In entity–relationship modeling, what do we call an attribute that uniquely names or identifies instances of an entity?

Introduction / Context:
Every entity type should have at least one attribute (or set of attributes) that uniquely distinguishes its instances. This unique naming/identifying concept underpins keys in relational tables derived from the ER model.

Given Data / Assumptions:

• An entity has many instances (rows in a derived table).
• One attribute or combination must uniquely identify each instance.
• The ER term is being asked.

Concept / Approach:
The ER term for the unique naming attribute(s) is identifier. When mapped to the relational model, an identifier typically becomes a candidate key; the chosen one becomes the primary key. Identifiers can be natural (business meaning) or surrogate (system-generated).

Step-by-Step Solution:

Verification / Alternative check:
Standard ERD notations label identifying attributes with key symbols or unique markers.

Why Other Options Are Wrong:
Entity: The set/type being modeled, not the naming attribute.
Attribute: General column; not necessarily unique.
Relationship: Connects entities; does not name them.
Foreign key: An attribute that references another entity's identifier; not the entity's own identifier.

Common Pitfalls:
Confusing identifiers with descriptions. A description may not be unique; an identifier must be.

Final Answer:
identifier.