Types of relations

A relation is nothing more than a rule for which elements of a set are linked to which — but Class 12 asks you to go further and classify that link. Is every element linked to itself? If aa is linked to bb, must bb be linked back to aa? If aa links to bb and bb links to cc, must aa link to cc? These three yes/no questions — reflexive, symmetric, transitive — are the vocabulary of the whole chapter, and when all three answers are "yes" together, the relation earns the special name equivalence relation, which quietly sorts the whole set into disjoint clusters called equivalence classes. This article builds that vocabulary from the definition of a relation as a subset of A×AA \times A, works two fully-solved examples — one on a small finite set, one on the integers under divisibility — and closes with the traps examiners love: vacuous truth, symmetric-looks-obvious, and the missing-element reflexivity trick.

The core insight

A relation on a set is reflexive, symmetric, and/or transitive depending on whether every element relates to itself, relations reverse, and relations chain — and a relation with all three properties is an equivalence relation whose equivalence classes partition the set into disjoint pieces.

The core idea

Let AA be a set and let RR be a relation on AA, meaning R⊆A×AR \subseteq A \times A. Write aRbaRb (equivalently (a,b)∈R(a,b) \in R) to say "aa is related to bb." Three properties classify how RR behaves.

  • Reflexive — every element relates to itself. Formally, for all a∈Aa \in A, (a,a)∈R(a,a) \in R.
  • Symmetric — the relation never goes one way only. Formally, (a,b)∈R  ⟹  (b,a)∈R(a,b) \in R \implies (b,a) \in R for all a,b∈Aa,b \in A.
  • Transitive — the relation chains. Formally, (a,b)∈R(a,b) \in R and (b,c)∈R  ⟹  (a,c)∈R(b,c) \in R \implies (a,c) \in R for all a,b,c∈Aa,b,c \in A.

A relation satisfying all three simultaneously is an equivalence relation:

R is an equivalence relation  ⟺  R is reflexive, symmetric, and transitive.R \text{ is an equivalence relation} \iff R \text{ is reflexive, symmetric, and transitive.}

where: AA is the underlying set, R⊆A×AR \subseteq A \times A is the relation, and each implication (  ⟹  \implies) must hold for every choice of the elements named, not just some. When RR is an equivalence relation, each a∈Aa \in A generates an equivalence class

[a]={x∈A:(x,a)∈R},[a] = \{x \in A : (x,a) \in R\},

where: [a][a] is the set of all elements related to aa. These classes have a beautiful structural property: they partition AA — every element of AA lies in exactly one class, so the classes are non-empty, pairwise disjoint, and their union is all of AA. Intuitively, RR groups AA into clusters of mutually related elements, and [a][a] is simply the cluster containing aa. A practical habit worth building early: to disprove one of the three properties, a single counterexample suffices — you never need to check every pair, only find the one pair that breaks the rule.

Figure
Three small graphs on four dots showing a reflexive relation with self-loops, a symmetric relation with two-way arrows, and a transitive relation whose chain closes up
Three properties, three pictures reflexive every dot loops to itself symmetric an arrow back for every arrow out transitive a to b, b to c forces a to c directly

Why it works: equivalence classes partition the set

Proving that the equivalence classes of an equivalence relation RR on AA partition AA into non-empty, pairwise disjoint pieces whose union is all of AA is what turns "reflexive, symmetric, transitive" from three isolated checks into one powerful structural guarantee, and every step below uses exactly one of the three defining properties, so it doubles as a demonstration of why all three are needed together.

Step 1. Every class is non-empty and covers AA. Take any a∈Aa \in A. By reflexivity, (a,a)∈R(a,a) \in R, so a∈[a]={x∈A:(x,a)∈R}a \in [a] = \{x \in A : (x,a) \in R\}. Hence aa belongs to its own class, [a][a] is never empty, and since this holds for every a∈Aa \in A, the union of all the classes is all of AA.

Step 2. If two classes share an element, they are identical. Suppose [a][a] and [b][b] share some element cc, so c∈[a]c \in [a] and c∈[b]c \in [b], meaning (c,a)∈R(c,a) \in R and (c,b)∈R(c,b) \in R. By symmetry, (c,a)∈R(c,a) \in R gives (a,c)∈R(a,c) \in R. By transitivity applied to (a,c)∈R(a,c) \in R and (c,b)∈R(c,b) \in R, we get (a,b)∈R(a,b) \in R.

Step 3. Show [a]=[b][a] = [b] using that (a,b)∈R(a,b) \in R. Take any x∈[a]x \in [a], so (x,a)∈R(x,a) \in R. Combined with (a,b)∈R(a,b) \in R from Step 2, transitivity gives (x,b)∈R(x,b) \in R, i.e. x∈[b]x \in [b]. So [a]⊆[b][a] \subseteq [b]. The identical argument with aa and bb swapped (using symmetry to get (b,a)∈R(b,a) \in R) gives [b]⊆[a][b] \subseteq [a]. Hence [a]=[b][a] = [b].

Step 4. Conclude the partition. Steps 1–3 together say: every element lies in some class (non-empty, covering), and any two classes are either identical or disjoint (since sharing even one element forces them equal). That is precisely the definition of a partition of AA — and it required reflexivity for coverage, symmetry to flip the shared relation, and transitivity to chain it into equality.

