Let A be a non-empty set. Every SUBSET of A×A is called a BINARY RELATION on A.
ILLUSTRATIVE EXAMPLES: (1) Let A={1,2,3} and R={(1,2),(2,2),(3,1),(3,2)}. Since A×A={(1,1),(1,2),(1,3),(2,1),(2,2),(2,3),(3,1),(3,2),(3,3)} and R⊂A×A, R is a binary relation on A. (2) Let N be the natural numbers and R={(a,b)/a,b∈N and 2a+b=10}. Since R⊂N×N, R is a binary relation on N; explicitly R={(1,8),(2,6),(3,4),(4,2)}, giving domain(R)={1,2,3,4} and (co-domain restricted to what's used, i.e.) range(R)={2,4,6,8}. (3) If R′ on Z×Z is defined by 'aR′b if (a+b) is even' and R is defined by 'aRb if (a−b) is even', these two definitions actually give the SAME relation — the underlying subset is {(a,b)/a,b both even, or both odd}.
Two special extreme cases: (i) since ϕ⊂A×A, the EMPTY set ϕ is itself always a valid relation on A, called the EMPTY or VOID relation. (ii) since A×A⊂A×A (trivially), the whole of A×A is itself always a valid relation on A, called the UNIVERSAL relation on A, i.e. R=A×A. For example, if A={2,4,6}, then R=A×A={(2,2),(2,4),(2,6),(4,2),(4,4),(4,6),(6,2),(6,4),(6,6)} is the universal relation on A. …