Can an atom contain a variable?

Yes. It is then nonground unless the variable has been replaced by an appropriate ground term.

Is NOT approved(box7) an atom?

In ordinary classical terminology, it is a negative literal built by negating the atom approved(box7).

Does an atom have to mention two objects?

No. Its number of arguments depends on the predicate's arity; unary and zero-argument forms are also possible.

AI Glossary / Graphs and symbolic AI / Beginner

Atom

Definition

An atom is a basic logical formula that is not built by combining simpler formulas with logical connectives or quantifiers. In predicate logic, it typically consists of a predicate applied to the appropriate number of terms, such as owns(maya, box7). Atoms provide the building blocks for more complex statements and reasoning rules.

What it means

A logical atom can have internal structure even though it is not a compound formula. In owns(maya, box7), owns is the predicate symbol, while maya and box7 are terms occupying its two argument positions.

The complete expression represents a relationship under an interpretation. Its name does not make the relationship true. A knowledge representation system needs agreed meanings for the symbols and information or assumptions about whether the relationship holds.

This entry uses atom in formal logic. The word also has meanings in chemistry and programming-language data types. Those meanings should not be substituted automatically for the logical one, even when the same software supports logic programming.

How it works

A logical language defines its symbols and formation rules. Predicate symbols have an arity: the number of argument positions they accept. Terms fill those positions. In a simple relational language, terms can be constants or variables; richer first-order languages can also contain function terms.

An atom with no variables is ground. An atom containing variables is nonground. For example, owns(maya, box7) is ground, while owns(Person, box7) contains the variable Person under the convention used here.

An interpretation connects symbols to a domain and its relationships. For a ground atom, that interpretation determines whether the indicated relation holds. For an atom containing variables, evaluating it also requires assignments to the free variables or an enclosing quantified context.

Connectives and quantifiers build larger formulas from atoms. Negating an atom produces a negative literal. Joining two atoms with AND produces a conjunction. These expressions contain atoms but are not themselves single atoms.

Three practical examples

1. An inventory rule system represents located(box7, shelf2) as a basic relationship that can participate in rules about picking and storage.

2. A scheduling model uses available(room3, afternoon) as one condition when determining whether a booking meets its constraints.

3. A logic tutor asks learners to distinguish an individual atom from a negated atom and a conjunction, helping them understand the structure of formal expressions.

Think of it this way

Think of an atom as one entry in a structured statement form: a relationship name followed by its filled-in positions. More elaborate sentences can combine several such entries.

The analogy concerns logical organization. It does not mean every atom is verified information, and a form with a variable position still needs an assignment or quantified meaning before its truth can be evaluated in context.

Worked example

Define owns as a predicate with two arguments: first the owner, then the owned object. Let maya and noah name people, and box7 and box8 name boxes. Suppose a particular interpretation makes owns true exactly for the pairs (maya, box7) and (noah, box8).

The expression owns(maya, box7) is a ground atom and is true in this interpretation. The expression owns(maya, box8) is also a well-formed ground atom, but it is false in the same interpretation. Being an atom is a syntactic property, not a certificate of truth.

The expression owns(Person, box7) is a nonground atom. Assigning Person to Maya makes it true here; assigning Person to Noah makes it false.

Finally, NOT owns(maya, box8) is a negative literal, and owns(maya, box7) AND owns(noah, box8) is a conjunction containing two atoms. Their truth follows from this specified interpretation, not from the labels alone.

Illustrative case study

Fictional case: Beacon Archive builds a small rule-based catalog assistant. Its developers define stored(Document, Cabinet) and approved(Document), then use those relationships in retrieval rules.

During review, one developer reverses the arguments of stored. Another assumes that every syntactically valid atom is an accepted catalog fact. The team corrects both errors by documenting argument meanings and separating expression validation from evidence about the catalog.

The fictional case illustrates why basic logical vocabulary matters in practical systems. A rule engine can manipulate expressions consistently while still answering the wrong question if predicate meanings, argument order, or source information are misrepresented.

Common misconceptions and mistakes

An atom is not necessarily a true fact. False statements can have exactly the same atomic grammatical form as true ones.

Atomic does not mean that the expression contains no parts. Predicate atoms can contain several arguments, and richer languages can include complex terms inside those arguments.

Do not confuse a predicate symbol with a complete atom. A two-place predicate needs two suitable terms to form an application. Conversely, a language may permit a proposition-like predicate with no arguments.

Also distinguish missing evidence from falsity. Whether an unrecorded atom is treated as false depends on the reasoning system's semantics, not on the definition of atom itself.

Business use

Understanding atoms helps teams design rule-based systems, knowledge representations, and formal checks. Define each predicate's intended meaning, argument order, permitted values, and relationship to evidence.

Keep the distinction between valid syntax and supported content visible in testing. Include cases where an expression is well formed but false in a chosen interpretation, as well as cases with the wrong number of arguments. This makes it easier to detect modeling errors before they propagate through larger rules and conclusions.

Check your understanding

In the worked interpretation, owns(maya, box8) is false. Does that stop it from being an atom?

Answer: No. It remains a well-formed ground atom. Its grammatical category and its truth in an interpretation are separate matters.

Related terms

Predicate (logic)
Term
Arity
Literal
Interpretation
First-order logic
predicate
ground atom
logical variable
formula

Learn next

Write three predicates with different arities and explain every argument position. Build ground and nonground atoms from them, then form a negation and a conjunction. Finally, specify a small interpretation so that truth judgments depend on explicit relationships rather than suggestive names.

Interpretation

Sources

Primary source consulted: David Poole and Alan Mackworth, Artificial Intelligence: Foundations of Computational Agents, predicate-calculus syntax and semantics, https://artint.info/3e/html/ArtInt3e.Ch15.S3.html . Scope is the logical atom, distinguished from programming data types. The ownership interpretation and fictional catalog case are original teaching examples.

Previous: ATN Next: Atomic data types