Can a ground term contain a function symbol?

Yes. A compound term is ground when all its arguments are ground.

Is a ground atom the same thing?

No. A ground atom is a variable-free atomic formula, usually formed by applying a predicate to terms.

Does replacing one variable always make a term ground?

No. Every variable occurrence must be eliminated through appropriate substitutions; other variables may remain inside nested arguments.

AI Glossary / Graphs and symbolic AI / Beginner

Ground term

Definition

A ground term is a term in formal logic that contains no variables. It can be a constant or a compound expression built from function symbols applied to other ground terms. Groundness is a property of the expression's syntax; it does not mean the expression is a verified fact, a real-world object name, or a measured value.

What it means

Terms are expressions used to refer to objects within a logical interpretation. A constant such as Parcel7 is ground. If owner is a function symbol, owner(Parcel7) is also ground because it contains no variable. By contrast, owner(x) is not ground when x is a variable.
Keep terms separate from formulas. Delivered(Parcel7), with Delivered a predicate, is a ground atomic formula rather than a ground term. It makes an assertion that can be true or false in a model. The term Parcel7 itself refers to an object and is not normally assigned a truth value.

How it works

To check groundness, inspect the entire expression recursively. A constant passes. A variable fails. A compound term passes only if every argument is ground. Function symbols may be nested, so checking only the outermost argument names is insufficient.
A substitution can turn a non-ground term into a ground one by replacing all its variables with ground terms. Replacing a variable with another variable does not accomplish that. Ground terms are useful when constructing specific instances of general expressions, but their interpretation remains a separate semantic question. Two syntactically different ground terms can refer to the same object in an interpretation.

Three practical examples

A logic-programming exercise substitutes a specific item identifier into a general term before processing a concrete query.
A theorem-proving demonstration constructs variable-free instances of a universally quantified rule.
A symbolic-data validator checks that an expression expected to represent a fully specified object description contains no unresolved variables.

Think of it this way

Think of a form with all placeholder fields filled in, including any placeholders hidden inside an attached subform. This captures the absence of variables. The analogy is limited because a filled form can still contain inaccurate information, and a ground term likewise gains no truth or real-world verification merely by being variable-free.

Worked example

Declare Box1 and DepotA to be constants, x and y to be variables, and pair and label to be function symbols. The term pair(Box1, DepotA) is ground. The term pair(Box1, x) is not, because x remains a variable. The nested term label(pair(Box1, DepotA)) is ground because every part is variable-free.
Start with t = pair(x, label(y)). Apply the substitution x maps to Box1 and y maps to DepotA. The resulting term is pair(Box1, label(DepotA)), which is ground. If the substitution instead maps x to Box1 and y to z, where z is a variable, the result remains non-ground.
Now introduce Stored as a predicate symbol. Stored(pair(Box1, DepotA)) is a ground atom, not a term. Its being ground says nothing about whether the storage assertion is true. Also, if the interpretation makes two constants refer to the same object, their different spellings do not force different referents. These distinctions separate syntax, expression category, and meaning.

Illustrative case study

In a fictional symbolic-reasoning lesson, Elm Lab labels every expression without an obvious blank as a completed fact. A reviewer points out two problems: one compound expression hides a variable inside a nested argument, and another is a term rather than an assertion.
The team introduces separate checks for expression type, variable occurrence, and whether a formula is supported by the knowledge base. The invented case illustrates a cleaner reasoning interface. It does not imply that groundness checking can establish the accuracy or completeness of a database.

Common misconceptions and mistakes

Ground does not mean true, executable, numerical, or externally verified. A variable-free compound term may be symbolic and need not be evaluated like a programming-language function call. A quantified formula can also have no free variables while still containing variables, so closed and ground are different descriptions.
Another mistake is assuming a finite vocabulary gives finitely many ground terms. With a constant a and an unrestricted unary function f, the terms a, f(a), f(f(a)), and so on continue indefinitely. Generating every ground instance can therefore be impossible to complete even from a small set of symbols.

Business use

Ground terms help explain how symbolic systems move from general patterns to particular expressions. They appear in logic programming, automated reasoning, and formal representations of structured objects. The distinction is especially useful when debugging why a supposedly concrete expression still contains unresolved variables.
For a project using symbolic rules, document the language's conventions for variables, constants, functions, and predicates. Do not rely on capitalization alone unless that convention belongs to the implementation. Keep syntactic validation separate from semantic validation and evidence checking. A complete expression may still refer ambiguously, violate a domain rule, or occur inside an unsupported assertion.

Check your understanding

Is pair(Box1, label(z)) ground when Box1 is a constant and z is a variable?

Answer: No. The variable z occurs inside label(z), so the whole compound term is non-ground even though its first argument is a constant.

Related terms

Term
Variable
Predicate (logic)
First-order logic
constant symbol
function symbol
predicate
ground atom
substitution
Herbrand universe

Learn next

Classify a small set of expressions as terms or formulas before checking groundness. Then apply substitutions to nested expressions and inspect every argument. Finally, compare a ground formula with a closed quantified formula to see why the two properties should not be conflated.

Sources

Sources: Russell and Norvig, Artificial Intelligence: A Modern Approach, first-order logic chapter, terms and ground terms: https://aima.cs.berkeley.edu/4th-ed/pdfs/newchap08.pdf

Previous: Ground resolution theorem Next: Ground truth