A property of expressions where they can be replaced with their values without changing the program's behavior, meaning the expression always produces the same output for the same inputs with no side effects. This enables powerful optimization and reasoning techniques.
The term was introduced by philosopher Willard Van Orman Quine in the 1960s, later adopted into programming language theory. 'Referential' comes from Latin 'referre' meaning to relate, while 'transparency' suggests the ability to see through to the underlying value.
Referential transparency is like having mathematical equations in your code - just like you can replace '2 + 3' with '5' anywhere without changing the meaning, you can replace transparent expressions with their results! This is why functional programmers love pure functions - they behave like math.
Complete word intelligence in one call. Free tier — 50 lookups/day.