Worked example

Let A={1,2,3,4}A = \{1,2,3,4\} and R={(1,1),(2,2),(3,3),(4,4),(1,2),(2,1)}R = \{(1,1),(2,2),(3,3),(4,4),(1,2),(2,1)\}. Check whether RR is reflexive, symmetric, and transitive, and if it is an equivalence relation, find its equivalence classes.

Step 1. Reflexive? We need (a,a)∈R(a,a) \in R for every a∈A={1,2,3,4}a \in A = \{1,2,3,4\}. Scanning RR: (1,1),(2,2),(3,3),(4,4)(1,1), (2,2), (3,3), (4,4) are all present. So RR is reflexive.

Step 2. Symmetric? We need: whenever (a,b)∈R(a,b) \in R, also (b,a)∈R(b,a) \in R. The only non-diagonal pair is (1,2)(1,2), and its reverse (2,1)(2,1) is also in RR. Every diagonal pair (a,a)(a,a) trivially reverses to itself. So RR is symmetric.

Step 3. Transitive? Check every chain. From (1,2)∈R(1,2) \in R and (2,1)∈R(2,1) \in R, transitivity demands (1,1)∈R(1,1) \in R — true. From (2,1)∈R(2,1) \in R and (1,2)∈R(1,2) \in R, it demands (2,2)∈R(2,2) \in R — true. All remaining chains only involve pairs of the form (a,a)(a,a), which close up trivially. So RR is transitive.

Step 4. Conclude. RR is reflexive, symmetric, and transitive, so it is an equivalence relation on AA. Applying the partition proof above: [1]=[2]={1,2}[1] = [2] = \{1,2\} (since 11 and 22 are mutually related), while 33 and 44 relate to nothing but themselves, so [3]={3}[3] = \{3\} and [4]={4}[4] = \{4\}. The equivalence classes are {1,2},{3},{4}\{1,2\}, \{3\}, \{4\} — three disjoint sets whose union is AA.

Competitive stretch

On Z\mathbb{Z}, define aRbaRb iff 5∣(a−b)5 \mid (a-b). Reflexive: a−a=0=5×0a-a=0=5\times 0. Symmetric: a−b=5k  ⟹  b−a=5(−k)a-b=5k \implies b-a=5(-k). Transitive: a−b=5k,b−c=5m  ⟹  a−c=5(k+m)a-b=5k, b-c=5m \implies a-c=5(k+m). So RR is an equivalence relation, and [2]={…,−8,−3,2,7,12,17,…}[2] = \{\ldots,-8,-3,2,7,12,17,\ldots\} — exactly the five classes [0],[1],[2],[3],[4][0],[1],[2],[3],[4] partition Z\mathbb{Z} into remainders mod 5, matching the partition proof above elements-for-elements: coverage (every integer has a remainder), and disjointness (no integer has two different remainders mod 5).

Common traps

  • Treating reflexivity as "obviously true" without checking it — verify (a,a)∈R(a,a) \in R explicitly for every element of the stated set AA, not just a couple.
  • Confusing "looks symmetric informally" with the actual pair check — (1,2)∈R(1,2) \in R does not make RR symmetric unless (2,1)∈R(2,1) \in R is also listed.
  • Forgetting that transitivity is tested only where a chain exists — if no two pairs in a finite relation actually chain together, transitivity holds vacuously (trivially true), which students often mistake for "not transitive."
  • Assuming symmetric-and-transitive automatically implies reflexive — this fails whenever some element of AA never appears as a coordinate of any pair in RR, a favourite exam trick.

Ask the tutor: Want to work through a trickier relation defined by a formula (like "aRba R b if a2=b2a^2 = b^2") to practice spotting whether it's reflexive, symmetric, and transitive?

Why does a relation being symmetric and transitive not automatically make it reflexive?

Symmetry and transitivity only make demands about elements that already appear as coordinates of some pair in R. If an element a of A never shows up in any ordered pair of R at all, there is nothing to force (a,a) into R, so reflexivity can fail silently for that lonely element. The classic trap example is A equals one, two, three with R equal the single pair two, three and three, two — symmetric and transitive, but not reflexive since one never appears and even two, two is missing.

What exactly does vacuous truth mean for transitivity, and why does it trip people up?

Transitivity says: whenever (a,b) and (b,c) are both in R, then (a,c) must be in R too. If no such matching pair of pairs ever occurs in R, the "whenever" clause is never triggered, so the implication is true by default — this is called vacuous truth. Students often see a small relation with few pairs, find no chains to check, and wrongly conclude "not transitive" when the honest answer is "transitive, vacuously." Always ask: does a chain exist at all? If not, the property holds automatically.

How is an equivalence class different from just "a subset of A"?

Any subset of A is just a subset, but an equivalence class [a] is generated by the relation itself: it is the set of every element related to a. What makes equivalence classes special is the partition property proved above — every element of A falls into exactly one class, so the classes never overlap and together they use up all of A. A random subset has no such guarantee; equivalence classes are forced into this clean, non-overlapping structure precisely because R is reflexive, symmetric, and transitive together.

Before this

Where this leads