the real world?

No. Satisfiability means its formal conditions can hold together; it does not verify that its premises are accurate or complete.

AI Glossary / Graphs and symbolic AI / Intermediate

Propositional logic

Definition

Propositional logic is a formal system for reasoning with statements that are treated as true or false and combined using logical connectives. It represents whole statements with symbols and expresses relationships through operations such as not, and, or, implication, and equivalence. It is a foundation for symbolic AI, satisfiability solving, and the analysis of explicit rules.

What it means

Let R mean “a document has been reviewed” and A mean “it has been approved.” The formula R implies A expresses a relationship between those statements. Propositional logic specifies exactly which truth-value combinations satisfy that formula.
The system does not inspect the internal structure of R or A. It cannot automatically express “every reviewer checks some document” using variables and quantifiers; that requires a richer language or a separate finite encoding. Its strength is a clear treatment of combinations of yes-or-no conditions. The model's usefulness still depends on whether those conditions accurately represent the intended situation.

How it works

Define atomic proposition symbols and combine them into formulas. An interpretation assigns true or false to every symbol. Evaluate each compound formula according to its connectives: a conjunction needs both parts, an inclusive disjunction needs at least one, and an implication is false only when its premise is true and its conclusion is false.
A knowledge base entails a conclusion when the conclusion is true in every interpretation satisfying the knowledge base. One can check this by enumerating assignments or use more efficient reasoning methods. Satisfiability asks whether at least one assignment makes a formula true; validity asks whether every assignment does. These questions are related but different.

Three practical examples

A configuration tool can encode that choosing one option requires another and that two options cannot be selected together.
A workflow prototype can check whether a release condition follows from recorded review and approval conditions.
A puzzle-solving lesson can represent clues as formulas and identify the assignments consistent with all clues.

Think of it this way

Think of a carefully defined checklist in which some boxes imply others and some combinations are forbidden. Logic examines the allowed combinations systematically. The analogy is limited because a checklist's informal wording can remain ambiguous, while the formal system operates only on its precise symbols and connectives; it does not automatically resolve what people intended those labels to mean.

Worked example

Consider an invented configuration with three options: S for standard package, E for express package, and T for tracking. Require exactly one package and require tracking whenever express is chosen. The knowledge base is (S or E) and not(S and E) and (E implies T). Here “or” is inclusive; the second condition separately excludes choosing both packages.
There are eight assignments for three symbols. Exactly-one-package constraints leave four: standard with or without tracking, and express with or without tracking. The final implication excludes express without tracking. Three valid configurations remain: S true, E false, T false; S true, E false, T true; and S false, E true, T true.
The knowledge base does not entail T, because standard without tracking is a counterexample. Adding E as a fact leaves only express with tracking, so T then follows. The formula is satisfiable because it has valid configurations, but it is not valid because some assignments violate its rules.

Illustrative case study

Fictional case: Elm Configurator encodes two product options using the word “or.” Business reviewers mean exactly one, while the implementation uses inclusive disjunction. The model consequently accepts customers selecting both options.
The team writes the exclusivity condition explicitly and checks all small combinations before extending the rule set. It also confirms that express requires tracking without assuming the converse. This fictional exercise shows why formalization needs a discussion of meaning as well as syntax. It does not claim a real product launch, customer outcome, or measured efficiency improvement.

Common misconceptions and mistakes

Do not confuse implication with causation. R implies A constrains truth assignments; it does not explain a physical process or prove that changing R causes A. Do not reverse an implication without a separate justification.
Ordinary propositional disjunction is inclusive, so both sides may be true. A biconditional is stronger than a one-way implication because it requires both directions. Missing premises do not automatically establish negation. An inconsistent classical knowledge base has no satisfying models and entails every formula, making consistency checking essential to interpreting its conclusions. More symbols can also make exhaustive truth-table enumeration impractical.

Business use

Propositional logic can help teams test small rule sets for contradictions, gaps, and unintended permissions. A useful output is often a counterexample: an allowed combination that stakeholders thought should be forbidden, or a forbidden combination they expected to allow.
Begin with precise definitions and a manageable set of conditions. Ask domain reviewers to inspect representative models and countermodels. Keep the formal rules connected to the corresponding business requirements, because a correct solver cannot detect an unstated intention. When relationships involve many objects, quantities, or changing time points, consider whether another representation would express the task more naturally.

Check your understanding

Does E implies T mean that every tracked package is express?

Answer: No. It rules out express without tracking. The worked model explicitly allows a standard package with tracking, which makes T true and E false while satisfying the implication. To require the converse, one would need T implies E as an additional rule.

Related terms

Learn next

Learn satisfiability and conjunctive normal form next. Add the rule S implies not T to the configuration example, then enumerate the remaining models. Determine whether tracking now identifies express uniquely, and justify the conclusion by checking every permitted assignment.

Sources

Stuart Russell and Peter Norvig, AIMA Chapter 7 teaching slides, propositional logic, entailment and satisfiability: https://aima.cs.berkeley.edu/slides-pdf/chapter07.pdf

Previous: Proposition symbol Next: Prosodic phrasing