Basic Consequences of the Group Axioms

The four basic consequences proved here are used, silently and constantly, in every branch of mathematics touching symmetry. In algebra itself, cancellation and unique solvability are what make it legitimate to "solve for $x$" in group equations, a step used without comment in the classification of finite simple groups and in the study of Subgroups and the Subgroup Criterion, where checking closure and inverses on a subset relies on these basic facts holding automatically inside the ambient group. In geometry, the uniqueness of inverses underlies the fact that every rigid motion (rotation, reflection, translation) has one and only one motion that undoes it, a fact taken for granted in crystallography and robotics when composing and reversing transformations. In physics, groups of symmetries (rotation groups, Lorentz group, gauge groups) rely on cancellation to correctly manipulate conserved-quantity equations and to define inverse transformations between reference frames unambiguously. In cryptography, especially in Diffie–Hellman key exchange and elliptic-curve cryptography, the order of a group element is the central quantity controlling security. In computer science, the well-definedness of order underlies algorithms for computing discrete logarithms and for detecting cycles in permutation-based hash functions and pseudorandom generators. In statistics and coding theory, cyclic groups built from elements of specific finite order (as in Example 2 above) underlie error-correcting codes and design of experiments via group-based combinatorial structures.

Ask the tutor: Why does the proof that inverses are unique break down completely without associativity, even if closure, identity, and (two-sided) inverses all still hold?

The group axioms are deliberately economical: they list only what an operation must do, not everything it happens to do in a specific example. The remarkable fact — and the content of this article — is that this bare minimum of rules already forces rigid, predictable behavior. Think of the axioms as a small set of load-bearing rules in a game; the theorems below show that certain "obvious" features of the game (there's only one starting position, every move has exactly one undo) are not separate rules at all, but are forced by the four rules already given. A useful mental picture is a Cayley table (multiplication table) for a finite group. Associativity and inverses are invisible constraints hiding behind the table's surface, but they force each row and each column of the table to be a permutation of $G$ — no repeats. This is exactly the cancellation law in disguise: if $ab = ac$ appeared as two entries equal to each other in row $a$, cancellation says $b=c$, so a row cannot repeat an entry. This rigidity is what lets us reason about groups abstractly, without knowing anything about the specific elements.

Order, meanwhile, is best pictured as the "period" of repeated application of an element: apply $g$ again and again, $g, g^2, g^3,\dots$, and ask when you first return to the identity. If you never return, the element has infinite order, like translation by $1$ in $(\mathbb{Z},+)$; if you return after exactly $n$ steps, you have found a finite cycle of length $n$ sitting inside the group.

Example 1: Cancellation and unique solvability in a non-abelian group, $S_3$. Let $G = S_3$, the symmetric group on $\{1,2,3\}$, with composition of functions (right-to-left) as the operation, so it is not commutative — a good stress test since the theorem claims no commutativity is needed. Let

Here $a$ is the 3-cycle sending $1\to2\to3\to1$, and $b,c$ are transpositions. We verify part (d) (left cancellation) concretely. First compute $a\ast b$ (apply $b$ first, then $a$): $b$ sends $1\to2,2\to1,3\to3$; then $a$ sends $2\to3$, $1\to2$, $3\to1$. Tracking $1\mapsto 2\mapsto 3$, $2\mapsto1\mapsto2$, $3\mapsto3\mapsto1$, we get

Now suppose, hypothetically, we are told $a\ast b = a\ast x$ for some unknown $x \in S_3$ and we want to recover $x$ using only the cancellation law, without recomputing directly. By part (e), the unique solution to $a\ast x = (1\,3)$ is $x = a^{-1}\ast(1\,3)$. Here $a^{-1} = (1\,3\,2)$ (the inverse 3-cycle, sending $1\to3\to2\to1$), since one checks $a\ast a^{-1}$: applying $a^{-1}$ then $a$ sends $1\to3\to1$, $2\to1\to2$, $3\to2\to3$, i.e., the identity permutation, confirming $a^{-1}=(1\,3\,2)$ is correct by direct verification of the defining property. Compute $x = a^{-1}\ast(1\,3)$: apply $(1\,3)$ first (swap $1,3$, fix $2$), then $a^{-1} = (1\,3\,2)$ (sends $1\to3,3\to2,2\to1$). Tracking: $1\mapsto 3\mapsto2$; $2\mapsto2\mapsto1$; $3\mapsto1\mapsto3$. So

Indeed $x = b = (1\,2)$, matching our original computation of $a\ast b=(1\,3)$ exactly as cancellation predicted: from $a\ast b = a\ast x$ we correctly recovered $x=b$. This confirms the cancellation law is not a formal trick — it genuinely reconstructs the unique correct answer even in a non-commutative setting where "dividing by $a$ on the left" must be done carefully (multiplying by $a^{-1}$ on the left of both sides, not on the right).

Example 2: Order of an element in $\mathbb{Z}/12\mathbb{Z}$, and a contrast with infinite order. Let $G = \mathbb{Z}/12\mathbb{Z} = \{0,1,\dots,11\}$ under addition mod $12$, with identity $e=0$. Take $g = 8$. "Powers" here are additive, so $g^n$ means $n\cdot 8 \bmod 12$ (repeated addition, matching the multiplicative notation of the general theorem with $\ast$ replaced by $+$). We compute successive multiples of $8$ mod $12$:

So $3\cdot 8 \equiv 0 \pmod{12}$, and no smaller positive multiple works ($1\cdot 8=8\ne0$, $2\cdot 8=4\ne 0$). By the definition in part (g), the set $S = \{n>0 : n\cdot 8 \equiv 0\}$ is nonempty (it contains $3$), and its least element is $3$. Hence

Notice this is asymmetric and non-obvious: $8$ is not a "small" or "symmetric" element of $\{0,\dots,11\}$, yet its order is a small number, $3$, illustrating that order depends on arithmetic (here, on $\gcd(8,12)=4$ and $12/4=3$) rather than on the numeric size of the element. Contrast this with $G = (\mathbb{Z}, +)$, the integers under addition, identity $e = 0$, and take $g = 8$ again. Now $S = \{n > 0 : n\cdot 8 = 0\}$. Since $\mathbb{Z}$ has no zero divisors in this additive sense (if $n>0$ and $8\ne0$ then $n\cdot 8 \ne 0$, an elementary fact about integer arithmetic used here as background), $S = \emptyset$. By part (g), $|8| = \infty$ in $(\mathbb{Z},+)$. This shows the same element $8$ can have finite order in one group and infinite order in another — order is a property of the element together with the ambient group and operation, not of the element in isolation, and the definition in part (g) correctly handles both outcomes via the well-ordering dichotomy.