For a relation on a single set A, the most revealing picture is the directed graph. Draw one node for each element of A. Draw one arrow from a to b for each pair (a, b) \in R. That is the entire construction. What makes this picture powerful is that every major property of the relation — reflexive, symmetric, transitive, antisymmetric — becomes a visual pattern you can recognise at a glance.
The picture
Take A = \{1, 2, 3, 4\}. A relation on A is any subset of A \times A = \{ all ordered pairs of elements from A \} — sixteen possible pairs in this case. The directed graph shows:
- Four nodes, one per element.
- Up to sixteen arrows — one for each possible pair, drawn only when the pair is in the relation.
- Self-loops (arrows from a node to itself) when (a, a) \in R.
- Two-way arrows (both (a, b) and (b, a) present) drawn as two separate arrows or as a double-headed edge, depending on convention.
How each property reads off the graph
Reflexive: every node has a self-loop.
Irreflexive: no node has a self-loop.
Symmetric: for every straight arrow a \to b (with a \neq b), the reverse arrow b \to a is also present. Self-loops are trivially symmetric.
Antisymmetric: no two distinct nodes have two-way arrows between them. An arrow from a to b excludes the reverse unless a = b.
Transitive: for every two-arrow path a \to b \to c, the shortcut arrow a \to c is also drawn. Self-loops participate in chains but never break them.
Equivalence relation: reflexive + symmetric + transitive. Visually: every self-loop drawn; every arrow has its reverse; every chain closes. The graph breaks into disjoint "cliques" — groups of nodes with complete back-and-forth arrows — one clique per equivalence class.
Partial order: reflexive + antisymmetric + transitive. Visually: self-loops everywhere, no two-way arrows between distinct nodes, and every chain closes. The graph looks like a hierarchy with a clear direction.
Five sample cases the slider walks through
Case 0 — empty relation. No arrows at all. Irreflexive, symmetric (vacuous), antisymmetric (vacuous), transitive (vacuous). Not reflexive (no self-loops), hence not an equivalence relation.
Case 1 — identity relation \{(1,1), (2,2), (3,3), (4,4)\}. Only self-loops. Reflexive, symmetric, antisymmetric, transitive. The smallest equivalence relation; every element is its own class.
Case 2 — same parity (the picture shown). Two cliques (\{1, 3\} and \{2, 4\}), each with self-loops and two-way arrows. Equivalence relation. Partitions A into odds and evens.
Case 3 — strict less-than, \{(1,2), (1,3), (1,4), (2,3), (2,4), (3,4)\}. Arrows only in one direction. Irreflexive (no self-loops), antisymmetric, transitive. Not a partial order because reflexivity fails — this is a strict partial order.
Case 4 — full relation A \times A. Every possible arrow. Reflexive, symmetric, transitive. An equivalence relation (everyone is in one big class). Also not antisymmetric (since two-way arrows exist between distinct nodes).
What the graph emphasises that the list hides
When a relation is given as a list of pairs, properties like symmetry and transitivity can be tedious to check — you scan pair by pair. The directed graph compresses the information spatially, so patterns pop:
- Symmetry becomes mirror symmetry (every arrow has a twin). Your eye spots missing twins fast.
- Reflexivity becomes "every node has a loop." A missing loop is a gap in an otherwise uniform pattern.
- Transitivity becomes "every chain closes." Chains appear as paths in the graph, and a missing shortcut is a visible gap.
- Equivalence becomes "disjoint cliques." The graph partitions visually into clumps.
Why: a directed graph lays out every pair in two dimensions rather than stringing them in a list. Patterns that are global (symmetry, transitivity) become local (twin arrows, closed triangles) when drawn spatially. The graph is the relation, rendered in a form that foregrounds its structure.
Conventions for drawing
For small sets (|A| \leq 6 or so), arrange the nodes in a circle or at the corners of a polygon. This keeps arrows from tangling and makes chains easier to trace.
For larger sets, or for relations with heavy structure, the reduction diagram (Hasse diagram for partial orders) drops redundant arrows — specifically, every transitive shortcut is erased and every self-loop is implicit. This is a cleaner picture for posets, but you lose the ability to read reflexivity and transitivity visually. Use the full directed graph when studying properties; use the Hasse diagram when the relation is a known partial order and you want to see its hierarchy.
Arrows are typically drawn as curves when there are two between the same pair of nodes (one for each direction), with arrowheads near the target. Self-loops are small circles or teardrop shapes attached to a single node.
Spotting properties — a walk-through
Take the relation R = \{(1,1), (2,2), (1,2), (2,1), (1,3)\} on A = \{1, 2, 3\}. Draw the graph:
- Node 1 has a self-loop (\checkmark).
- Node 2 has a self-loop (\checkmark).
- Node 3 has no self-loop. Not reflexive. Counter-example: (3, 3) \notin R.
- Arrows 1 \to 2 and 2 \to 1: both present. 1 \to 3: present, 3 \to 1: missing. Not symmetric. Counter-example: (1, 3) \in R but (3, 1) \notin R.
- Chain 2 \to 1 \to 3: needs (2, 3). Missing. Not transitive. Counter-example: (2, 1), (1, 3) \in R but (2, 3) \notin R.
None of the three big properties hold. This relation fits no named category — it is just a generic relation.
Contrast with R = \{(1,1), (2,2), (3,3)\}: all three self-loops, no other arrows. Reflexive. Symmetric (vacuously, no non-loop arrows to check reverses for). Transitive (vacuously). Equivalence relation. The identity relation.
Why the graph picture generalises
Everything you learn from drawing directed graphs of small relations transfers directly to:
- Network analysis. Social networks, citation graphs, and web link graphs are directed graphs in disguise — each edge is a "relates to" pair.
- Graph theory. Vertices and edges are the building blocks of an entire branch of mathematics, and a relation on A is exactly a simple directed graph with vertex set A.
- Computer science. Dependency graphs (which tasks must happen before which), state machines (which state transitions are allowed), and finite automata all use directed graphs to encode relations.
The habit of reading properties off pictures, trained on small relations, generalises to every domain where "which things relate to which" is the central question. It is a core visual literacy for modern mathematics and computing.
The reflex
- Sketch the directed graph on rough paper. Nodes in a convenient layout, arrows for every pair.
- Scan for self-loops → reflexivity check.
- Scan for twin arrows → symmetry check.
- Scan for two-hop chains and their shortcuts → transitivity check.
- Identify connected clumps with full internal connectivity → equivalence-class pattern.
Once drawn, every major property is visible in under thirty seconds for small sets. For JEE-level questions, small sets are the norm, and the graph picture is routinely the fastest path to the answer.
Related: Relations · Reflexivity Tester: Missing Self-Loops Glow Red · Symmetry Check: Missing Reverse Arrows Glow Red · Transitivity Chain Finder